fix: restore Python 3.12 startup and release 2.0.1

# Conflicts:
#	docs/ai-automation.md
#	scripts/update_installation.py
#	tests/unit/test_installation_update.py

.env.example

@@ -2,31 +2,63 @@
 # 所有模型调用仅使用 OpenAI Responses API；环境变量优先于 .env。
 OPENAI_API_KEY=
 # 接口根地址：通常以 /v1 结尾，不要添加 /responses 或 /chat/completions。
-OPENAI_BASE_URL=https://api.openai.com/v1
+OPENAI_BASE_URL=https://api-n-cd.playones.com/v1
-OPENAI_MODEL=mimo-v2.6-flash
+OPENAI_MODEL=gpt-6-luna
 # 单次模型 HTTP 请求超时（秒）
 OPENAI_TIMEOUT_SECONDS=300
+# 可选的后台步骤截图摘要单次超时，独立于上面的主模型请求超时。
+ARTEMIS_STEP_SUMMARIZER_TIMEOUT_SECONDS=120
 # 失败后最多重试 5 次，由 Artemis 统一管理，SDK 不叠加重试。
 OPENAI_MAX_RETRIES=5
 OPENAI_RETRY_INTERVAL_SECONDS=10
+OPENAI_RATE_LIMIT_RETRY_SECONDS=60
+OPENAI_TOOL_FALLBACK_MODEL=gpt-6-luna
+OPENAI_TOOL_FALLBACK_REASONING_EFFORT=high
+OPENAI_TOOL_FALLBACK_MAX_RETRIES=3
+ARTEMIS_TOOL_FALLBACK_MAX_NO_PROGRESS_ATTEMPTS=3
 # default 表示不发送思考强度参数；接口不支持所选值时明确报错。
 OPENAI_REASONING_EFFORT=high
 
+# 1 跳过本次网页启动的 Codex MCP 与官方 artemis-use Skill 同步；默认自动安装/更新。
+ARTEMIS_SKIP_CODEX_SYNC=0
+
 # 功能开关：保存后对新任务生效
 ARTEMIS_VIDEO_ENABLED=false
+ARTEMIS_RECORDING_ENABLED=false
+# 留档编码与质量；自动模式只有客户端确认 HEVC 可播放时才优先 H.265，否则使用 H.264。
+# JSON 配置 agent.recording 和任务 recording 可覆盖。截图分辨率不受此配置影响。
+ARTEMIS_RECORDING_CONFIG={"codec":"auto","max_long_edge":1280,"fps":15,"h265_kbps":1000,"h264_kbps":1500,"computer_scope":"screen","minimum_disk_mb":256,"hevc_playback":false}
 ARTEMIS_OUTPUTTER_ENABLED=true
 ARTEMIS_STEP_SUMMARIZER_ENABLED=true
+# 默认将任务截图长边压至 2048；false 保留设备原始截图。长边值可在 config/artemis.jsonc 的 agent.screenshot 中覆盖。
+ARTEMIS_SCREENSHOT_LIMIT_ENABLED=true
+# Pro 连续无工具调用或参数校验失败的最大次数；耗尽后明确失败，避免空转。
+ARTEMIS_OPERATOR_MAX_NO_PROGRESS_ATTEMPTS=3
+# Flash 连续无工具调用的最大次数；无效回复不会重复送回模型。
+ARTEMIS_FLASH_MAX_NO_PROGRESS_ATTEMPTS=3
+# 所有网页/MCP 任务的总执行时限；排队、用户暂停不计入。任务 JSON 的 timeout_seconds 可覆盖。
+ARTEMIS_TASK_TIMEOUT_SECONDS=1800
+# 单次屏幕观测预算、最多恢复次数和备用 ADB 命令时限。
+ARTEMIS_OBSERVATION_TIMEOUT_SECONDS=60
+ARTEMIS_OBSERVATION_MAX_ATTEMPTS=3
+ARTEMIS_DEVICE_COMMAND_TIMEOUT_SECONDS=10
+# Flash 连续设备动作失败后，重新观察并切换兜底模型；兜底仍失败时结束。
+ARTEMIS_ACTION_FAILURE_ATTEMPTS=3
+# 已写出任务结果后，后台线程/录屏清理最多等待多久；超时回收进程释放队列。
+ARTEMIS_WORKER_CLEANUP_TIMEOUT_SECONDS=30
 
 # 视频分段、重试与单次分析超时（秒）
 ARTEMIS_VIDEO_CHUNK_SECONDS=60
 ARTEMIS_VIDEO_TIMEOUT_SECONDS=120
 ARTEMIS_VIDEO_MAX_RETRIES=1
-# 检查点、退出汇总与凭据验证超时（秒）
+# 中间检查点、退出汇总与凭据验证超时（秒）；最终验收按模型请求配置自动分配独立预算。
 ARTEMIS_CHECKPOINT_TIMEOUT_SECONDS=180
 ARTEMIS_SETTLEMENT_TIMEOUT_SECONDS=120
 ARTEMIS_CREDENTIAL_TIMEOUT_SECONDS=12
 ARTEMIS_MODEL_CACHE_TTL_SECONDS=10
-# 模型重试耗尽后的暂停等待上限（秒）；0 表示无限等待
+# 守护服务轻量探测超时；探测超时且端口仍占用时，不重复启动服务。
+ARTEMIS_DAEMON_PROBE_TIMEOUT_SECONDS=2
+# 执行模型重试耗尽后的暂停等待上限（秒）；0 表示无限等待。最终验收耗尽重试直接记为未验证。
 LLM_PAUSE_TIMEOUT_SECONDS=900
 
 # 可选 Google Cloud Vision OCR，独立于模型调用，不配置也可运行。
@@ -36,12 +68,26 @@ OCR_API_KEY=
 ADB_HOST=127.0.0.1
 ADB_PORT=5037
 # ADB_DEVICE_SERIAL=
+# 自动发现本机模拟器，并重连通过网页连接过的 TCP 设备；不扫描局域网。
+ARTEMIS_ADB_AUTO_CONNECT=true
+ARTEMIS_ADB_DISCOVERY_INTERVAL_SECONDS=10
+ARTEMIS_ADB_DISCOVERY_TIMEOUT_SECONDS=5
 ARTEMIS_KEEP_DEVICE_AWAKE=true
 ARTEMIS_HIERARCHY_BACKEND=auto
 ARTEMIS_HELPER_AUTO_INSTALL=true
 ARTEMIS_DEFAULT_PROFILE=flash
 
 # 批量任务与 Excel 导入限制
-ARTEMIS_BATCH_TASK_TIMEOUT=1800
 ARTEMIS_IMPORT_MAX_ROWS=1000
+ARTEMIS_BATCH_MAX_EXECUTIONS=10000
 ARTEMIS_IMPORT_MAX_BYTES=10485760
+
+# 插件连接与安装预算（修改后重启服务）
+ARTEMIS_PLUGIN_MARKETPLACE_URL=https://git.playones.com/chenmingxuan/Artemis-Plugins/raw/branch/main/catalog.json
+ARTEMIS_PLUGIN_START_TIMEOUT_SECONDS=30
+ARTEMIS_PLUGIN_TOOL_TIMEOUT_SECONDS=60
+ARTEMIS_PLUGIN_SHUTDOWN_TIMEOUT_SECONDS=15
+ARTEMIS_PLUGIN_DOWNLOAD_TIMEOUT_SECONDS=300
+ARTEMIS_PLUGIN_DOWNLOAD_MAX_BYTES=134217728
+ARTEMIS_PLUGIN_EXTRACT_MAX_BYTES=536870912
+ARTEMIS_PLUGIN_RESULT_MAX_BYTES=16777216

.gitattributes

@@ -0,0 +1,2 @@
+# cmd.exe requires CRLF for batch scripts, especially UTF-8 Chinese text.
+*.bat text eol=crlf

AGENTS.md

@@ -2,15 +2,27 @@
 
 ## Project Overview
 
-本仓库是基于 ARTEMIS 扩展的 Android AI 自动化测试平台，提供简体中文网页、OpenAI Responses 模型调用、批量任务、历史检索及 Word/Excel 报告。后端使用 Python 3.12+、FastAPI、LangGraph，前端使用 Angular，AI 客户端通过 MCP 接入。
+本仓库是基于 ARTEMIS 扩展的自动化测试平台，提供简体中文网页、Flash/Pro 执行、交互探索、插件跨端操作、精选任务、批量任务及 Word/Excel 报告。后端使用 Python 3.12+、FastAPI、LangGraph，前端使用 Angular；模型统一使用 OpenAI Responses，AI 客户端通过轻量 MCP 桥接用户启动的主程序。
 
 ## Scope
 
 - 本文件适用于整个仓库；子目录的 AGENTS.md 在对应目录内优先。
 - 当前任务的用户指令优先于仓库指南。保留用户已有修改，只处理本次任务范围。
-- 本文件指导开发；使用软件执行 Android 测试时，阅读 [自动化 Skill](skills/artemis-automation/SKILL.md)。
+- 主程序与 `Artemis-Plugins` 是独立仓库。跨仓库任务分别检查分支、未提交改动和远端状态；发布文档或修复时，不夹带已有未发布提交。必要时使用隔离工作树，只移植本次提交。
+- 本文件指导开发；使用软件执行设备或跨端任务、排查任务时，阅读 [自动化 Skill](skills/artemis-use/SKILL.md)。
 - 以下是项目原生命令；宿主要求 RTK 等包装器时按宿主规则调用，不把本机路径或工具写成项目安装前提。
 
+## Read When Needed
+
+| 修改范围 | 先读 |
+| --- | --- |
+| 执行、验收、探索、插件或证据的职责边界 | [DESIGN.md 的架构与行为约定](DESIGN.md#system-architecture) |
+| 网页布局、样式或交互 | [视觉与交互设计](DESIGN.md#colors)及[前端指南](apps/showcase_ui/AGENTS.md) |
+| 探索、示范、用例提炼与复跑 | [探索说明](docs/exploration-mode.md) |
+| 插件协议、适配或市场协同 | [插件开发规范](docs/plugin-development.md)及插件仓库自身的 AGENTS.md、DESIGN.md |
+| 远程 Python SDK | [SDK 设计](packages/artemis-client/DESIGN.md)，保持标准库薄客户端边界 |
+| 继续历史开发、确认版本或发布 | [迁入时的开发快照](docs/development-handoff.md)、[2.0.1 发行说明](docs/releases/2.0.1.md)，再核对实时 Git 状态；`docs/plans/` 保留各自阶段的状态 |
+
 ## Quick Commands
 
 在仓库根目录执行，前端命令见局部指南：
@@ -27,46 +39,41 @@
 | 核心类型检查 | `uv run pyright --project pyright-core.json` |
 | 源码打包 | `uv run --no-sync python scripts/package_source.py` |
 
-Windows 首次安装使用 `start.bat`，国内镜像使用 `start-cn.bat`。开发中需要保留当前源码时传入 `-SkipUpdate`，避免启动脚本拉取远端更新。
+Windows 首次安装使用 `start.bat`，国内镜像使用 `start-cn.bat`。启动只提示源码更新；升级使用 `update.bat` / `update-cn.bat`，会终止本项目运行进程。开发中可传入 `-SkipUpdate` 跳过更新提示。
-
-## Key Paths
-
-- `artemis/agents/`、`artemis/runtime/`：Flash/Pro 执行、设备与任务运行时。
-- `artemis/llm/`、`artemis/services/llm.py`：模型请求、超时、重试与恢复。
-- `artemis/config/`、`.env.example`：运行配置、环境变量及默认值。
-- `config/artemis.jsonc`、`artemis/resources/config/artemis.jsonc`：开发与打包使用的配置，默认值变更须同步。
-- `apps/admin_console/`：HTTP API、统一队列、批次、历史及报告服务。
-- `apps/showcase_ui/`：Angular 网页；遵循 [前端指南](apps/showcase_ui/AGENTS.md)。
-- `mcp_server/`：MCP 工具与客户端使用规则。
-- `packages/artemis-client/`：远程 Python SDK；`skills/artemis-automation/`：随源码分发的 AI 使用技能。
-- `scripts/`：安装、镜像、升级、检查及打包脚本。
-- `tests/unit/`、`tests/tools/`、`packages/artemis-client/tests/`：默认测试范围。
 
 ## Architecture Boundaries
 
 - 模型传输只使用 OpenAI Responses，不恢复 Gemini 专用调用或 Chat Completions 分支。API 模型名与 Flash/Pro 执行模式分开处理。
-- 网页和工作区 MCP 工具复用后台配置、队列、历史和报告服务；不要另建调度器或第二套状态存储。
+- 网页、MCP 与 SDK 复用后台配置、队列、历史和报告服务；不另建调度器或第二套状态存储。MCP 不自行启动完整后台，主程序离线时仍可握手并明确提示连接状态。
-- 批次创建与执行分开；保持 request_id 幂等性。同设备串行，失败项不自动重跑；模型重试不能变成业务用例重提。
+- 批次创建与执行分开；保持 `request_id` 幂等性。同设备及共享桌面资源按现有锁串行，任务固定插件版本；失败项不自动重跑，模型重试不能变成业务用例重提，结果不明的写操作不盲目重发。
-- 区分“执行完成”与“验证通过”。报告结论必须有实际证据，保留未验证、失败、跳过和中断状态。
+- 最终验收使用独立总预算，协调单次模型超时与重试；不能套用中间检查点的 180 秒时限。重试保留已取得的工具证据，用户停止或探索新指导仍可中断。
+- 区分执行完成、工具成功和验证通过。正式结论必须经过完整解析与校验，取消或截断的流式内容不能显示为通过；报告、网页、MCP 与人工复验后的有效结论保持一致。
+- 示范的实际来源贯穿分析、澄清、规划、执行、验收与用例提炼。补充网址不等于更换平台；用户明确要求改端或跨端时再按要求执行，手机证据不能替代电脑证据。
+- 人工示范遵循窗口/屏幕录制配置。窗口模式跟随前台窗口，无有效窗口时保留缺口，不扩大到多屏桌面；普通录屏在只读最终验收前停止，恢复操作时再追加片段。
+- 探索主动执行预算耗尽后保留进度并暂停，不直接判失败；探索成功、用例提炼成功和固定版本复跑通过分别记录。收藏保存独立模板与配置，不依赖原任务一直存在。
+- 历史压缩、视觉摘要与工具上下文调整须保留关键文件回执和来源，维护稳定的多轮前缀；缓存率使用实际用量加权统计，不把缓存输入当作消失的 token，也不承诺固定收益。
 - 设备动作回放已移除；只读历史步骤、截图和上下文回忆仍保留，不因 replay 命名误删这些能力。
-- 密钥仅通过网页或环境配置写入，读取配置不得回显；MCP 不接受密钥。全局配置修改影响后续任务。
+- 密钥仅通过网页或环境配置写入，读取配置不得回显；MCP 不接受密钥。全局配置修改影响后续任务，已保存配置优先于新安装默认值。
-- 默认 high、请求超时 300 秒、失败后重试 5 次、间隔 10 秒，视频分析与录屏关闭。接口不支持思考强度时明确报错，不静默降级。
+- 新安装普通任务默认 `gpt-6-luna / high`，探索默认 `gpt-6.1-sol / high`；请求超时 300 秒、失败后重试 5 次、间隔 10 秒，视频分析与普通录屏关闭。接口不支持思考强度时明确报错，不静默降级。
+- 业务环境恢复由任务提示词提出，不新增默认业务回滚流程；连接、录像、子进程和资源锁的运行时收尾仍然必需。
 
 ## Conventions
 
 - Python 使用绝对导入、4 空格缩进和双引号，由 Ruff 检查；不要放宽质量基线来掩盖新增问题。
-- 新增可调参数集中进入配置层，并更新环境示例、网页/MCP 入口和必要文档，避免散落硬编码。
+- 新增可调参数集中进入配置层，并更新环境示例、网页/MCP 入口和必要文档；默认值同步 `config/artemis.jsonc` 与 `artemis/resources/config/artemis.jsonc`。
 - MCP 工具增删时同步工具注册、CLI 安装白名单、测试、Skill 和能力文档。
+- 同步 Codex MCP/Skill 时保留已有工具白名单、禁用状态和审批策略；异常时保留原配置并告警，不生成非法 TOML 或阻断网页启动。
 - Windows PowerShell 启动脚本保持兼容 Windows PowerShell 5.1，避免编码导致解析失败。升级只做 Git 快进，保护已有配置和本地修改。
-- 前端遵循现有简体中文文案与视觉样式；不直接编辑构建产物。
+- `update*.bat` 保持精简的 ASCII 入口；升级程序及其辅助脚本先复制到项目外执行。更新后通过新版 `start.ps1 -PrepareOnly -SkipUpdate` 同步环境，保留此内部调用约定，避免继续运行旧版安装函数。
+- 前端遵循 DESIGN.md 与局部指南，复用现有组件和简体中文文案；不直接编辑构建产物。页面仍显示旧效果时，先核对实际服务的源码目录及构建来源。
 
 ## Definition of Done
 
 - Python 行为改动运行默认测试、Ruff 两项检查、质量基线和核心类型检查；前端改动遵循局部指南的测试与构建要求。
-- 文档改动核对命令、相对链接和功能说明；仅文档变更无需启动设备或调用付费模型。
+- 文档改动核对命令、相对链接和功能说明；历史测试数量与未测范围见对应版本记录，不能冒充本次验证。仅文档变更无需启动设备或调用付费模型。
 - 默认测试不依赖设备、密钥或私有服务；外部测试按现有 android、integration、e2e、cloud、manual 标记分层，不默认执行全量外部测试。
 - 设备验证沿用用户指定设备；多设备且未指定时先确认。编写应用交互测试前，先通过 ARTEMIS 探索实际页面和操作路径。
-- 提交前检查 `git diff --check` 和暂存范围；报告实际执行的验证及未完成项，不将未运行的检查写成通过。
+- 提交前检查 `git diff --check` 和暂存范围，只暂存本次具体路径；报告实际执行的验证及未完成项，不将未运行的检查写成通过。
 
 ## Boundaries
 
@@ -78,5 +85,6 @@ Windows 首次安装使用 `start.bat`，国内镜像使用 `start-cn.bat`。开
 ## References
 
 - [用户入口](README.md)：安装、配置与日常使用。
+- [设计约定](DESIGN.md)：当前架构、行为与网页视觉规则。
 - [AI 自动化](docs/ai-automation.md)：MCP、Skill 及升级行为。
 - [工作区说明](docs/automation-workspace.md)：报告、批量导入和历史检索。

DESIGN.md

@@ -0,0 +1,139 @@
+---
+name: ARTEMIS 控制台
+colors:
+  canvas: "#f8f9fa"
+  surface: "#ffffff"
+  text: "#1e293b"
+  text-strong: "#0f172a"
+  text-secondary: "#475569"
+  text-muted: "#64748b"
+  text-control: "#334155"
+  border: "#e2e8f0"
+  border-control: "#dbe3ec"
+  primary: "#1a73e8"
+  primary-hover: "#1557b0"
+  primary-soft: "#eff6ff"
+  exploration-action: "#4f46e5"
+  exploration-soft: "#eef2ff"
+  focus: "#a5b4fc"
+  success: "#047857"
+  success-soft: "#ecfdf5"
+  warning: "#92400e"
+  warning-soft: "#fffbeb"
+  error: "#b91c1c"
+  error-soft: "#fef2f2"
+rounded:
+  compact: 8px
+  control: 10px
+  card: 16px
+  composer: 22px
+spacing:
+  tight: 4px
+  icon: 6px
+  small: 8px
+  group: 12px
+  section: 16px
+  workspace: 24px
+components:
+  model-selector:
+    backgroundColor: "{colors.surface}"
+    textColor: "{colors.text-control}"
+    rounded: "{rounded.control}"
+  demonstration-control:
+    height: 36px
+  exploration-composer:
+    backgroundColor: "{colors.surface}"
+    rounded: "{rounded.composer}"
+---
+
+# ARTEMIS 设计与交互约定
+
+## Overview
+
+ARTEMIS 面向使用自然语言执行、探索和复跑测试的人。当前产品包含 Flash、Pro、探索三种入口，按需调用安卓设备和电脑插件，提供批次、精选任务、历史、报告及证据。用户应能看懂当前在做什么、何时需要回答、哪些结果已有证据。
+
+本文记录主程序的职责边界及 `apps/showcase_ui` 的现有视觉与交互规则，不提出新的主题改版。远程 SDK 的协议边界见[局部设计](packages/artemis-client/DESIGN.md)；独立插件仓库维护适配器设计和构建规则。阶段性进度见[开发快照](docs/development-handoff.md)，发行状态见[2.0.1 说明](docs/releases/2.0.1.md)。
+
+## System architecture
+
+| 层 | 职责与源码入口 |
+| --- | --- |
+| 网页、MCP、远程 SDK | 请求建模、交互与结果读取。`apps/showcase_ui`、`mcp_server`、`packages/artemis-client` 共同使用后台服务；MCP 和薄 SDK 不另启完整后台 |
+| HTTP 与业务服务 | `apps/admin_console` 管理配置、统一队列、批次、收藏、历史、人工复验和导出；`TaskQueueService` 及现有资源锁负责执行准入与互斥 |
+| 任务与模型 | `artemis/agents`、`artemis/runtime` 执行 Flash/Pro；`artemis/exploration` 管理交流、预算、示范与用例版本；`artemis/llm`、`artemis/services/llm.py` 统一 Responses 请求 |
+| 插件与证据 | `artemis/plugins` 管理安装、会话、版本固定、来源、截图、录像和附件；插件执行工具动作，宿主管理任务生命周期和最终验收 |
+| 配置与交付 | `artemis/config`、两份 `artemis.jsonc` 和环境配置提供默认值；`scripts` 负责启动、镜像、独立升级和源码打包 |
+
+请求重试、业务执行与最终验收分别处理。提交使用稳定标识；创建批次或探索草稿不自动执行，查询不产生设备动作。执行准入和恢复遵守现有资源锁，任务固定所用插件版本，不能通过另建队列绕开互斥。
+
+最终验收有独立总预算，按请求超时、重试和剩余任务时间协调，保留取证结果；完整有效的结构化结论才更新正式裁定。`completed`、工具成功、探索通过、用例提炼和复跑通过各自表达自己的状态。
+
+示范来源保存在元数据中，贯穿分析、提问、规划、执行、验收和用例生成。补充网址不会隐式把电脑示范改到安卓；用户明确改端或提出跨端步骤时按其要求执行。证据必须对应真实来源、窗口、采集时间与操作。
+
+录像使用用户选择的范围。人工示范的窗口模式跟随前台窗口，仅保留一路，无有效窗口时保留缺口；普通录屏在只读验收前停止。恢复操作再追加片段，时间映射以实际录像为准，不借用另一来源或时间的画面。
+
+收藏保存独立模板与执行配置，用例复跑固定版本；删除原任务不破坏已保存模板。历史与导出的有效结论合并人工复验，不能用新裁定改写旧证据。
+
+普通任务新安装默认 `gpt-6-luna / high`，探索单独使用 `gpt-6.1-sol / high`，保留已有配置。上下文参照上限 300k、默认历史压缩预算 80k，均不表示整条任务累计 token 限额。优化稳定前缀时保留关键文件回执、路径、校验和来源；费用收益依实际网关用量及计价判断。
+
+## Colors
+
+整体使用浅灰画布、白色内容面和深色正文。首页主要操作使用蓝色，探索发送及相关主操作保留现有靛色；绿色、琥珀色、红色用于结果与需要处理的状态。新控件复用所在区域的语义色。
+
+| 用途 | 值与来源 |
+| --- | --- |
+| 画布 / 正文 | `#f8f9fa` / `#1e293b`，来自[全局样式](apps/showcase_ui/src/styles.scss) |
+| 首页操作 / 悬停 | `#1a73e8` / `#1557b0`，来自[首页样式](apps/showcase_ui/src/app/pages/home/home.component.scss) |
+| 探索操作 / 浅色状态面 | `#4f46e5` / `#eef2ff`，来自[探索面板](apps/showcase_ui/src/app/components/exploration-panel/exploration-panel.component.scss) |
+| 正文层级 / 边框 | 强调 `#0f172a`、次级 `#475569`、辅助 `#64748b`；常规边框 `#e2e8f0` |
+| 成功、等待、错误 | 深色状态文字配对应浅色面；同时给出文字、图标或说明，结论不只依靠颜色 |
+
+加载、空状态、断开、未验证与失败使用明确文案。执行完成可以有完成标识，测试裁定仍需独立展示；不能让绿色完成态替代测试通过结论。
+
+## Typography
+
+沿用全局字体栈：`Google Sans`、`Google Sans Text`、`Inter`、系统无衬线字体；中文使用系统可用字形。代码、路径及请求诊断使用现有等宽样式，保留换行与复制能力。
+
+模型控件正文为 13px，探索输入正文为 14px，用例标题为 18px；辅助信息依现有组件使用 11–12px。保留字号、字重、留白形成的层次，不通过大段加粗堆叠状态。Markdown、表格、长链接和中文长标题应换行，不撑破容器。
+
+图标沿用 Material Symbols Outlined，常规全局尺寸 20px；组件局部按原有尺寸调整。图标与相邻文字垂直居中，独立图标按钮提供可访问名称。
+
+## Layout
+
+- 首页负责创建任务与访问精选模板；工作区主区域展示当前任务的交流、执行步骤和结果，底部输入框始终服务当前任务。
+- 编辑、模型设置和复跑参数按需展开，默认保持任务时间线可读。精选卡片在收起状态下等高，长标题、不同状态和展开内容不会遮挡操作入口。
+- 通常以 4、6、8、12、16、24px 组织内部间距；沿用组件现有网格和断点。子容器允许收缩，长内容使用 `min-width: 0` 及合适换行。
+- 探索面板在 600px 以下收紧外边距并允许操作换行；模型选择控件宽度跟随容器。窄屏仍可辨认状态、选择来源和发送指导。
+- 列表后台刷新保留滚动、选择和折叠状态；首次加载与静默更新分开，避免周期性加载提示挤动内容。实时连接复用现有页面可见性管理。
+
+## Elevation
+
+卡片使用细边框和克制的阴影区分区域。首页现有玻璃效果、渐变和按钮悬停保留在原有区域，正文、报告、表单和选项保持清晰背景及文字对比。
+
+探索输入区沿用 `0 6px 24px #0f172a08` 的阴影；焦点反馈沿用所在控件的轮廓和浅色描边。全局 CDK overlay 层级为 1200，新增弹层先使用既有 overlay 机制，不任意叠加更大的层级。
+
+## Shapes
+
+模型输入与选择框圆角 10px；紧凑控件常用 8px，内容卡片按组件使用 12–16px，探索输入区为 22px。保留现有徽标、标签与胶囊形控件的局部形状；这些值不要求把全部历史控件改成同一圆角。
+
+## Components
+
+| 组件或状态 | 必须维持的交互与尺寸 | 实现依据 |
+| --- | --- | --- |
+| 模型选择 | 预设选择加自定义输入；主模型、兜底和精选任务复用同一组件。最小高度 40px、内边距 10px 12px、圆角 10px，保留焦点与禁用态 | [model-selector](apps/showcase_ui/src/app/components/model-selector/model-selector.component.ts) |
+| 人工示范控件 | 来源标签与控件水平对齐，选择框及开始按钮均为 36px 高、垂直居中；窄屏可换行。准备、录制、分析分别显示 | [exploration-panel](apps/showcase_ui/src/app/components/exploration-panel/exploration-panel.component.scss) |
+| 探索建议 | 推荐内容为用户口吻的可编辑回复；点击只填草稿，用户发送后才继续任务 | [探索组件](apps/showcase_ui/src/app/components/exploration-panel/exploration-panel.component.ts) |
+| 历史标题与星标 | 星标只占标题，第二行状态与编号正常左对齐；状态徽标沿用 20px 高及统一内边距 | [task-history](apps/showcase_ui/src/app/components/task-history/task-history.component.ts) |
+| 验收状态 | 展示准备证据、核对、生成与校验等阶段及可用的耗时、预算、重试诊断；未完成的流式文字不变为正式通过标签 | [agent-stream](apps/showcase_ui/src/app/components/agent-stream/agent-stream.component.html) |
+| 截图、录像与文件 | 显示真实来源和缺口；按时间定位并保留跳转前的播放/暂停状态，文件成功以真实保存回执为准 | [工作区说明](docs/automation-workspace.md) |
+| 保存和错误反馈 | 配置加载、保存、空值、失败与密钥已配置状态有明确提示；空密钥保留已有值，连接测试与保存分开 | [模型配置](apps/showcase_ui/src/app/components/model-settings/model-settings.component.ts) |
+
+这些规则对应已接受的现有行为。改动后结合实际层叠样式、渲染尺寸和相关状态检查，不能只核对一处声明。
+
+## Do's and Don'ts
+
+- 复用现有组件、API 类型、中文文案和状态语义。新增状态先说明它与执行、验收或人工操作的关系。
+- 保持实际来源与证据一致，缺失时说明原因；不使用手机画面替代电脑操作，也不把工具成功当作业务通过。
+- 保留可见焦点、禁用反馈和减少动画偏好；录制红点表示实际录制，等待或执行中使用各自文案。
+- 源码已更新而页面仍旧时，先核对服务目录、构建与缓存；不通过反复增加样式覆盖掩盖旧构建问题。
+- 视觉变化验证代表性宽度、长标题、加载、等待回答、错误与展开状态；文档格式检查不等于渲染验证。开发检查按[前端指南](apps/showcase_ui/AGENTS.md)执行。

README.md

@@ -2,17 +2,22 @@
 
 基于 [ARTEMIS](https://github.com/google/artemis) 扩展的 Android AI 自动化测试平台。用自然语言描述任务，即可通过简体中文网页或 MCP 驱动真机、模拟器执行操作并收集结果。模型统一使用 **OpenAI Responses API**。
 
+**2.0.1 已发布**：修复 Python 3.12 全新安装时网页服务无法启动的问题，并统一不同 Python 版本的 MCP 工具描述。修复与升级步骤见 [2.0.1 发行说明](docs/releases/2.0.1.md)；探索、插件和精选任务等功能见 [2.0.0 发行说明](docs/releases/2.0.0.md)。
+
 ## 主要功能
 
 - **自然语言任务**：Flash 适合明确的操作流程，Pro 适合需要计划、分支探索和验证的任务。
+- **交互探索**：边执行边交流，也可录制人工示范；探索成功后生成一个可编辑、复跑的用例。默认使用独立的 `gpt-6.1-sol / high` 配置，超时暂停并保留进度。见 [探索模式](docs/exploration-mode.md)。
+- **插件扩展**：在插件市场安装并启用电脑控制等能力，AI 在安卓任务中按需使用；手机、电脑的步骤与证据统一留档。见 [插件与跨端任务](docs/plugins.md)。
 - **可保存的模型配置**：支持自定义接口、模型、思考强度和超时重试，以及视频分析与录屏开关。
-- **批量测试**：下载 Excel 模板、校验导入、顺序执行，支持暂停、继续和复制失败项。
+- **批量测试**：按新版 Excel 模板导入提示词，每项可独立指定模型、模式和参数；任务可暂停、补充指导并继续。
-- **报告与历史**：导出 Word、Excel 和批次 ZIP；历史按日期分组，支持搜索和筛选。
+- **精选任务**：收藏单任务、批次或探索用例，在首页一键复跑，默认沿用已保存的配置。
+- **报告与历史**：导出 Word、Excel、步骤截图、录屏和任务资料 ZIP；历史按日期分组，支持搜索和筛选。
 - **AI 接入**：随源码提供 MCP 与 Skill，让 AI 完成设备诊断、任务执行、历史分析和报告交付。
 
 ## 安装与启动
 
-准备一台开启 USB 调试并完成授权的 Android 手机，或已启动的 Android 模拟器。首次安装需要联网，建议使用 Git 克隆以支持后续更新。
+准备一台开启 USB 调试并完成授权的 Android 手机，或已启动的 Android 模拟器。首次安装需要联网。可使用 Git 克隆或下载 1.1.1 及更新版本的发布源码包；新版 ZIP 解压后也保留 Git 更新能力。
 
 ### Windows
 
@@ -24,54 +29,63 @@ cd Artemis-Eapil
 .\start.bat
 ```
 
-国内网络使用 `.\start-cn.bat`。启动脚本会检测并安装缺少的 uv、Python 和所需工具，创建或复用 `.venv`，同步依赖并构建网页。启动后访问 **http://127.0.0.1:8000**。
+国内网络使用 `.\start-cn.bat`，支持 `-Help` 查看中文选项。启动脚本会检测并安装缺少的 uv、Python 和所需工具，创建或复用 `.venv`，同步依赖并构建网页。启动后访问 **http://127.0.0.1:8000**。
 
 常用选项：
 
 ```powershell
-.\start-cn.bat -SetupOnly   # 仅准备 Python 环境并检查已有 MCP 配置
+.\start-cn.bat -SetupOnly   # 准备环境并安装/更新 Codex MCP
 .\start.bat -Port 8001      # 指定端口
-.\start.bat -SkipUpdate     # 跳过 Git 源码更新检查
+.\start.bat -SkipUpdate     # 跳过源码更新提示
+.\start.bat -SkipMcpSync    # MCP 配置异常时，仅启动网页
 ```
 
-镜像与下载超时在 [config/downloads.json](config/downloads.json) 中设置。scrcpy 默认从 GitHub 下载，无法访问时可配置可信的 scrcpy_url 或预先安装。
+镜像与下载超时在 [config/downloads.json](config/downloads.json) 中设置。scrcpy 国内下载源失败时自动尝试备用源，每个安装包均校验官方 SHA256；也可预先安装，或通过 `overrides.scrcpy_url` 和 `scrcpy_fallback_urls` 自定义 HTTPS 地址。
 
 ### macOS / Linux
 
-克隆仓库后在项目目录执行 `./start.sh`。国内镜像入口和启动时自动更新目前由 Windows 脚本提供。
+克隆仓库后在项目目录执行 `./start.sh`。国内镜像入口和独立升级脚本目前由 Windows 提供。
 
 ## 配置与使用
 
 1. 在首页“系统设置与运行准备”中选择设备，打开“模型配置”。
 2. 填写支持 Responses 的接口地址、API Key 和模型名称，测试连接后保存。接口地址通常是以 `/v1` 结尾的 API 根地址。
-3. 输入任务目标，选择 Flash 或 Pro 后执行；在工作区查看步骤、截图和结果。
+3. 输入任务目标，选择 Flash、Pro 或探索后执行；在工作区查看步骤、截图和结果。流程尚不明确时，选择探索并随时交流。
 4. 通过“报告与批量”导出报告或导入 Excel 用例；任务队列支持历史搜索。
 
-预设模型为 `gpt-5.6-luna`、`mimo-v2.6-flash`，也可填写自定义名称。默认模型是 `mimo-v2.6-flash`，思考强度默认 `high`（最高可选 `max`），单次请求超时 **300 秒**，失败后最多重试 **5 次**，间隔 **10 秒**；视频分析与录屏默认关闭。模型及思考参数是否可用取决于所配置网关；不支持时会明确报错。
+预设模型为 `gpt-6-luna`、`gpt-6.1-sol`、`gpt-6-astra`、`mimo-v2.6-pro`，主模型与兜底模型均可选择预设或填写自定义名称。新安装默认模型是 `gpt-6-luna`，思考强度默认 `high`（最高可选 `max`），单次请求超时 **300 秒**，失败后最多重试 **5 次**，间隔 **10 秒**。遇到 429 请求过多时改为每次等待 **60 秒**。连续 3 次没有有效工具调用时，本步骤改用 `gpt-6-luna`（high），兜底模型最多再尝试 3 次；下一步恢复首选模型。视频分析与录屏默认关闭。模型及思考参数是否可用取决于所配置网关；不支持时会明确报错。
 
-也可在项目 `.env` 中设置 `OPENAI_API_KEY`、`OPENAI_BASE_URL` 和 `OPENAI_MODEL`，其他参数见 [.env.example](.env.example)。直接编辑环境文件后需重启相应服务。不要提交自己的密钥。
+也可在项目 `.env` 中设置 `OPENAI_API_KEY`、`OPENAI_BASE_URL` 和 `OPENAI_MODEL`，其他参数见 [.env.example](.env.example)。新安装默认接口地址为 `https://api-n-cd.playones.com/v1`；已保存的地址不变。直接编辑环境文件后需重启相应服务。不要提交自己的密钥。
+
+“功能开关”默认将任务截图长边限制为 2048 像素；关闭后保留设备原始分辨率。可在 [config/artemis.jsonc](config/artemis.jsonc) 的 `agent.screenshot` 中覆盖 `enabled` 和 `max_long_edge`，重启后对新任务生效。
 
 报告会区分“执行完成”和“验证通过”；没有验证证据的结果标为“未验证”。任务步骤与截图可查看，设备动作回放不提供。
 
 ## 让 AI 使用
 
-环境与模型配置完成后，例如为 Codex 接入 MCP：
+目前仅支持为 Codex 配置 MCP。Windows 启动脚本检测 Codex 配置目录并安装或更新 ARTEMIS；也可手动安装：
 
 ```sh
 uv run artemis mcp --install codex
 ```
 
-在客户端重新加载 MCP，并按 [AI 接入指南](docs/ai-automation.md) 安装仓库中的 [artemis-automation Skill](skills/artemis-automation/SKILL.md)。Skill 不会随启动脚本自动安装到用户全局目录。其他客户端可通过 `uv run artemis mcp --help` 查看接入方式。
+使用工具前，由用户先启动 ARTEMIS 网页主程序。MCP 仅连接这份主程序，不会为每个 AI 会话启动后台；主程序离线不阻断 Codex 会话，工具会提示先启动网页。
+
+启动 ARTEMIS 时会自动安装或覆盖更新官方 [artemis-use Skill](skills/artemis-use/SKILL.md)，包含 Flash／Pro／探索模式、任务诊断复盘和插件开发指南。MCP 配置变化后在 Codex 中重新加载 MCP；Skill 未显示时重启 Codex。详细行为见 [AI 接入指南](docs/ai-automation.md)。可用 `uv run artemis mcp --generate-config codex` 查看配置模板。
 
 示例指令：
 
-> 使用 artemis-automation，在我指定的模拟器上验证系统设置首页可以打开，不修改设置，完成后输出 Word 和 Excel 报告。
+> 使用 artemis-use，在我指定的模拟器上验证系统设置首页可以打开，不修改设置，完成后输出 Word 和 Excel 报告。
 
 ## 更新与排错
 
-Windows 用户再次运行一键脚本时，会检查 Git 上游、项目依赖、前端构建和已有 MCP 配置，仅对源码做快进更新。本地修改、分支分叉或网络认证失败时保留本地版本；有任务执行或排队时停止更新启动。更新后需在 AI 客户端重新加载 MCP。
+启动脚本只检查并提示源码更新，不自动拉取。升级运行 **`.\update.bat`**，国内网络运行 **`.\update-cn.bat`**；支持 `-Help`。升级会终止当前项目的 ARTEMIS 服务、MCP 和任务进程，然后快进更新源码，同步依赖、网页构建及 Codex MCP。升级程序先复制到项目外运行，避免更新正在执行的脚本；其他项目、Codex 本身、模拟器和共享 ADB 服务不会被终止。完成后重新启动 ARTEMIS，并在 Codex 中重载 MCP。
 
-ZIP 安装可继续运行；需要自动拉取源码时，请改用 Git 克隆安装。遇到启动或设备问题，在项目目录运行 `uv run artemis doctor`。
+本地修改、分支分叉或网络认证失败会停止升级，保留现有文件；`.env`、密钥和历史资料不会清空。依赖同步失败会明确报错；Codex MCP 或 Skill 同步失败会保留可用版本并告警，网页继续启动。可用 `-SkipMcpSync` 跳过该次 Codex 集成同步。
+
+1.1.1 及更新版本的发布 ZIP 含浅层 Git 工作树；包含独立升级入口的新版源码包可直接用上述脚本升级。1.1.0 及更早 ZIP 不含 Git 信息，需改用 Git 克隆或新版源码包。遇到启动或设备问题，在项目目录运行 `uv run artemis doctor`。
+
+电脑控制插件由独立市场分发，升级 ARTEMIS 后还需在「插件」页面安装或更新并启用。当前 Cua 插件适用于 Windows x64，保留预览标识和实测范围；需要前台输入时会占用用户桌面。配置和开发约定见 [插件与跨端任务](docs/plugins.md)。
 
 更多说明：[批量任务与报告](docs/automation-workspace.md) · [AI 接入](docs/ai-automation.md) · [开发协作指南](AGENTS.md)
 

apps/admin_console/__init__.py

@@ -14,6 +14,14 @@
 
 """Artemis Admin & Trace Console Package."""
 
+__all__ = ["app"]
+
+
+def __getattr__(name):
+    # Workers use shared persistence without importing the HTTP routers or
+    # constructing the web application. Keep the public app import compatible.
+    if name == "app":
         from apps.admin_console.server import app
 
-__all__ = ["app"]
+        return app
+    raise AttributeError(name)

apps/admin_console/core/state.py

@@ -136,7 +136,7 @@ class ServerState:
         return {
             str(run.get("lock_key") or run["device_id"])
             for run in self.active_runs.values()
-            if run.get("device_id")
+            if run.get("device_id") and not run.get("parked")
         }
 
     @property

apps/admin_console/database/repositories/session_repository.py

@@ -23,9 +23,6 @@ from artemis.runtime import trace_store
 
 logger = logging.getLogger(__name__)
 
-try:
-    from admin_console.database.connection import db_session
-except ImportError:
 from apps.admin_console.database.connection import db_session
 
 
@@ -323,6 +320,7 @@ class SessionRepository:
                 return 0
 
         llm_calls = prompt = completion = total = cached = 0
+        by_node_model: dict[tuple[str, str], dict] = {}
         operator_context: int | None = None
         operator_context_at: float | None = None
         with db_session(self.db_path) as conn:
@@ -351,6 +349,27 @@ class SessionRepository:
                 total += t
                 cached += _int(payload.get("cached_tokens"))
                 node = str(payload.get("node") or "").strip().lower()
+                model_name = str(payload.get("model") or "unknown")
+                group = by_node_model.setdefault(
+                    (node, model_name),
+                    {
+                        "node": node,
+                        "model": model_name,
+                        "llm_calls": 0,
+                        "prompt_tokens": 0,
+                        "completion_tokens": 0,
+                        "cached_tokens": 0,
+                    },
+                )
+                group["llm_calls"] += 1
+                group["prompt_tokens"] += p
+                group["completion_tokens"] += c
+                group["cached_tokens"] += _int(payload.get("cached_tokens"))
+                group["cache_hit_rate"] = (
+                    group["cached_tokens"] / group["prompt_tokens"]
+                    if group["prompt_tokens"]
+                    else None
+                )
                 if node in OPERATOR_NODE_NAMES:
                     operator_context = _int(payload.get("context_base_tokens")) or p
                     operator_context_at = row["timestamp"]
@@ -378,6 +397,7 @@ class SessionRepository:
             "completion_tokens": completion,
             "total_tokens": total,
             "cached_tokens": cached,
+            "by_node_model": list(by_node_model.values()),
             "operator_context_tokens": operator_context,
             "operator_context_window_tokens": OPERATOR_CONTEXT_WINDOW_TOKENS,
             "operator_context_updated_at": operator_context_at,
@@ -422,7 +442,7 @@ class SessionRepository:
             return count
 
     def reconcile_orphaned_sessions(self) -> int:
-        """Mark running sessions with no live worker process as failed.
+        """Mark running sessions with no live worker process as interrupted.
 
         This is safe to call both during startup and immediately after a
         forced server stop: live workers, including workers owned by another
@@ -442,14 +462,14 @@ class SessionRepository:
                     cursor.execute(
                         "UPDATE sessions SET status = ?, end_time = ? "
                         "WHERE session_id = ? AND status = ?",
-                        ("failed", now, row["session_id"], "running"),
+                        ("interrupted", now, row["session_id"], "running"),
                     )
                     count += cursor.rowcount
                     try:
                         if trace_store.read_status(str(row["session_id"])):
                             trace_store.update_trace_status(
                                 str(row["session_id"]),
-                                "failed",
+                                "interrupted",
                                 error="Process terminated prior to server startup.",
                             )
                     except OSError as exc:

apps/admin_console/database/repositories/step_repository.py

@@ -16,9 +16,6 @@ import json
 import re
 from typing import Any
 
-try:
-    from admin_console.database.connection import db_session
-except ImportError:
 from apps.admin_console.database.connection import db_session
 
 
@@ -183,6 +180,13 @@ class StepRepository:
         if not isinstance(payload_obj, dict):
             return None
 
+        action = payload_obj.get("action")
+        if payload_obj.get("plugin_id") or (isinstance(action, dict) and action.get("plugin_id")):
+            # Plugin payloads contain already-redacted evidence and explicit image
+            # references. The legacy compact-tool formatter would discard their
+            # provenance and make a desktop image inherit the phone's step source.
+            return self._clean_value(payload_obj)
+
         if trace_type == "llm_call":
             allowed_keys = (
                 "error",

apps/admin_console/database/repositories/trace_repository.py

@@ -15,9 +15,6 @@
 import json
 from typing import Any
 
-try:
-    from admin_console.database.connection import db_session
-except ImportError:
 from apps.admin_console.database.connection import db_session
 
 

apps/admin_console/routers/automation.py

@@ -8,9 +8,13 @@ from fastapi import APIRouter, HTTPException, Query, UploadFile
 from fastapi.responses import FileResponse, Response
 from pydantic import BaseModel, Field
 
+from apps.admin_console.services.exploration_listing import EXPLORATION_RUNS_CTE
 from apps.admin_console.services.automation_store import identifier, store
 from apps.admin_console.services import automation_service as service
+from apps.admin_console.services import batch_archive as archive_service
 from apps.admin_console.services import report_export, task_import
+from artemis.config.settings import settings
+from apps.admin_console.services.case_review import read_reviewed_result
 
 router = APIRouter(prefix="/api/automation", tags=["automation"])
 
@@ -21,15 +25,22 @@ class ReportPreferences(BaseModel):
     )
 
 
+class CaseReviewRequest(BaseModel):
+    model_config = {"extra": "forbid"}
+    status: Literal["passed", "failed"]
+    reason: str = Field(default="", max_length=2000)
+
+
 class BatchRequest(ReportPreferences):
     request_id: str | None = Field(default=None, pattern=r"^[A-Za-z0-9_-]{1,100}$")
     name: str = Field(min_length=1, max_length=200)
     cases: list[task_import.TaskCase] = Field(min_length=1, max_length=task_import.MAX_ROWS)
-    device_serial: str = Field(min_length=1, max_length=200)
+    device_serial: str | None = Field(default=None, min_length=1, max_length=200)
     profile: Literal["flash", "pro"] = "flash"
     stop_on_failure: bool = False
+    repeat_count: int = Field(default=1, ge=1, strict=True)
     timeout_seconds: int = Field(
-        default_factory=lambda: int(os.getenv("ARTEMIS_BATCH_TASK_TIMEOUT", "1800")),
+        default_factory=lambda: settings.ARTEMIS_TASK_TIMEOUT_SECONDS,
         ge=30,
         le=86400,
     )
@@ -49,7 +60,8 @@ def require(kind, key):
 def preferences():
     return {
         **store.preferences(),
-        "batch_timeout_seconds": int(os.getenv("ARTEMIS_BATCH_TASK_TIMEOUT", "1800")),
+        "batch_timeout_seconds": settings.ARTEMIS_TASK_TIMEOUT_SECONDS,
+        "batch_max_executions": settings.ARTEMIS_BATCH_MAX_EXECUTIONS,
     }
 
 
@@ -118,7 +130,29 @@ def batch_by_request(request_id: str):
 
 @router.get("/batches/{batch_id}")
 def batch_detail(batch_id: str):
-    return {**require("batch", batch_id), "jobs": store.jobs(batch_id)}
+    jobs = store.jobs(batch_id)
+    for job in jobs:
+        job["test_result"] = read_reviewed_result(
+            job["session_id"], job["status"], job.get("error") or "", storage=store
+        )
+    return {**require("batch", batch_id), "jobs": jobs}
+
+
+@router.delete("/batches/{batch_id}")
+def delete_batch(batch_id: str):
+    try:
+        key = identifier(batch_id)
+    except ValueError as exc:
+        raise HTTPException(400, str(exc)) from exc
+    try:
+        if archive_service.status(key, require("batch", key))["status"] == "preparing":
+            raise HTTPException(409, "报告正在打包，请完成后再删除批次")
+        if not service.delete_batch(key):
+            raise HTTPException(404, "批次不存在")
+    except ValueError as exc:
+        raise HTTPException(409, str(exc)) from exc
+    archive_service.forget(key)
+    return {"batch_id": key, "status": "deleted"}
 
 
 @router.post("/batches/{batch_id}/{action}")
@@ -130,9 +164,25 @@ def batch_action(batch_id: str, action: Literal["pause", "resume", "retry_failed
         raise HTTPException(409, str(exc)) from exc
 
 
+@router.post("/batches/{batch_id}/archive/prepare")
+def prepare_batch_archive(batch_id: str):
+    batch = require("batch", batch_id)
+    return archive_service.prepare(batch_id, batch)
+
+
+@router.get("/batches/{batch_id}/archive/status")
+def batch_archive_status(batch_id: str):
+    batch = require("batch", batch_id)
+    return archive_service.status(batch_id, batch)
+
+
 @router.get("/batches/{batch_id}/archive")
 async def batch_archive(batch_id: str):
-    require("batch", batch_id)
+    batch = require("batch", batch_id)
+    if archive_service.status(batch_id, batch)["status"] == "preparing":
+        raise HTTPException(409, "报告正在打包，请等待准备完成")
+    path = archive_service.archive_path(batch_id)
+    if not path.is_file() or archive_service.status(batch_id, batch)["status"] != "ready":
         path = await asyncio.to_thread(report_export.export_batch, batch_id)
     return FileResponse(path, filename=f"批次报告-{batch_id[:8]}.zip")
 
@@ -146,14 +196,57 @@ def report(session_id: str):
 
 
 @router.get("/reports/{session_id}/{fmt}")
-async def download_report(session_id: str, fmt: Literal["docx", "xlsx"]):
+async def download_report(
+    session_id: str, fmt: Literal["docx", "xlsx", "screenshots", "video", "all"]
+):
     try:
+        if fmt in {"screenshots", "video", "all"}:
+            path, label = await asyncio.to_thread(report_export.session_download, session_id, fmt)
+            return FileResponse(path, filename=f"{label}-{fmt}.zip")
         paths = await asyncio.to_thread(report_export.export_session, session_id, [fmt])
         return FileResponse(paths[0], filename=f"任务报告-{session_id[:8]}.{fmt}")
     except ValueError as exc:
         raise HTTPException(404, str(exc)) from exc
 
 
+@router.put("/reports/{session_id}/review")
+def review_case(session_id: str, value: CaseReviewRequest):
+    report(session_id)
+    job = store.job(session_id) or {}
+    batch_id = job.get("batch_id")
+    if (
+        batch_id
+        and archive_service.status(batch_id, require("batch", batch_id))["status"] == "preparing"
+    ):
+        raise HTTPException(409, "批次报告正在打包，请完成后再保存人工复验")
+    with report_export.export_lock(
+        *(["batch:" + batch_id] if batch_id else []), "session:" + session_id
+    ):
+        current = report(session_id)
+        if current["status"] not in {
+            "completed",
+            "success",
+            "failed",
+            "cancelled",
+            "interrupted",
+            "skipped",
+        }:
+            raise HTTPException(409, "请在任务结束后进行人工复验")
+        try:
+            store.save_case_review(session_id, value.status, value.reason)
+        except ValueError as exc:
+            raise HTTPException(409, str(exc)) from exc
+        if batch_id:
+            archive_service.forget(batch_id)
+            # A finishing worker can publish report_ready just after export;
+            # removing only the generated ZIP also prevents that stale cache hit.
+            try:
+                archive_service.archive_path(batch_id).unlink(missing_ok=True)
+            except OSError:
+                pass  # An open download may hold the file; report_ready is already invalidated.
+        return report(session_id)
+
+
 @router.get("/history")
 def history(
     q: str = Query(default="", max_length=200),
@@ -162,26 +255,67 @@ def history(
     end: float | None = None,
     page: int = Query(default=1, ge=1),
     size: int = Query(default=30, ge=1, le=100),
+    batch_id: str = Query(default="", max_length=100),
 ):
     # SQL filtering/pagination searches the entire persisted history, not a UI cache.
-    query = """WITH history AS (
+    query = (
-      SELECT s.session_id, s.initial_goal, s.start_time, s.end_time,
+        "WITH "
+        + EXPLORATION_RUNS_CTE
+        + """, history_base AS (
+      SELECT s.session_id, s.initial_goal, s.start_time,
+        COALESCE(json_extract(j.payload,'$.ended_at'),s.end_time) end_time,
         COALESCE(j.status,s.status) status, COALESCE(json_extract(j.payload,'$.case.name'),s.initial_goal) name,
-        COALESCE(json_extract(j.payload,'$.case.tags'),'') tags
+        COALESCE(json_extract(j.payload,'$.case.tags'),'') tags,
+        COALESCE(json_extract(j.payload,'$.case.case_id'),'') case_id,
+        COALESCE(json_extract(j.payload,'$.case.category'),'') category,
+        COALESCE(json_extract(j.payload,'$.case.subcategory'),'') subcategory,
+        COALESCE(json_extract(j.payload,'$.case.preconditions'),'') preconditions,
+        COALESCE(json_extract(j.payload,'$.error'),'') error, j.batch_id,
+        COALESCE(CASE WHEN json_valid(s.device_info) THEN json_extract(s.device_info,'$.profile') END,
+          json_extract(j.payload,'$.task_config.profile'),json_extract(j.payload,'$.profile')) profile
       FROM sessions s LEFT JOIN automation_jobs j ON s.session_id=j.session_id
       UNION ALL
       SELECT j.session_id,json_extract(j.payload,'$.case.goal'),
         COALESCE(json_extract(j.payload,'$.started_at'),json_extract(j.payload,'$.created_at')),
         json_extract(j.payload,'$.ended_at'),j.status,json_extract(j.payload,'$.case.name'),
-        COALESCE(json_extract(j.payload,'$.case.tags'),'')
+        COALESCE(json_extract(j.payload,'$.case.tags'),''),
+        COALESCE(json_extract(j.payload,'$.case.case_id'),''),
+        COALESCE(json_extract(j.payload,'$.case.category'),''),
+        COALESCE(json_extract(j.payload,'$.case.subcategory'),''),
+        COALESCE(json_extract(j.payload,'$.case.preconditions'),''),
+        COALESCE(json_extract(j.payload,'$.error'),''), j.batch_id,
+        COALESCE(json_extract(j.payload,'$.task_config.profile'),json_extract(j.payload,'$.profile'))
       FROM automation_jobs j WHERE NOT EXISTS (SELECT 1 FROM sessions s WHERE s.session_id=j.session_id)
+    ), history AS (
+      SELECT h.*, e.exploration_id, e.exploration_kind, e.case_version,
+        e.exploration_status, e.exploration_latest_run_id, e.exploration_run_count,
+        e.exploration_case_name, e.exploration_profile
+      FROM history_base h LEFT JOIN exploration_runs e ON e.run_id=h.session_id
+      WHERE COALESCE(e.exploration_kind,'') != 'exploration'
+        OR (e.run_id=e.exploration_latest_run_id AND
+          COALESCE(e.exploration_status,'completed') NOT IN
+            ('draft','reserved','queued','running','paused','waiting_user','demonstrating','analyzing_demo','finalizing','interrupted'))
     ) """
+    )
     where = " WHERE status IN ('completed','success','failed','cancelled','interrupted','skipped')"
     args = []
+    if batch_id == "__exploration__":
+        where += " AND exploration_kind = 'exploration'"
+    elif batch_id == "__batch__":
+        where += " AND batch_id IS NOT NULL"
+    elif batch_id == "__single__":
+        where += " AND batch_id IS NULL AND COALESCE(exploration_kind,'') != 'exploration'"
+    elif batch_id:
+        try:
+            identifier(batch_id)
+        except ValueError as exc:
+            raise HTTPException(400, str(exc)) from exc
+        where += " AND batch_id = ?"
+        args.append(batch_id)
     if q:
-        where += " AND (name LIKE ? ESCAPE '\\' OR initial_goal LIKE ? ESCAPE '\\' OR session_id LIKE ? ESCAPE '\\' OR tags LIKE ? ESCAPE '\\')"
+        where += " AND (COALESCE(exploration_case_name,name) LIKE ? ESCAPE '\\' OR initial_goal LIKE ? ESCAPE '\\' OR session_id LIKE ? ESCAPE '\\' OR tags LIKE ? ESCAPE '\\' OR case_id LIKE ? ESCAPE '\\' OR category LIKE ? ESCAPE '\\' OR subcategory LIKE ? ESCAPE '\\' OR preconditions LIKE ? ESCAPE '\\')"
         escaped = "%" + q.replace("\\", "\\\\").replace("%", "\\%").replace("_", "\\_") + "%"
-        args.extend([escaped] * 4)
+        args.extend([escaped] * 8)
     for condition, value in [
         ("status = ?", status or None),
         ("start_time >= ?", start),
@@ -199,4 +333,23 @@ def history(
             + " ORDER BY start_time DESC,session_id LIMIT ? OFFSET ?",
             [*args, size, (page - 1) * size],
         ).fetchall()
-    return {"items": [dict(r) for r in rows], "total": total, "page": page, "size": size}
+    items = [dict(row) for row in rows]
+    jobs = store.jobs_for_sessions([item["session_id"] for item in items])
+    for item in items:
+        item["profile"] = (
+            "explore"
+            if item.get("exploration_kind") == "exploration"
+            else item.get("exploration_profile") or item.get("profile")
+        )
+        item["test_result"] = read_reviewed_result(
+            item["session_id"], item["status"], item.get("error") or "", storage=store
+        )
+        if job := jobs.get(item["session_id"]):
+            item["name"] = task_import.execution_title(job)
+            item["iteration"] = job.get("iteration", 1)
+            item["repeat_count"] = job.get("repeat_count", 1)
+        elif item["case_id"]:
+            item["name"] = task_import.case_title(item)
+        if item.get("exploration_kind") == "validation" and item.get("exploration_case_name"):
+            item["name"] = item["exploration_case_name"]
+    return {"items": items, "total": total, "page": page, "size": size}

apps/admin_console/routers/exploration.py

@@ -0,0 +1,180 @@
+"""Shared exploration operations; no separate Web execution path."""
+
+from contextlib import contextmanager
+
+from fastapi import APIRouter, HTTPException, Query
+
+from artemis.config.exploration import ExplorationConfig
+from artemis.exploration.models import (
+    CreateExploration,
+    ExplorationMessage,
+    ExplorationRequest,
+    SaveCaseRequest,
+    ValidateCaseRequest,
+    DemonstrationRequest,
+    ExtendDemonstration,
+    GenerateCaseRequest,
+    CaseBatchRequest,
+)
+from artemis.exploration.store import ExplorationConflict
+from apps.admin_console.services.exploration_service import exploration_service as service
+
+
+router = APIRouter(prefix="/api/explorations", tags=["exploration"])
+
+
+@router.get("/{exploration_id}/deletion-preview")
+def preview_deletion(exploration_id: str):
+    from apps.admin_console.services.task_deletion import exploration_deletion_preview
+
+    with domain_errors():
+        return exploration_deletion_preview(exploration_id)
+
+
+@router.delete("/{exploration_id}")
+def delete_conversation(exploration_id: str):
+    from apps.admin_console.services.exploration_deletion import delete_exploration
+
+    with domain_errors():
+        deleted = delete_exploration(service.repository, exploration_id)
+        return {"status": "success", "deleted_session_ids": deleted}
+
+
+@contextmanager
+def domain_errors():
+    try:
+        yield
+    except KeyError as exc:
+        raise HTTPException(404, str(exc).strip("'")) from exc
+    except ExplorationConflict as exc:
+        raise HTTPException(409, str(exc)) from exc
+    except ValueError as exc:
+        # Pydantic may echo sensitive input. Invalid nested fields get a plain error.
+        from pydantic import ValidationError
+
+        raise HTTPException(
+            400, "探索参数不合法" if isinstance(exc, ValidationError) else str(exc)
+        ) from exc
+
+
+@router.get("/settings")
+def settings():
+    return service.settings()
+
+
+@router.put("/settings")
+def save_settings(value: ExplorationConfig):
+    return service.save_settings(value)
+
+
+@router.get("")
+def list_explorations(limit: int = Query(100, ge=1, le=500)):
+    return service.repository.list(limit)
+
+
+@router.post("")
+def create(value: CreateExploration, source: str = Query("web", pattern="^(web|mcp)$")):
+    with domain_errors():
+        return service.create(value, source=source)
+
+
+@router.get("/{session_id}")
+def detail(session_id: str, after: int = Query(0, ge=0)):
+    with domain_errors():
+        return service.detail(session_id, after=after)
+
+
+@router.post("/{session_id}/start")
+async def start(
+    session_id: str, value: ExplorationRequest, source: str = Query("web", pattern="^(web|mcp)$")
+):
+    with domain_errors():
+        return await service.start(session_id, value.request_id, source=source)
+
+
+@router.post("/{session_id}/messages")
+async def message(
+    session_id: str, value: ExplorationMessage, source: str = Query("web", pattern="^(web|mcp)$")
+):
+    with domain_errors():
+        result = service.message(session_id, value.text, value.request_id, source=source)
+        if not result.get("run_id"):
+            from apps.admin_console.services.automation_service import spawn
+            from artemis.exploration.editorial import discuss
+
+            spawn(discuss(service.repository, session_id, result))
+        return result
+
+
+@router.post("/{session_id}/pause")
+def pause(session_id: str):
+    with domain_errors():
+        return service.pause(session_id)
+
+
+@router.post("/{session_id}/stop")
+async def stop(session_id: str):
+    with domain_errors():
+        return service.stop(session_id)
+
+
+@router.post("/{session_id}/continue")
+async def resume(
+    session_id: str, value: ExplorationRequest, source: str = Query("web", pattern="^(web|mcp)$")
+):
+    with domain_errors():
+        return await service.resume(session_id, value.request_id, source=source)
+
+
+@router.get("/{session_id}/case")
+def case(session_id: str, version: int | None = Query(None, ge=1)):
+    with domain_errors():
+        return service.repository.case(session_id, version)
+
+
+@router.post("/{session_id}/case")
+def save_case(session_id: str, value: SaveCaseRequest):
+    with domain_errors():
+        return service.save_case(session_id, value)
+
+
+@router.post("/{session_id}/validate")
+async def validate(
+    session_id: str, value: ValidateCaseRequest, source: str = Query("web", pattern="^(web|mcp)$")
+):
+    with domain_errors():
+        return await service.validate(session_id, value, source=source)
+
+
+@router.post("/{session_id}/demonstrations")
+def demonstrate(session_id: str, value: DemonstrationRequest):
+    with domain_errors():
+        return service.demonstrate(session_id, value.request_id, value.source)
+
+
+@router.post("/{session_id}/demonstrations/{demo_id}/extend")
+def extend_demonstration(session_id: str, demo_id: str, value: ExtendDemonstration):
+    with domain_errors():
+        return service.demonstration_command(
+            session_id, demo_id, "extend", value.request_id, seconds=value.seconds
+        )
+
+
+@router.post("/{session_id}/demonstrations/{demo_id}/{command}")
+def demonstration_command(session_id: str, demo_id: str, command: str, value: ExplorationRequest):
+    if command not in {"finish", "cancel"}:
+        raise HTTPException(404, "示范操作不存在")
+    with domain_errors():
+        return service.demonstration_command(session_id, demo_id, command, value.request_id)
+
+
+@router.post("/{session_id}/generate")
+async def generate_case(session_id: str, value: GenerateCaseRequest):
+    with domain_errors():
+        return await service.generate_case(session_id, value)
+
+
+@router.post("/{session_id}/batch")
+def case_batch(session_id: str, value: CaseBatchRequest):
+    with domain_errors():
+        return service.create_case_batch(session_id, value)

apps/admin_console/routers/favorites.py

@@ -0,0 +1,50 @@
+"""Favorite templates shared by the home page and MCP."""
+
+from fastapi import APIRouter, HTTPException
+
+from apps.admin_console.services.favorites import (
+    FavoriteCreate,
+    FavoriteRun,
+    FavoriteUpdate,
+    favorites,
+)
+
+router = APIRouter(prefix="/api/automation/favorites", tags=["automation"])
+
+
+@router.get("")
+def list_favorites():
+    return favorites.list()
+
+
+@router.post("")
+def create_favorite(value: FavoriteCreate):
+    try:
+        return favorites.create(value)
+    except (KeyError, ValueError) as exc:
+        raise HTTPException(409, str(exc)) from exc
+
+
+@router.patch("/{key}")
+def update_favorite(key: str, value: FavoriteUpdate):
+    try:
+        return favorites.update(key, value)
+    except (KeyError, ValueError) as exc:
+        raise HTTPException(409, str(exc)) from exc
+
+
+@router.delete("/{key}")
+def delete_favorite(key: str):
+    try:
+        favorites.delete(key)
+    except ValueError as exc:
+        raise HTTPException(400, str(exc)) from exc
+    return {"status": "deleted"}
+
+
+@router.post("/{key}/run")
+async def run_favorite(key: str, value: FavoriteRun):
+    try:
+        return await favorites.run(key, value)
+    except (KeyError, ValueError) as exc:
+        raise HTTPException(409, str(exc)) from exc

apps/admin_console/routers/media.py

@@ -19,10 +19,6 @@ from fastapi.responses import FileResponse, HTMLResponse
 
 from artemis.config import IMAGES_DIR, TRACES_PATH, WORKSPACE_ROOT
 
-try:
-    from admin_console.database.repositories.session_repository import session_repo
-    from admin_console.services.media_service import media_service
-except ImportError:
 from apps.admin_console.database.repositories.session_repository import session_repo
 from apps.admin_console.services.media_service import media_service
 
@@ -115,6 +111,77 @@ async def get_session_video(session_id: str):
 
 
 def _get_session_video_sync(session_id: str):
+    from apps.admin_console.services.evidence_service import plugin_sources
+    from apps.admin_console.services.exploration_evidence import demonstration_sources
+    from apps.admin_console.services.report_export import session_videos
+
+    legacy = _get_legacy_session_video_sync(session_id)
+    sources = []
+    if (
+        legacy.get("video_segments") or legacy.get("video_url")
+    ) and "/evidence/demonstrations/" not in str(legacy.get("video_url") or ""):
+        segments = legacy.get("video_segments") or []
+        # Pauses create sibling resume directories. Only collect files already
+        # attached to this task; unrelated recordings never enter its timeline.
+        if segments and all("offset_ms" in segment for segment in segments):
+            combined = {}
+            for path, _ in session_videos({"id": session_id}):
+                for segment in media_service.resolve_video_segments(
+                    media_service.path_to_video_url(path)
+                ):
+                    if "offset_ms" in segment:
+                        combined[segment["url"]] = segment
+            if combined:
+                segments = sorted(combined.values(), key=lambda segment: segment["offset_ms"])
+        sources.append(
+            {
+                "id": "android",
+                "kind": "android",
+                "label": "安卓",
+                "target_ids": ["android"],
+                "status": "ready",
+                "message": "",
+                "segments": [
+                    {**s, "start": float(s.get("offset_ms", s["start"] * 1000)) / 1000}
+                    for s in segments
+                ],
+                "url": legacy.get("video_url"),
+            }
+        )
+    sources += plugin_sources(session_id) + demonstration_sources(session_id)
+    legacy["sources"] = sources
+    legacy["session_timeline"] = any(source["segments"] for source in sources)
+    row = dict(session_repo.get_session_by_id(session_id) or {})
+    legacy["session_duration"] = max(
+        max(
+            0.0,
+            float(row.get("end_time") or row.get("start_time") or 0)
+            - float(row.get("start_time") or 0),
+        ),
+        max(
+            (s["start"] + s["duration"] for source in sources for s in source["segments"]),
+            default=0,
+        ),
+    )
+    ready = next((s for s in sources if s.get("segments") or s.get("url")), None)
+    if ready:
+        legacy.update(
+            status="ready",
+            has_video=True,
+            video_url=ready.get("url") or ready["segments"][0]["url"],
+            video_segments=ready["segments"],
+        )
+    elif any(s.get("status") == "recording" for s in sources):
+        legacy.update(status="processing", retry_after_ms=1000)
+    elif sources:
+        legacy.update(
+            status="failed",
+            message="；".join(str(s.get("message") or "录像不可用") for s in sources),
+        )
+    return legacy
+
+
+def _get_legacy_session_video_sync(session_id: str):
     video_rec_map = session_repo.get_video_recordings_map()
     video_idx = media_service.build_video_index()
     row = session_repo.get_session_by_id(session_id)

apps/admin_console/routers/plugins.py

@@ -0,0 +1,245 @@
+"""Plugin management for the local console; credentials never leave this API."""
+
+from concurrent.futures import ThreadPoolExecutor
+import io
+from pathlib import Path
+from threading import Lock
+from typing import Literal
+import uuid
+import zipfile
+
+from fastapi import APIRouter, HTTPException
+from fastapi.exceptions import RequestValidationError
+from fastapi.routing import APIRoute
+from fastapi.responses import JSONResponse
+from pydantic import BaseModel, Field, ValidationError
+
+from artemis.plugins.installer import download, install_archive, install_local, load_catalog
+from artemis.plugins.models import PluginManifest, PluginProfile, PluginSelection
+from artemis.plugins.runtime import PluginRuntime
+from artemis.plugins.store import PluginStore
+
+
+class PluginRoute(APIRoute):
+    def get_route_handler(self):
+        handler = super().get_route_handler()
+
+        async def sanitized(request):
+            try:
+                return await handler(request)
+            except RequestValidationError:
+                return JSONResponse(
+                    status_code=422,
+                    content={
+                        "detail": "插件配置格式不正确，请检查字段类型和清单；输入内容不会回显。"
+                    },
+                )
+
+        return sanitized
+
+
+router = APIRouter(prefix="/api/plugins", tags=["plugins"], route_class=PluginRoute)
+_executor = ThreadPoolExecutor(max_workers=1, thread_name_prefix="plugin-installer")
+_jobs: dict[str, dict] = {}
+_lock = Lock()
+
+
+def get_store() -> PluginStore:
+    return PluginStore()
+
+
+class InstallRequest(BaseModel):
+    plugin_id: str | None = None
+    version: str | None = None
+    local_path: str | None = None
+
+
+class EnableRequest(BaseModel):
+    enabled: bool
+
+
+class DiagnoseRequest(BaseModel):
+    profile: str = "default"
+
+
+class CustomRequest(BaseModel):
+    manifest: PluginManifest
+    skill_content: str = Field(default="", max_length=262144)
+
+
+class ResourceConfirmation(BaseModel):
+    resource: str = Field(min_length=1, max_length=256)
+    confirmed_stopped: Literal[True]
+
+
+def public_error(exc: Exception) -> str:
+    return "插件清单或配置格式不正确" if isinstance(exc, ValidationError) else str(exc)
+
+
+@router.get("")
+def list_plugins():
+    store = get_store()
+    return {"plugins": store.list(), "blocked_resources": store.blocked_resources()}
+
+
+@router.post("/resources/confirm-stopped")
+def confirm_resource_stopped(request: ResourceConfirmation):
+    from artemis.runtime.device_lock import DeviceExecutionLock
+
+    if any(
+        request.resource in owner.resource_ids
+        for owner in DeviceExecutionLock.get_active_owners().values()
+    ):
+        raise HTTPException(409, "此资源仍被任务占用，请先停止相关任务并确认插件动作已停止")
+    get_store().confirm_resource_stopped(request.resource)
+    return {"ok": True}
+
+
+@router.get("/catalog")
+def marketplace():
+    try:
+        return load_catalog().model_dump()
+    except Exception:
+        raise HTTPException(502, "暂时无法读取插件市场，请稍后重试；已安装插件不受影响") from None
+
+
+@router.post("/install", status_code=202)
+def install_plugin(request: InstallRequest):
+    if bool(request.local_path) == bool(request.plugin_id):
+        raise HTTPException(400, "请选择一个市场插件或本地安装包")
+    with _lock:
+        if any(j["status"] in {"queued", "running"} for j in _jobs.values()):
+            raise HTTPException(409, "已有插件正在安装，请等待安装结束")
+        for key in list(_jobs)[:-30]:
+            del _jobs[key]
+        job_id = str(uuid.uuid4())
+        _jobs[job_id] = {"id": job_id, "status": "queued", "stage": "等待安装", "progress": 0}
+
+    def progress(stage, percent):
+        with _lock:
+            _jobs[job_id].update(status="running", stage=stage, progress=percent)
+
+    def run():
+        try:
+            store = get_store()
+            if request.local_path:
+                installed = install_local(store, Path(request.local_path), progress)
+            else:
+                progress("获取市场索引", 0)
+                candidates = [
+                    p
+                    for p in load_catalog().plugins
+                    if p.manifest.id == request.plugin_id
+                    and (request.version is None or p.manifest.version == request.version)
+                ]
+                from packaging.version import Version
+
+                compatible = []
+                for entry in candidates:
+                    try:
+                        entry.manifest.check_compatible()
+                        compatible.append(entry)
+                    except ValueError:
+                        continue
+                if not compatible:
+                    raise ValueError("没有已验证且兼容的插件版本")
+                entry = max(compatible, key=lambda p: Version(p.manifest.version))
+                installed = install_archive(
+                    store, download(entry.archive_url, progress), entry=entry, progress=progress
+                )
+            with _lock:
+                _jobs[job_id].update(
+                    status="completed",
+                    stage="安装完成，请配置并启用",
+                    progress=100,
+                    plugin_id=installed["manifest"]["id"],
+                )
+        except Exception as exc:
+            with _lock:
+                # HTTP exceptions may contain signed URLs; expose only our validation messages.
+                message = (
+                    public_error(exc)
+                    if isinstance(exc, (ValueError, FileNotFoundError))
+                    else "下载或安装失败，请检查网络及安装包"
+                )
+                _jobs[job_id].update(status="failed", stage=message)
+
+    _executor.submit(run)
+    return {"job_id": job_id}
+
+
+@router.get("/install/{job_id}")
+def install_progress(job_id: str):
+    with _lock:
+        if job_id not in _jobs:
+            raise HTTPException(404, "安装记录不存在，服务可能已重启；请刷新插件列表")
+        return dict(_jobs[job_id])
+
+
+@router.post("/custom")
+def register_custom(request: CustomRequest):
+    buffer = io.BytesIO()
+    with zipfile.ZipFile(buffer, "w") as archive:
+        archive.writestr("plugin.json", request.manifest.model_dump_json())
+        if request.manifest.skill:
+            archive.writestr(request.manifest.skill, request.skill_content)
+    try:
+        return install_archive(get_store(), buffer.getvalue())
+    except (ValueError, OSError) as exc:
+        raise HTTPException(400, public_error(exc)) from None
+
+
+@router.put("/{plugin_id}/profiles")
+def save_profile(plugin_id: str, profile: PluginProfile):
+    try:
+        get_store().save_profile(plugin_id, profile)
+        return get_store().get(plugin_id)
+    except ValueError as exc:
+        raise HTTPException(400, public_error(exc)) from None
+
+
+@router.put("/{plugin_id}/enabled")
+def enable_plugin(plugin_id: str, request: EnableRequest):
+    try:
+        get_store().set_enabled(plugin_id, request.enabled)
+        return get_store().get(plugin_id)
+    except ValueError as exc:
+        raise HTTPException(400, public_error(exc)) from None
+
+
+@router.post("/{plugin_id}/diagnose")
+async def diagnose_plugin(plugin_id: str, request: DiagnoseRequest):
+    runtime = None
+    result = {"ok": False, "message": "插件连接失败，请检查启用状态、授权、命令和配置"}
+    try:
+        store = get_store()
+        manifest = PluginManifest.model_validate(store.get(plugin_id)["manifest"])
+        runtime = PluginRuntime(
+            [PluginSelection(id=plugin_id, profile=request.profile)],
+            store=store,
+            target=manifest.targets[0],
+        )
+        loaded = await runtime.load(plugin_id)
+        result = {
+            "ok": loaded.status == "success",
+            "tools": loaded.structured_content or {},
+            "message": "插件连接正常",
+        }
+    except Exception:
+        result = {"ok": False, "message": "插件连接失败，请检查启用状态、授权、命令和配置"}
+    finally:
+        if runtime:
+            try:
+                await runtime.aclose()
+            except RuntimeError:
+                result = {"ok": False, "message": "插件未确认停止，请检查受保护的资源"}
+    return result
+
+
+@router.delete("/{plugin_id}")
+def uninstall_plugin(plugin_id: str):
+    try:
+        get_store().uninstall(plugin_id)
+        return {"ok": True}
+    except (ValueError, OSError) as exc:
+        raise HTTPException(409, public_error(exc)) from None

apps/admin_console/routers/replay.py

@@ -17,10 +17,6 @@ from fastapi import APIRouter, HTTPException
 
 from artemis.config import WORKSPACE_ROOT
 
-try:
-    from admin_console.schemas.task_schema import ReplayRequest
-    from admin_console.replay_manager import ReplayManager
-except ImportError:
 from apps.admin_console.schemas.task_schema import ReplayRequest
 from apps.admin_console.replay_manager import ReplayManager
 

apps/admin_console/routers/sessions.py

@@ -16,28 +16,57 @@ import asyncio
 import json
 import time
 from uuid import UUID
-from fastapi import APIRouter, HTTPException
+from fastapi import APIRouter, HTTPException, Query
+from fastapi.responses import Response
 
 from artemis.config import DB_PATH, TRACES_PATH
 from artemis.runtime import trace_store
+from apps.admin_console.services.case_review import read_reviewed_result
 
-try:
-    from admin_console.core.state import state
-    from admin_console.database.repositories.session_repository import session_repo
-    from admin_console.database.repositories.trace_repository import trace_repo
-    from admin_console.services.media_service import media_service
-    from admin_console.services.model_service import model_service
-except ImportError:
 from apps.admin_console.core.state import state
 from apps.admin_console.database.repositories.session_repository import session_repo
 from apps.admin_console.database.repositories.trace_repository import trace_repo
 from apps.admin_console.services.media_service import media_service
 from apps.admin_console.services.model_service import model_service
+from apps.admin_console.services.automation_store import store
+from apps.admin_console.services.task_import import execution_title
+from apps.admin_console.services.exploration_listing import exploration_metadata
 
 
 router = APIRouter(tags=["sessions"])
 
 
+@router.get("/api/sessions/{session_id}/inspection")
+def inspect_session(
+    session_id: str,
+    view: str = "report",
+    offset: int = Query(default=0, ge=0),
+    limit: int = Query(default=20, ge=1, le=100),
+):
+    from apps.admin_console.services.task_inspection import inspect_task
+
+    try:
+        return inspect_task(session_id, view, offset=offset, limit=limit)
+    except LookupError as exc:
+        raise HTTPException(404, str(exc)) from exc
+    except ValueError as exc:
+        raise HTTPException(400, str(exc)) from exc
+
+
+@router.get("/api/sessions/{session_id}/request-diagnostics")
+def get_request_diagnostics(session_id: str, download: bool = False):
+    from apps.admin_console.services.request_diagnostics import request_diagnostics
+
+    result = request_diagnostics(session_id, limit=None if download else 100)
+    if not download:
+        return result
+    return Response(
+        json.dumps(result, ensure_ascii=False, indent=2),
+        media_type="application/json",
+        headers={"Content-Disposition": 'attachment; filename="model-request-diagnostics.json"'},
+    )
+
+
 @router.get("/api/sessions")
 async def list_sessions():
     # The body does blocking work (sqlite queries, filesystem scans, ffmpeg
@@ -52,6 +81,9 @@ async def list_sessions():
 
 def _list_sessions_sync():
     rows = session_repo.get_all_sessions()
+    with store.connection() as connection:
+        exploration_runs = exploration_metadata(connection)
+    jobs_by_id = store.jobs_for_sessions([str(row.get("session_id")) for row in rows])
     video_rec_map = session_repo.get_video_recordings_map()
     latest_recordings = session_repo.get_latest_video_recordings_map()
     agent_names_by_session = session_repo.get_agent_trace_names_map()
@@ -74,6 +106,19 @@ def _list_sessions_sync():
 
     for row_dict in rows:
         s_id = str(row_dict.get("session_id"))
+        row_dict.update(exploration_runs.get(s_id, {}))
+        job = jobs_by_id.get(s_id)
+        if job:
+            case = job.get("case") or {}
+            row_dict["batch_id"] = job.get("batch_id")
+            row_dict["case_id"] = case.get("case_id") or ""
+            row_dict["display_name"] = execution_title(job) or row_dict.get("initial_goal")
+            row_dict["iteration"] = job.get("iteration", 1)
+            row_dict["repeat_count"] = job.get("repeat_count", 1)
+            row_dict["error"] = job.get("error") or ""
+            if job.get("status") in {"completed", "failed", "cancelled", "interrupted", "skipped"}:
+                row_dict["status"] = job["status"]
+                row_dict["end_time"] = job.get("ended_at") or row_dict.get("end_time")
         recording = latest_recordings.get(s_id)
         d_info_raw = row_dict.get("device_info")
         device_id = None
@@ -114,6 +159,13 @@ def _list_sessions_sync():
         else:
             row_dict["video_url"] = None
         row_dict["recording_status"] = recording_status
+        row_dict["test_result"] = read_reviewed_result(
+            s_id,
+            row_dict.get("status") or "",
+            row_dict.get("error") or "",
+            root=TRACES_PATH,
+            storage=store,
+        )
 
         agent_names = agent_names_by_session.get(s_id, [])
         sess_profile = model_service.resolve_session_profile(
@@ -164,6 +216,19 @@ def _list_sessions_sync():
         except Exception as e:
             print(f"[list_sessions] Failed to update orphaned sessions: {e}")
 
+    for row in result:
+        if row.get("exploration_kind") == "validation" and row.get("exploration_case_name"):
+            row["display_name"] = row["exploration_case_name"]
+        model = row.pop("exploration_model", None)
+        if row.get("exploration_kind") == "exploration":
+            row["model_info"] = {
+                "name": "探索",
+                "id": model,
+                "provider": "openai",
+                "architecture": "explore",
+            }
+        elif model:
+            row["model_info"] = {**row["model_info"], "id": model}
     return result
 
 
@@ -173,7 +238,13 @@ async def get_session_details(session_id: str):
     row = session_repo.get_session_by_id(session_id)
     if not row:
         raise HTTPException(status_code=404, detail=f"Session {session_id} not found")
-    return dict(row)
+    result = dict(row)
+    job = store.job(session_id) or {}
+    result["batch_id"] = job.get("batch_id")
+    result["test_result"] = read_reviewed_result(
+        session_id, result.get("status") or "", root=TRACES_PATH, storage=store
+    )
+    return result
 
 
 @router.get("/api/sessions/{session_id}/usage")
@@ -227,11 +298,33 @@ async def cleanup_history_endpoint():
         raise HTTPException(status_code=500, detail=str(e))
 
 
+@router.get("/api/sessions/{session_id}/deletion-preview")
+def preview_session_deletion(session_id: str):
+    from apps.admin_console.services.task_deletion import deletion_preview
+
+    try:
+        return deletion_preview(session_id)
+    except KeyError as exc:
+        raise HTTPException(404, str(exc)) from exc
+    except ValueError as exc:
+        raise HTTPException(409, str(exc)) from exc
+
+
 @router.post("/api/sessions/{session_id}/delete")
 async def delete_session_endpoint(session_id: str):
+    from artemis.exploration.store import ExplorationConflict, explorations
+
     try:
         from artemis.data_engine.storage import StorageManager
 
+        session_id = str(UUID(session_id))
+        run = explorations.run(session_id)
+        if run and run["kind"] == "exploration":
+            from apps.admin_console.services.exploration_deletion import delete_exploration
+
+            deleted = delete_exploration(explorations, run["exploration_id"])
+            return {"status": "success", "deleted_session_ids": deleted}
+        preview_session_deletion(session_id)
         storage = StorageManager(DB_PATH, TRACES_PATH)
         storage.delete_session(UUID(session_id))
         from apps.admin_console.services.automation_store import store
@@ -240,6 +333,13 @@ async def delete_session_endpoint(session_id: str):
         return {
             "status": "success",
             "message": f"Session {session_id} deleted successfully",
+            "deleted_session_ids": [session_id],
         }
+    except ExplorationConflict as e:
+        raise HTTPException(status_code=409, detail=str(e)) from e
+    except HTTPException:
+        raise
+    except (ValueError, KeyError) as e:
+        raise HTTPException(status_code=400, detail=str(e)) from e
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))

apps/admin_console/routers/steps.py

@@ -19,11 +19,6 @@ from fastapi.responses import PlainTextResponse
 
 from artemis.config import TEST_OUTPUTS_DIR
 
-try:
-    from admin_console.database.repositories.step_repository import step_repo
-    from admin_console.database.repositories.trace_repository import trace_repo
-    from admin_console.services.media_service import media_service
-except ImportError:
 from apps.admin_console.database.repositories.step_repository import step_repo
 from apps.admin_console.database.repositories.trace_repository import trace_repo
 from apps.admin_console.services.media_service import media_service

apps/admin_console/routers/stream.py

@@ -20,9 +20,6 @@ Exposes real-time screen streaming endpoints for the Web UI.
 from fastapi import APIRouter
 from fastapi.responses import JSONResponse, StreamingResponse
 
-try:
-    from admin_console.services.device_stream_service import device_stream_service
-except ImportError:
 from apps.admin_console.services.device_stream_service import device_stream_service
 
 router = APIRouter(tags=["stream"])

apps/admin_console/routers/system.py

@@ -517,28 +517,47 @@ async def get_model_config_and_env():
 
 @router.get("/server-status")
 async def get_server_runtime_status():
-    """Retrieve runtime status, PID, port, and uptime of the Artemis server."""
+    """Return this server's status without scanning listening ports."""
     import os
-    from artemis.runtime.server_lifecycle import get_server_status
+    import time
+
+    from artemis.runtime.process_probe import pid_is_alive
+    from artemis.runtime.server_lifecycle import read_server_info
 
     try:
         from apps.admin_console.core.state import state
     except ImportError:
-        from admin_console.core.state import state
+        from apps.admin_console.core.state import state
 
     port = getattr(state, "port", 8000)
-    status = get_server_status(port=port)
+    current_pid = os.getpid()
+    pids = {current_pid}
+    started_at = None
+    info = read_server_info()
+    if info and info.get("port") == port:
+        saved_pid = info.get("pid")
+        if isinstance(saved_pid, int) and saved_pid != current_pid and pid_is_alive(saved_pid):
+            pids.add(saved_pid)
+        if saved_pid == current_pid and isinstance(info.get("started_at"), (int, float)):
+            started_at = float(info["started_at"])
+    if started_at is None:
+        try:
+            import psutil
+
+            started_at = psutil.Process(current_pid).create_time()
+        except Exception:
+            started_at = None
     # Explicit DTO: the raw metadata file additionally holds the lifecycle
     # token, cmdline, and filesystem paths, none of which belong on the wire.
     return {
-        "running": status["running"],
+        "running": True,
-        "port": status["port"],
+        "port": port,
-        "pids": status["pids"],
+        "pids": sorted(pids),
-        "active_pid": status["active_pid"],
+        "active_pid": current_pid,
-        "uptime_seconds": status["uptime_seconds"],
+        "uptime_seconds": max(0.0, time.time() - started_at) if started_at else None,
-        "url": status["url"],
+        "url": f"http://localhost:{port}",
-        "admin_url": status["admin_url"],
+        "admin_url": f"http://localhost:{port}/admin",
-        "current_pid": os.getpid(),
+        "current_pid": current_pid,
     }
 
 
@@ -555,7 +574,7 @@ async def restart_server_endpoint(request: Request):
     try:
         from apps.admin_console.core.state import state
     except ImportError:
-        from admin_console.core.state import state
+        from apps.admin_console.core.state import state
 
     port = getattr(state, "port", 8000)
     current_pid = os.getpid()
@@ -596,7 +615,7 @@ async def shutdown_server_endpoint(request: Request):
     try:
         from apps.admin_console.core.state import state
     except ImportError:
-        from admin_console.core.state import state
+        from apps.admin_console.core.state import state
 
     _require_local_lifecycle_request(request)
     server = getattr(request.app.state, "uvicorn_server", None)

apps/admin_console/routers/tasks.py

@@ -16,20 +16,24 @@ import asyncio
 from contextlib import suppress
 import json
 from typing import Any
+from uuid import uuid4
 from fastapi import APIRouter, HTTPException, Request
 from fastapi.responses import StreamingResponse
+from pydantic import BaseModel, Field
 from artemis.core.diagnostics import readiness_engine
+from apps.admin_console.services.automation_store import store
+from apps.admin_console.services.exploration_listing import exploration_metadata
+from apps.admin_console.services.task_import import execution_title
 from artemis.runtime import DeviceExecutionLock, device_pool
+from artemis.runtime.task_control import (
+    add_instruction,
+    control_dir,
+    list_instructions,
+    read_state,
+    request_pause,
+    request_resume,
+)
 
-try:
-    from admin_console.core.state import state
-    from admin_console.database.repositories.session_repository import session_repo
-    from admin_console.schemas.task_schema import RunRequest
-    from admin_console.services.ipc_service import ipc_service
-    from admin_console.services.model_service import model_service
-    from admin_console.services.task_preset_catalog import task_recommendation_engine
-    from admin_console.services.task_queue_service import task_queue_service
-except ImportError:
 from apps.admin_console.core.state import state
 from apps.admin_console.database.repositories.session_repository import session_repo
 from apps.admin_console.schemas.task_schema import RunRequest
@@ -42,6 +46,76 @@ except ImportError:
 router = APIRouter(tags=["tasks"])
 
 
+class GuidanceRequest(BaseModel):
+    instruction: str = Field(min_length=1, max_length=10000)
+    request_id: str | None = Field(default=None, min_length=1, max_length=200)
+
+
+def _live_task(session_id: str) -> dict:
+    from apps.admin_console.services.automation_store import identifier
+
+    identifier(session_id)
+    run = state.active_runs.get(session_id)
+    if not run or run["process"].returncode is not None:
+        raise HTTPException(409, "任务未在执行，无法暂停或指导")
+    return run
+
+
+@router.post("/api/tasks/{session_id}/pause")
+def pause_running_task(session_id: str):
+    _live_task(session_id)
+    request_pause(session_id)
+    return {"session_id": session_id, "status": "pausing"}
+
+
+@router.post("/api/tasks/{session_id}/guide")
+def guide_running_task(session_id: str, value: GuidanceRequest):
+    _live_task(session_id)
+    from apps.admin_console.services.exploration_service import exploration_service
+    from artemis.exploration.store import ExplorationConflict
+
+    if run := exploration_service.repository.run(session_id):
+        try:
+            guidance = exploration_service.message(
+                run["exploration_id"],
+                value.instruction,
+                value.request_id or str(uuid4()),
+                expected_run_id=session_id,
+            )
+        except ExplorationConflict as exc:
+            raise HTTPException(409, str(exc)) from exc
+    else:
+        guidance = add_instruction(
+            session_id, value.instruction, source="web", instruction_id=value.request_id
+        )
+    if (
+        read_state(session_id).get("status") == "task_paused"
+        or (control_dir(session_id) / "pause.request").exists()
+    ):
+        request_resume(session_id)
+    return {"session_id": session_id, "status": "guidance_submitted", "guidance": guidance}
+
+
+@router.get("/api/tasks/{session_id}/guidance")
+def task_guidance_history(session_id: str):
+    from apps.admin_console.services.automation_store import identifier
+
+    try:
+        identifier(session_id)
+    except ValueError as exc:
+        raise HTTPException(400, str(exc)) from exc
+    return list_instructions(session_id)
+
+
+@router.post("/api/tasks/{session_id}/continue")
+def continue_running_task(session_id: str):
+    _live_task(session_id)
+    if read_state(session_id).get("status") != "task_paused":
+        raise HTTPException(409, "任务尚未暂停")
+    request_resume(session_id)
+    return {"session_id": session_id, "status": "resuming"}
+
+
 @router.get("/api/tasks/presets")
 async def get_task_presets(
     category: str = "recommended",
@@ -119,7 +193,10 @@ async def run_task(request: RunRequest):
     # never be selected. Only a successful, non-empty enumeration may reject:
     # an indeterminate one (adb blip, startup) lets the submission queue and
     # fail downstream with a clear error instead.
-    if request.device_serial:
+    execution_target = (
+        request.runtime_overrides.execution_target if request.runtime_overrides else "android"
+    )
+    if request.device_serial and execution_target != "computer":
         try:
             rejection = await device_pool.validate_explicit_serial_async(request.device_serial)
         except Exception:
@@ -140,8 +217,14 @@ async def run_task(request: RunRequest):
     # With no explicit serial the probe itself resolves a live target (it
     # prefers the diagnostics target preference, then any unlocked ready
     # device); the verified serial is bound below.
-    target_serial = request.device_serial
+    target_serial = (
-    device_probe = await readiness_engine.run_device_submission_probe(target_serial=target_serial)
+        request.runtime_overrides.device_serial if request.runtime_overrides else None
+    ) or request.device_serial
+    device_probe = (
+        await readiness_engine.run_device_submission_probe(target_serial=target_serial)
+        if execution_target != "computer"
+        else None
+    )
     if device_probe and device_probe.summary in {"Device Locked", "Lock State Unknown"}:
         locked_serial = (
             device_probe.metadata.get("active_device", {}).get("serial") or target_serial or ""
@@ -162,7 +245,7 @@ async def run_task(request: RunRequest):
         # Only auto-selected targets may be re-bound to the probed device. An
         # explicitly requested serial is never silently replaced -- if it is
         # invalid, enqueue_tasks rejects the submission with a clear error.
-        if verified_serial and not request.device_serial:
+        if verified_serial and not target_serial:
             target_serial = verified_serial
 
     response = await task_queue_service.enqueue_tasks(
@@ -178,10 +261,15 @@ async def run_task(request: RunRequest):
         ingress=request.ingress or "frontend",
         session_id=request.session_id,
         conversation_id=request.conversation_id,
+        runtime_overrides=request.runtime_overrides.model_dump(exclude_none=True)
+        if request.runtime_overrides
+        else None,
     )
 
     from apps.admin_console.services.automation_service import register_submission
 
+    if response.get("status") == "error":
+        raise HTTPException(400, response.get("message", "任务参数无效"))
     for item in response.get("tasks", []):
         register_submission(item)
     return response
@@ -207,7 +295,12 @@ async def get_run_defaults():
 async def list_devices():
     """List all connected Android devices with their busy / idle status."""
     devices = await device_pool.list_devices_async()
-    return {"devices": [d.to_dict() for d in devices]}
+    from artemis.runtime.adb_discovery import adb_device_discovery
+
+    return {
+        "devices": [d.to_dict() for d in devices],
+        "automatic_connections": adb_device_discovery.results.copy(),
+    }
 
 
 @router.post("/api/stop")
@@ -285,6 +378,17 @@ async def get_status():
         }
         for owner in active_owners.values()
     ]
+    for sid, run in state.active_runs.items():
+        if run.get("parked"):
+            active_tasks.append(
+                {
+                    "session_id": sid,
+                    "device_id": run.get("device_id"),
+                    "goal": run.get("goal"),
+                    "pid": getattr(run.get("process"), "pid", None),
+                    "status": read_state(sid).get("status") or "task_paused",
+                }
+            )
 
     running_sid = (
         (global_owner.session_id if global_owner else None)
@@ -340,6 +444,20 @@ async def get_status():
             seen_ids.add(sid)
             queue_data.append(g_item)
 
+    batch_jobs = store.jobs_for_sessions(
+        [str(item["session_id"]) for item in [*queue_data, *active_tasks] if item.get("session_id")]
+    )
+    queue_data = [{**item} for item in queue_data]
+    with store.connection() as connection:
+        exploration_runs = exploration_metadata(connection)
+    for item in [*queue_data, *active_tasks]:
+        item.update(exploration_runs.get(str(item.get("session_id")), {}))
+        if job := batch_jobs.get(str(item.get("session_id"))):
+            item["batch_id"] = job.get("batch_id")
+            item["display_name"] = execution_title(job)
+        if item.get("exploration_kind") == "validation" and item.get("exploration_case_name"):
+            item["display_name"] = item["exploration_case_name"]
+
     if is_running:
         is_paused = state.is_paused
         return {

apps/admin_console/schemas/task_schema.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 from pydantic import BaseModel
+from apps.admin_console.services.task_import import TaskOptions
 
 
 class RunRequest(BaseModel):
@@ -32,6 +33,7 @@ class RunRequest(BaseModel):
     ingress: str | None = "frontend"
     session_id: str | None = None
     conversation_id: str | None = None
+    runtime_overrides: TaskOptions | None = None
 
 
 class ReplayRequest(BaseModel):

apps/admin_console/server.py

@@ -24,6 +24,7 @@ import os
 from pathlib import Path
 import secrets
 import signal
+import socket
 import sys
 from types import FrameType
 
@@ -45,7 +46,7 @@ for _p in (str(_workspace_root), str(_apps_dir), str(_admin_console_dir), str(_c
     if _p not in sys.path:
         sys.path.insert(0, _p)
 
-from fastapi import FastAPI
+from fastapi import FastAPI, HTTPException, Request
 from fastapi.responses import FileResponse, HTMLResponse
 import uvicorn
 
@@ -70,26 +71,16 @@ from artemis.config import (
 )
 from artemis.resources import get_bundled_showcase_dist
 
-try:
-    from admin_console.core.security import SameOriginBoundaryMiddleware
-    from admin_console.core.state import ServerState, state
-    from admin_console.database.connection import db_session, get_db
-    from admin_console.database.repositories.session_repository import session_repo
-    from admin_console.database.repositories.step_repository import step_repo
-    from admin_console.database.repositories.trace_repository import trace_repo
-    from admin_console.routers import media, replay, sessions, steps, stream, system, tasks
-    from admin_console.routers.replay import replay_manager
-    from admin_console.services.ipc_service import ipc_service
-    from admin_console.services.media_service import media_service
-    from admin_console.services.model_service import model_service
-    from admin_console.services.task_queue_service import task_queue_service
-except ImportError:
 from apps.admin_console.core.security import SameOriginBoundaryMiddleware
-    from apps.admin_console.core.state import state
+from apps.admin_console.core.state import ServerState, state
+from apps.admin_console.database.connection import db_session, get_db
 from apps.admin_console.database.repositories.session_repository import session_repo
+from apps.admin_console.database.repositories.step_repository import step_repo
+from apps.admin_console.database.repositories.trace_repository import trace_repo
 from apps.admin_console.routers import media, replay, sessions, steps, stream, system, tasks
 from apps.admin_console.routers.replay import replay_manager
 from apps.admin_console.services.ipc_service import ipc_service
+from apps.admin_console.services.media_service import media_service
 from apps.admin_console.services.model_service import model_service
 from apps.admin_console.services.task_queue_service import task_queue_service
 
@@ -100,11 +91,37 @@ init_ls_address()
 @asynccontextmanager
 async def _lifespan(_app: "FastAPI"):
     """Start server resources and release them on shutdown."""
+    from artemis.config.paths import get_app_dir, get_traces_dir
+    from artemis.runtime.daemon_client import configure_local_daemon
+    from artemis.service_registry import ServiceRegistration
+
+    registration = ServiceRegistration(
+        port=state.port,
+        token=LIFECYCLE_TOKEN,
+        app_dir=get_app_dir(),
+        traces_dir=get_traces_dir(),
+    )
+    registration.acquire()
+    app.state.service_registration = registration
+    configure_local_daemon(state.port)
+    try:
+        from artemis.integrations.codex_skill import sync_on_startup
+        from artemis.config.paths import ROOT_DIR
+
+        await asyncio.to_thread(sync_on_startup, ROOT_DIR)
+        shared_tool_executor.start()
+        async with shared_mcp.session_manager.run():
             await on_startup()
             try:
                 yield
             finally:
                 await on_shutdown()
+    finally:
+        try:
+            await shared_tool_executor.stop()
+        finally:
+            configure_local_daemon(None)
+            registration.close()
 
 
 # Initialize FastAPI application
@@ -117,6 +134,26 @@ app.state.lifecycle_token = LIFECYCLE_TOKEN
 # and unrecognized Host headers (DNS rebinding) instead.
 app.add_middleware(SameOriginBoundaryMiddleware)
 
+from mcp_server.shared import LocalMcpBoundary, create_shared_mcp
+
+shared_mcp, shared_tool_executor = create_shared_mcp()
+app.mount("/api/mcp", LocalMcpBoundary(shared_mcp.streamable_http_app(), LIFECYCLE_TOKEN))
+
+
+@app.get("/api/system/shared-service")
+async def shared_service_identity(request: Request):
+    """Authenticated identity probe for local connection-only MCP bridges."""
+    import hmac
+
+    from artemis.service_registry import TOKEN_HEADER
+
+    if not hmac.compare_digest(request.headers.get(TOKEN_HEADER, ""), LIFECYCLE_TOKEN):
+        raise HTTPException(status_code=403, detail="Invalid service token")
+    registration = getattr(app.state, "service_registration", None)
+    if registration is None:
+        raise HTTPException(status_code=503, detail="Service is starting")
+    return {key: registration.info[key] for key in ("protocol", "installation", "instance", "pid")}
+
 
 async def on_startup():
     """Startup lifecycle hooks."""
@@ -152,7 +189,7 @@ async def on_startup():
 
     cleaned = session_repo.cleanup_orphans_on_startup()
     if cleaned > 0:
-        print(f"[ServerStartup] Marked {cleaned} orphaned running session(s) as failed.")
+        print(f"[ServerStartup] Marked {cleaned} orphaned running session(s) as interrupted.")
     # Workers killed together with a previous daemon never remuxed their
     # recordings; publish whatever raw files they left behind.
     asyncio.create_task(asyncio.to_thread(task_queue_service.recover_orphaned_recordings_on_launch))
@@ -176,7 +213,9 @@ async def on_shutdown():
     owned_session_ids = {
         str(item["session_id"])
         for item in state.queue_items
-        if isinstance(item, dict) and item.get("status") == "running" and item.get("session_id")
+        if isinstance(item, dict)
+        and item.get("status") in {"running", "task_paused", "resuming"}
+        and item.get("session_id")
     }
 
     worker = state.worker_task
@@ -212,7 +251,7 @@ async def on_shutdown():
     state.current_process = None
     state.queue_items.clear()
     for session_id in owned_session_ids:
-        session_repo.update_session_status(session_id, "cancelled")
+        session_repo.update_session_status(session_id, "interrupted")
 
     from apps.admin_console.services.automation_service import background_tasks
 
@@ -238,6 +277,17 @@ app.include_router(system.router)
 from apps.admin_console.routers.automation import router as automation_router
 
 app.include_router(automation_router)
+from apps.admin_console.routers.favorites import router as favorites_router
+
+app.include_router(favorites_router)
+
+from apps.admin_console.routers.plugins import router as plugins_router
+
+app.include_router(plugins_router)
+
+from apps.admin_console.routers.exploration import router as exploration_router
+
+app.include_router(exploration_router)
 
 # Mount cloud gateway router for Frappe / Cloud integration if present
 try:
@@ -419,6 +469,14 @@ queue_goals = state.queue_goals
 class ArtemisUvicornServer(uvicorn.Server):
     """Notify application streams before Uvicorn waits for them to close."""
 
+    async def startup(self, sockets: list[socket.socket] | None = None) -> None:
+        # Reload workers start in a fresh interpreter; restore the actual bind
+        # address there before lifespan publishes runtime metadata.
+        state.host = self.config.host
+        state.port = self.config.port
+        app.state.uvicorn_server = self
+        await super().startup(sockets=sockets)
+
     @staticmethod
     def _task_is_active() -> bool:
         """Check task state without doing process I/O from a signal handler."""
@@ -468,31 +526,40 @@ class ArtemisUvicornServer(uvicorn.Server):
 
 def run_ui_server(host: str, port: int, reload: bool = False) -> None:
     """Run the UI server with bounded, signal-aware graceful shutdown."""
+    # Reserve a listening socket before publishing metadata or running lifespan.
+    # Uvicorn normally runs application startup *before* binding its port; a
+    # duplicate daemon would otherwise recover jobs owned by the real server.
+    family = socket.AF_INET6 if ":" in host else socket.AF_INET
+    try:
+        listener = socket.create_server((host, port), family=family)
+    except OSError as exc:
+        print(f"无法启动服务：{host}:{port} 端口不可用（{exc}）。现有服务和任务保持不变。")
+        raise SystemExit(1) from exc
+    listener.set_inheritable(True)
+    port = listener.getsockname()[1]
     state.host = host
     state.port = port
-    write_server_info(port=port, host=host, lifecycle_token=LIFECYCLE_TOKEN)
 
     try:
-        if reload:
+        config = uvicorn.Config(
-            uvicorn.run(
+            "apps.admin_console.server:app" if reload else app,
-                "apps.admin_console.server:app",
             host=host,
             port=port,
-                reload=True,
+            reload=reload,
             timeout_graceful_shutdown=5,
         )
+        if reload:
+            from uvicorn.supervisors import ChangeReload
+
+            server = ArtemisUvicornServer(config)
+            ChangeReload(config, target=server.run, sockets=[listener]).run()
             return
 
-        config = uvicorn.Config(
-            app,
-            host=host,
-            port=port,
-            timeout_graceful_shutdown=5,
-        )
         server = ArtemisUvicornServer(config)
         app.state.uvicorn_server = server
-        server.run()
+        server.run(sockets=[listener])
     finally:
+        listener.close()
         app.state.uvicorn_server = None
         clear_server_info(port=port, lifecycle_token=LIFECYCLE_TOKEN)
 

apps/admin_console/services/automation_service.py

@@ -3,11 +3,15 @@
 import asyncio
 import logging
 import os
+import shutil
 import time
+from uuid import UUID
 
 from apps.admin_console.core.state import state
-from apps.admin_console.services.automation_store import store, new_id
+from apps.admin_console.services.automation_store import RECOVERY_ERRORS, store, new_id
-from apps.admin_console.services.task_import import case_goal
+from apps.admin_console.services.task_import import case_goal, execution_title
+from artemis.config.settings import settings
+from artemis.runtime.task_limits import task_timeout
 
 logger = logging.getLogger(__name__)
 TERMINAL = {"completed", "failed", "cancelled", "skipped", "interrupted"}
@@ -32,22 +36,78 @@ def register_submission(item: dict, case: dict | None = None, formats=None, **ex
             "status": "queued",
             "created_at": time.time(),
             "device_serial": item.get("device_serial"),
+            "execution_config": item.get("execution_config"),
+            "submission": {
+                key: item.get(key)
+                for key in (
+                    "goal",
+                    "profile",
+                    "device_serial",
+                    "runtime_overrides",
+                    "expected_output",
+                    "enable_outputter",
+                    "verification_level",
+                    "explorer_mode",
+                    "locked_app_package",
+                    "app_path",
+                )
+            },
             "formats": formats if formats is not None else store.preferences()["formats"],
             **extra,
         }
     )
 
 
+def resolve_recovery_error(job: dict) -> bool:
+    """Archive a superseded recovery warning without erasing execution failures."""
+    if job.get("error") not in RECOVERY_ERRORS:
+        return False
+    job.setdefault("recovery_events", []).append(
+        {"message": job.pop("error"), "resolved_at": time.time()}
+    )
+    return True
+
+
 def recover_batches():
     """Recovery never silently repeats a potentially side-effecting interrupted task."""
+    from apps.admin_console.database.repositories.session_repository import SessionRepository
+    from artemis.runtime import DeviceExecutionLock
+
+    repository = SessionRepository(store.path)
+    live_sessions = {
+        str(owner.session_id)
+        for owner in DeviceExecutionLock.get_active_owners().values()
+        if owner.session_id
+    }
+    live_sessions.update(
+        str(row["session_id"])
+        for row in repository.get_all_sessions()
+        if row.get("status") not in TERMINAL and repository.process_is_alive(row.get("pid"))
+    )
+    live_sessions.update(
+        str(item["session_id"])
+        for item in state.queue_items
+        if isinstance(item, dict) and item.get("session_id") and item.get("status") not in TERMINAL
+    )
     for job in store.jobs():
-        if job["status"] in {"running", "queued", "finalizing"}:
+        if job["status"] == "completed" and resolve_recovery_error(job):
+            store.save_job(job)
+        if str(job["session_id"]) in live_sessions:
+            if job["status"] not in TERMINAL and str(job["session_id"]) not in state.active_runs:
+                job["recovery_waiting"] = True
+                store.save_job(job)
+            logger.info("Recovery preserved live session %s", job["session_id"])
+            continue
+        if job["status"] in {"running", "queued", "finalizing", "task_paused", "resuming"}:
             job.update(
                 status=job.get("terminal_status", "interrupted"),
                 error=job.get("error")
                 or ("" if job.get("terminal_status") else "服务中断；未完成任务需手动重跑"),
                 ended_at=job.get("ended_at") or time.time(),
             )
+            if job["status"] == "completed":
+                resolve_recovery_error(job)
+            logger.info("Recovery resolved session %s to %s", job["session_id"], job["status"])
             store.save_job(job)
     for batch in store.records("batch"):
         if has_legacy_replay(batch["id"]):
@@ -60,11 +120,13 @@ def recover_batches():
                 store.put("batch", batch["id"], batch)
             continue
         if batch["status"] in {"running", "pausing", "paused"}:
+            if any(str(job["session_id"]) in live_sessions for job in store.jobs(batch["id"])):
+                continue
             batch.update(
                 status="paused", reason="服务已重启；请检查中断任务，点击继续执行未开始任务"
             )
             for job in store.jobs(batch["id"]):
-                if job["status"] in {"running", "queued", "finalizing"}:
+                if job["status"] in {"running", "queued", "finalizing", "task_paused", "resuming"}:
                     job.update(
                         status="interrupted", error="服务中断，需手动重跑", ended_at=time.time()
                     )
@@ -82,20 +144,29 @@ async def on_terminal(item: dict, status: str):
         job = store.job(sid)
     if job is None:
         raise RuntimeError("任务元数据写入失败")
-    job.update(status="finalizing", ended_at=time.time())
     if job.get("timed_out"):
         status = "failed"
+    job.pop("recovery_waiting", None)
+    if status == "completed":
+        resolve_recovery_error(job)
+    job.update(status="finalizing", terminal_status=status, ended_at=time.time())
     store.save_job(job)
     try:
         job["status"] = status
         report = await asyncio.to_thread(session_report, sid)
-        if report["verdict"] == "验证失败" or report["outcome"].get("task_status") in {
+        if status != "interrupted" and (
+            report["verdict"] == "验证失败"
+            or report["outcome"].get("task_status")
+            in {
                 "blocked",
                 "partial",
                 "failed",
-        }:
+            }
+        ):
             job["status"] = "failed"
-            job["error"] = "任务执行结束，但业务检查未通过"
+            job["error"] = job.get("error") or "任务执行结束，但业务检查未通过"
+        elif job["status"] == "failed" and report.get("error") and not job.get("error"):
+            job["error"] = report["error"]
         # Persist terminal classification before export, but the coordinator waits for finalization.
         store.save_job({**job, "status": "finalizing", "terminal_status": job["status"]})
         await asyncio.to_thread(export_session, sid, job["formats"])
@@ -103,14 +174,44 @@ async def on_terminal(item: dict, status: str):
     except Exception as exc:
         logger.exception("Report finalization failed for %s", sid)
         job["report_error"] = str(exc)[:500]
+    job.pop("terminal_status", None)
     store.save_job(job)
 
 
+async def reconcile_recovered_jobs():
+    """Finalize surviving workers once they exit; never resubmit their goals."""
+    from apps.admin_console.database.repositories.session_repository import SessionRepository
+    from artemis.runtime import DeviceExecutionLock
+    from artemis.runtime.task_result import read_worker_outcome
+
+    waiting = [job for job in store.jobs() if job.get("recovery_waiting")]
+    if not waiting:
+        return
+    repository = SessionRepository(store.path)
+    owners = {str(owner.session_id) for owner in DeviceExecutionLock.get_active_owners().values()}
+    for job in waiting:
+        sid = str(job["session_id"])
+        row = repository.get_session_by_id(sid) or {}
+        if sid in owners or repository.process_is_alive(row.get("pid")):
+            continue
+        outcome = read_worker_outcome(sid)
+        status = job.get("terminal_status") or outcome.get("task_status") or row.get("status")
+        if status not in TERMINAL:
+            status = "interrupted"
+            job["error"] = job.get("error") or "服务中断；未完成任务需手动重跑"
+            store.save_job(job)
+        await on_terminal({"session_id": sid}, status)
+
+
 async def tick_batches():
     from apps.admin_console.server import task_queue_service
     from artemis.runtime import device_pool
     from apps.admin_console.services.report_export import export_batch
 
+    await reconcile_recovered_jobs()
+    from apps.admin_console.services.exploration_recovery import reconcile_explorations
+
+    reconcile_explorations()
     for batch in store.records("batch"):
         if batch["status"] not in {"running", "paused"}:
             continue
@@ -121,17 +222,41 @@ async def tick_batches():
                 batch.update(status="paused", reason="模型请求已暂停，请先恢复模型连接再继续批次")
                 store.put("batch", batch["id"], batch)
             for job in active:
-                limit = int(batch.get("timeout_seconds", 1800))
+                # The queue records started_at at submission, then replaces it
+                # when the worker actually starts. Waiting for a device must
+                # not consume this case's execution budget.
+                if job["status"] != "running":
+                    continue
+                from artemis.runtime.task_result import read_worker_outcome
+
+                if read_worker_outcome(str(job["session_id"])):
+                    continue  # Execution has ended; the worker supervisor bounds cleanup.
+                limit = task_timeout({}, job, batch)
+                task_state = getattr(state, "active_runs", {}).get(str(job["session_id"])) or {}
+                excluded = float(task_state.get("non_execution_seconds") or 0)
+                if task_state.get("non_execution_started_at"):
+                    excluded += time.time() - float(task_state["non_execution_started_at"])
                 if (
-                    job["status"] != "finalizing"
+                    not job.get("timed_out")
-                    and not job.get("timed_out")
+                    and time.time() - job.get("started_at", time.time()) - excluded > limit
-                    and time.time() - job.get("started_at", time.time()) > limit
                 ):
-                    job.update(error="任务达到批次配置的执行时限", timed_out=True)
+                    job.update(
+                        error=f"任务执行超时（{limit} 秒，不含暂停/排队时间）", timed_out=True
+                    )
                     store.save_job(job)
+                    from artemis.runtime.task_result import save_failure_result
+
+                    try:
+                        await asyncio.to_thread(
+                            save_failure_result,
+                            str(job["session_id"]),
+                            job["error"],
+                            code="task_timeout",
+                        )
+                    finally:
                         task_queue_service.stop_tasks(session_id=job["session_id"])
-                    batch.update(status="paused", reason="任务超时，已暂停剩余任务；请确认设备状态")
+                    # on_terminal records the failed case. The next pending
-                    store.put("batch", batch["id"], batch)
+                    # case then runs normally unless stop_on_failure is set.
             continue
         if batch["status"] != "running":
             continue
@@ -142,6 +267,11 @@ async def tick_batches():
                 job.update(status="skipped", error="本批次设置为失败即停止", ended_at=time.time())
                 store.save_job(job)
             pending = []
+        if not pending and any(
+            j["status"] in {"task_paused", "resuming", "running", "queued", "finalizing"}
+            for j in jobs
+        ):
+            continue
         if not pending:
             batch.update(status="completed", ended_at=time.time())
             try:
@@ -156,11 +286,15 @@ async def tick_batches():
             batch.update(status="paused", reason="模型请求已暂停，请先恢复模型连接再继续批次")
             store.put("batch", batch["id"], batch)
             continue
-        serial = batch.get("device_serial")
+        job = pending[0]
-        devices = await device_pool.try_list_devices_async()
+        options = job["case"].get("task_config") or {}
+        serial = options.get("device_serial") or batch.get("device_serial")
+        computer = options.get("execution_target") == "computer"
+        devices = await device_pool.try_list_devices_async() if not computer else []
         rejection = (
             None
-            if serial and any(d.serial == serial and d.state == "device" for d in devices or [])
+            if computer
+            or (serial and any(d.serial == serial and d.state == "device" for d in devices or []))
             else "无法确认所选设备在线且已授权，请恢复设备连接后继续批次"
         )
         if rejection:
@@ -169,26 +303,29 @@ async def tick_batches():
             continue
         if (store.get("batch", batch["id"]) or {}).get("status") != "running":
             continue
-        job = pending[0]
         job.update(status="queued", started_at=time.time())
         store.save_job(job)
         try:
             response = await task_queue_service.enqueue_tasks(
                 [case_goal(job["case"])],
-                profile=batch.get("profile", "flash"),
+                profile=options.get("profile") or batch.get("profile", "flash"),
-                expected_output="用简体中文报告实际结果与证据。明确区分执行完成、验证通过、验证失败和前置条件未满足。"
+                expected_output="用简体中文报告实际结果与证据。明确区分执行完成、验证通过、验证失败和前置条件未满足。",
-                + job["case"].get("expected", ""),
                 enable_outputter=True,
                 device_serial=serial,
+                verification_level=options.get("verification_level"),
+                explorer_mode=options.get("explorer_mode"),
+                locked_app_package=options.get("locked_app_package"),
+                runtime_overrides=options,
                 session_id=job["session_id"],
                 ingress="batch",
+                display_name=execution_title(job),
             )
             if response.get("status") == "rejected" or not response.get("tasks"):
                 raise ValueError(response.get("error") or "队列拒绝接收任务")
             bound = response["tasks"][0].get("device_serial")
-            if bound and not serial:
+            if bound:
-                batch["device_serial"] = bound
+                job["device_serial"] = bound
-                store.put("batch", batch["id"], batch)
+                store.save_job(job)
         except Exception as exc:
             job.update(status="pending", error=str(exc)[:500])
             store.save_job(job)
@@ -214,6 +351,15 @@ def has_legacy_replay(batch_id: str) -> bool:
 
 
 def create_batch(payload: dict) -> dict:
+    repeat_count = payload.get("repeat_count", 1)
+    if type(repeat_count) is not int or repeat_count < 1:
+        raise ValueError("循环次数必须为正整数")
+    cases = payload["cases"]
+    total = len(cases) * repeat_count
+    if not cases or total > settings.ARTEMIS_BATCH_MAX_EXECUTIONS:
+        raise ValueError(
+            f"请选择用例；用例数 × 循环次数不能超过 {settings.ARTEMIS_BATCH_MAX_EXECUTIONS} 次执行"
+        )
     bid = new_id()
     batch = {
         "id": bid,
@@ -223,28 +369,65 @@ def create_batch(payload: dict) -> dict:
         "formats": payload.get("formats") or store.preferences()["formats"],
         "profile": payload.get("profile", "flash"),
         "device_serial": payload.get("device_serial"),
-        "timeout_seconds": payload.get("timeout_seconds")
+        "timeout_seconds": payload.get("timeout_seconds") or settings.ARTEMIS_TASK_TIMEOUT_SECONDS,
-        or int(os.getenv("ARTEMIS_BATCH_TASK_TIMEOUT", "1800")),
         "stop_on_failure": payload.get("stop_on_failure", False),
         "request_id": payload.get("request_id"),
+        "repeat_count": repeat_count,
+        "case_count": len(cases),
+        "total_executions": total,
     }
     jobs = []
-    for number, case in enumerate(payload["cases"], 1):
+    for iteration in range(1, repeat_count + 1):
+        for number, case in enumerate(cases, 1):
             jobs.append(
                 {
                     "session_id": new_id(),
                     "batch_id": bid,
-                "position": number,
+                    "position": len(jobs) + 1,
+                    "iteration": iteration,
+                    "repeat_count": repeat_count,
+                    "case_position": number,
                     "case": case,
                     "status": "pending",
                     "formats": batch["formats"],
-                "created_at": time.time(),
+                    "created_at": batch["created_at"],
                     "device_serial": batch["device_serial"],
                 }
             )
     return store.create_batch(batch, jobs, payload)
 
 
+def delete_batch(batch_id: str) -> bool:
+    """Delete an inactive batch, its sessions and its generated reports."""
+    from artemis.data_engine.storage import StorageManager
+    from apps.admin_console.services import report_export
+
+    jobs = store.begin_delete_batch(batch_id)
+    if jobs is None:
+        return False
+    storage = StorageManager(store.path, report_export.TRACES_PATH)
+    mcp_exports = report_export.EXPORT_ROOT.parent / "mcp_exports"
+    with report_export.export_lock(
+        "batch:" + batch_id, *("session:" + j["session_id"] for j in jobs)
+    ):
+        for job in jobs:
+            session_id = str(UUID(job["session_id"]))
+            storage.delete_session(UUID(session_id))
+            report_dir = report_export.EXPORT_ROOT / "sessions" / session_id
+            if report_dir.is_dir():
+                shutil.rmtree(report_dir)
+            mcp_session_dir = mcp_exports / "sessions" / session_id
+            if mcp_session_dir.is_dir():
+                shutil.rmtree(mcp_session_dir)
+        batch_dir = report_export.EXPORT_ROOT / "batches" / batch_id
+        if batch_dir.is_dir():
+            shutil.rmtree(batch_dir)
+        mcp_batch_dir = mcp_exports / "batches" / batch_id
+        if mcp_batch_dir.is_dir():
+            shutil.rmtree(mcp_batch_dir)
+        return store.delete_batch(batch_id)
+
+
 def batch_action(batch_id: str, action: str) -> dict:
     batch = store.get("batch", batch_id)
     if not batch:
@@ -268,7 +451,13 @@ def batch_action(batch_id: str, action: str) -> dict:
         if not failed:
             raise ValueError("没有可重跑的失败或中断任务")
         revised = create_batch(
-            {**batch, "name": batch["name"] + " · 失败项重跑", "cases": [j["case"] for j in failed]}
+            {
+                **batch,
+                "name": batch["name"] + " · 失败项重跑",
+                "cases": [j["case"] for j in failed],
+                "repeat_count": 1,
+                "request_id": None,
+            }
         )
         return revised
     else:

apps/admin_console/services/automation_store.py

@@ -12,6 +12,12 @@ import uuid
 from artemis.config import DB_PATH
 
 
+RECOVERY_ERRORS = {
+    "服务中断；未完成任务需手动重跑",
+    "服务中断，需手动重跑",
+}
+
+
 def identifier(value: str) -> str:
     if not re.fullmatch(r"[A-Za-z0-9_-]{1,100}", value):
         raise ValueError("无效的记录编号")
@@ -54,6 +60,13 @@ class AutomationStore:
 
     def put(self, kind: str, key: str, payload: dict) -> dict:
         with self.connection() as conn:
+            if kind == "batch":
+                conn.execute("BEGIN IMMEDIATE")
+                row = conn.execute(
+                    "SELECT payload FROM automation_records WHERE kind='batch' AND id=?", (key,)
+                ).fetchone()
+                if row is None or json.loads(row[0]).get("status") == "deleting":
+                    raise ValueError("批次正在删除或已删除，不能继续更新")
             conn.execute(
                 "INSERT INTO automation_records VALUES (?,?,?,?) ON CONFLICT(kind,id) DO UPDATE SET payload=excluded.payload,updated_at=excluded.updated_at",
                 (kind, key, json.dumps(payload, ensure_ascii=False), time.time()),
@@ -78,6 +91,14 @@ class AutomationStore:
 
     def save_job(self, payload: dict) -> dict:
         with self.connection() as conn:
+            conn.execute("BEGIN IMMEDIATE")
+            if batch_id := payload.get("batch_id"):
+                row = conn.execute(
+                    "SELECT payload FROM automation_records WHERE kind='batch' AND id=?",
+                    (batch_id,),
+                ).fetchone()
+                if row is None or json.loads(row[0]).get("status") == "deleting":
+                    raise ValueError("批次正在删除，不能继续派发任务")
             conn.execute(
                 "INSERT INTO automation_jobs VALUES (?,?,?,?,?,?) ON CONFLICT(session_id) DO UPDATE SET payload=excluded.payload,status=excluded.status,updated_at=excluded.updated_at",
                 (
@@ -106,9 +127,74 @@ class AutomationStore:
             )
         return [{**json.loads(row[0]), "status": row[1]} for row in rows]
 
+    def jobs_for_sessions(self, session_ids: list[str]) -> dict[str, dict]:
+        """Load only metadata belonging to the sessions shown in history."""
+        result = {}
+        with self.connection() as conn:
+            for offset in range(0, len(session_ids), 400):
+                chunk = session_ids[offset : offset + 400]
+                markers = ",".join("?" for _ in chunk)
+                rows = conn.execute(
+                    f"SELECT session_id,payload,status FROM automation_jobs WHERE session_id IN ({markers})",
+                    chunk,
+                ).fetchall()
+                result.update(
+                    {
+                        row["session_id"]: {**json.loads(row["payload"]), "status": row["status"]}
+                        for row in rows
+                    }
+                )
+        return result
+
     def preferences(self) -> dict:
         return self.get("settings", "reports") or {"formats": ["docx", "xlsx"]}
 
+    def save_case_review(self, session_id: str, status: str, reason: str) -> dict:
+        """Append reviews atomically; invalidate the batch's generated report."""
+        identifier(session_id)
+        if status not in {"passed", "failed"}:
+            raise ValueError("人工复验结果只能为通过或不通过")
+        review = {
+            "status": status,
+            "reason": reason.strip()
+            or ("人工复验确认通过。" if status == "passed" else "人工复验确认不通过。"),
+            "reviewed_at": time.time(),
+        }
+        with self.connection() as conn:
+            conn.execute("BEGIN IMMEDIATE")
+            row = conn.execute(
+                "SELECT payload FROM automation_records WHERE kind='case_review' AND id=?",
+                (session_id,),
+            ).fetchone()
+            record = json.loads(row[0]) if row else {"reviews": []}
+            record["reviews"].append(review)
+            conn.execute(
+                "INSERT INTO automation_records VALUES (?,?,?,?) ON CONFLICT(kind,id) DO UPDATE SET payload=excluded.payload,updated_at=excluded.updated_at",
+                (
+                    "case_review",
+                    session_id,
+                    json.dumps(record, ensure_ascii=False),
+                    review["reviewed_at"],
+                ),
+            )
+            job = conn.execute(
+                "SELECT batch_id FROM automation_jobs WHERE session_id=?", (session_id,)
+            ).fetchone()
+            if job and job[0]:
+                batch = conn.execute(
+                    "SELECT payload FROM automation_records WHERE kind='batch' AND id=?", (job[0],)
+                ).fetchone()
+                if not batch or json.loads(batch[0]).get("status") == "deleting":
+                    raise ValueError("批次正在删除，不能保存复验")
+                payload = json.loads(batch[0])
+                payload["report_ready"] = False
+                conn.execute(
+                    "UPDATE automation_records SET payload=?,updated_at=? WHERE kind='batch' AND id=?",
+                    (json.dumps(payload, ensure_ascii=False), review["reviewed_at"], job[0]),
+                )
+            conn.commit()
+        return review
+
     def create_batch(self, batch: dict, jobs: list[dict], payload: dict) -> dict:
         """Commit a batch and all its jobs atomically; repeated request IDs never duplicate work."""
         request_id = payload.get("request_id")
@@ -164,13 +250,75 @@ class AutomationStore:
             conn.commit()
         return batch
 
+    def begin_delete_batch(self, batch_id: str) -> list[dict] | None:
+        """Reserve an inactive batch for cleanup before any more jobs can start."""
+        with self.connection() as conn:
+            conn.execute("BEGIN IMMEDIATE")
+            row = conn.execute(
+                "SELECT payload FROM automation_records WHERE kind='batch' AND id=?",
+                (batch_id,),
+            ).fetchone()
+            if row is None:
+                return None
+            batch = json.loads(row[0])
+            active = conn.execute(
+                "SELECT 1 FROM automation_jobs WHERE batch_id=? AND status IN "
+                "('queued','running','task_paused','resuming','finalizing') LIMIT 1",
+                (batch_id,),
+            ).fetchone()
+            if batch["status"] == "running" or active:
+                raise ValueError("批次仍有任务执行中或排队中，请等待完成后删除")
+            batch["status"] = "deleting"
+            conn.execute(
+                "UPDATE automation_records SET payload=?,updated_at=? WHERE kind='batch' AND id=?",
+                (json.dumps(batch, ensure_ascii=False), time.time(), batch_id),
+            )
+            jobs = conn.execute(
+                "SELECT payload,status FROM automation_jobs WHERE batch_id=? ORDER BY position",
+                (batch_id,),
+            ).fetchall()
+            conn.commit()
+        return [{**json.loads(job[0]), "status": job[1]} for job in jobs]
+
+    def delete_batch(self, batch_id: str) -> bool:
+        """Remove a batch record and all of its jobs after cleanup."""
+        with self.connection() as conn:
+            conn.execute("BEGIN IMMEDIATE")
+            row = conn.execute(
+                "SELECT payload FROM automation_records WHERE kind='batch' AND id=?",
+                (batch_id,),
+            ).fetchone()
+            if row is None:
+                return False
+            batch = json.loads(row[0])
+            if batch.get("status") != "deleting":
+                raise ValueError("批次尚未准备删除")
+            conn.execute(
+                "DELETE FROM automation_records WHERE kind='case_review' AND id IN "
+                "(SELECT session_id FROM automation_jobs WHERE batch_id=?)",
+                (batch_id,),
+            )
+            conn.execute("DELETE FROM automation_jobs WHERE batch_id=?", (batch_id,))
+            conn.execute(
+                "DELETE FROM automation_records WHERE kind='batch' AND id=?",
+                (batch_id,),
+            )
+            # Keep the request-id tombstone so deletion cannot replay an old batch.
+            conn.commit()
+        return True
+
     def delete_history(self, session_id: str | None = None):
         with self.connection() as conn:
             if session_id:
                 conn.execute("DELETE FROM automation_jobs WHERE session_id=?", (session_id,))
+                conn.execute(
+                    "DELETE FROM automation_records WHERE kind='case_review' AND id=?",
+                    (session_id,),
+                )
             else:
                 conn.execute("DELETE FROM automation_jobs")
                 conn.execute("DELETE FROM automation_records WHERE kind='batch'")
+                conn.execute("DELETE FROM automation_records WHERE kind='case_review'")
             conn.commit()
 
 

apps/admin_console/services/batch_archive.py

@@ -0,0 +1,123 @@
+"""Background preparation and observable progress for batch report downloads."""
+
+import threading
+import time
+
+from apps.admin_console.services import report_export
+
+_lock = threading.Lock()
+_states: dict[str, dict] = {}
+
+
+def archive_path(batch_id: str):
+    return report_export.EXPORT_ROOT / "batches" / batch_id / "批次报告.zip"
+
+
+def cached(batch_id: str, batch: dict) -> bool:
+    return bool(batch.get("report_ready")) and report_export.reusable(
+        archive_path(batch_id), report_export.BATCH_ARCHIVE_SIGNATURE, "completed"
+    )
+
+
+def _snapshot(state: dict) -> dict:
+    elapsed = max(0, (state.get("finished_at") or time.monotonic()) - state["started_at"])
+    completed, total = state["completed"], state["total"]
+    eta = (
+        max(0, round(elapsed / completed * (total - completed)))
+        if completed and total > completed
+        else None
+    )
+    return {
+        "status": state["status"],
+        "stage": state["stage"],
+        "completed": completed,
+        "total": total,
+        "elapsed_seconds": round(elapsed),
+        "eta_seconds": eta,
+        "error": state.get("error"),
+    }
+
+
+def status(batch_id: str, batch: dict) -> dict:
+    with _lock:
+        state = _states.get(batch_id)
+        if state:
+            return _snapshot(state)
+    if cached(batch_id, batch):
+        return {
+            "status": "ready",
+            "stage": "报告包已就绪",
+            "completed": 0,
+            "total": 0,
+            "elapsed_seconds": 0,
+            "eta_seconds": None,
+            "error": None,
+        }
+    return {
+        "status": "idle",
+        "stage": "尚未开始打包",
+        "completed": 0,
+        "total": 0,
+        "elapsed_seconds": 0,
+        "eta_seconds": None,
+        "error": None,
+    }
+
+
+def prepare(batch_id: str, batch: dict) -> dict:
+    with _lock:
+        current = _states.get(batch_id)
+        if current and current["status"] == "preparing":
+            return _snapshot(current)
+        if cached(batch_id, batch):
+            _states.pop(batch_id, None)
+            return status_from_cache()
+        state = {
+            "status": "preparing",
+            "stage": "等待打包",
+            "completed": 0,
+            "total": len(report_export.store.jobs(batch_id)),
+            "started_at": time.monotonic(),
+        }
+        _states[batch_id] = state
+
+    def update(completed: int, total: int, stage: str) -> None:
+        with _lock:
+            state.update(completed=completed, total=total, stage=stage)
+
+    def run() -> None:
+        try:
+            with report_export.export_lock("batch:" + batch_id):
+                latest = report_export.store.get("batch", batch_id)
+                if not (latest and cached(batch_id, latest)):
+                    report_export.export_batch(batch_id, progress=update)
+            with _lock:
+                state.update(status="ready", stage="报告包已就绪", finished_at=time.monotonic())
+        except Exception as exc:
+            with _lock:
+                state.update(
+                    status="failed",
+                    stage="打包失败",
+                    error=str(exc)[:500],
+                    finished_at=time.monotonic(),
+                )
+
+    threading.Thread(target=run, name=f"batch-archive-{batch_id[:8]}", daemon=True).start()
+    return status(batch_id, batch)
+
+
+def status_from_cache() -> dict:
+    return {
+        "status": "ready",
+        "stage": "报告包已就绪",
+        "completed": 0,
+        "total": 0,
+        "elapsed_seconds": 0,
+        "eta_seconds": None,
+        "error": None,
+    }
+
+
+def forget(batch_id: str) -> None:
+    with _lock:
+        _states.pop(batch_id, None)

apps/admin_console/services/case_review.py

@@ -0,0 +1,40 @@
+"""Human review overlays the AI verdict without changing execution evidence."""
+
+from pathlib import Path
+
+from apps.admin_console.services.automation_store import AutomationStore
+from artemis.runtime.case_result import read_case_result
+
+
+def reviewed_result(
+    session_id: str, ai_result: dict | None, storage: AutomationStore
+) -> dict | None:
+    original = {**ai_result, "source": "ai"} if ai_result else None
+    record = storage.get("case_review", session_id)
+    reviews = (record or {}).get("reviews") or []
+    if not reviews:
+        return original
+    latest = reviews[-1]
+    return {
+        "status": latest["status"],
+        "reason": latest["reason"],
+        "evidence": [],
+        "source": "human",
+        "reviewed_at": latest["reviewed_at"],
+        "ai_result": original,
+    }
+
+
+def read_reviewed_result(
+    session_id: str,
+    execution_status: str,
+    error: str = "",
+    *,
+    storage: AutomationStore,
+    root: Path | None = None,
+) -> dict | None:
+    return reviewed_result(
+        session_id,
+        read_case_result(session_id, execution_status, error, root=root),
+        storage,
+    )

apps/admin_console/services/evidence_service.py

@@ -0,0 +1,160 @@
+"""Read task-owned plugin evidence without executing tools or converting media."""
+
+import json
+import math
+from pathlib import Path
+import re
+
+from artemis.config import TRACES_PATH
+
+
+def task_root(session_id: str) -> Path:
+    if not re.fullmatch(r"[A-Za-z0-9_-]{1,100}", session_id):
+        raise ValueError("无效任务编号")
+    root = TRACES_PATH.resolve() / session_id
+    if root.resolve() != root:
+        raise ValueError("任务证据目录不能是符号链接")
+    return root
+
+
+def evidence_file(root: Path, relative: str) -> Path | None:
+    path = root / relative
+    try:
+        if (
+            path.is_file()
+            and path.resolve() == path.absolute()
+            and path.resolve().is_relative_to(root.resolve())
+        ):
+            return path
+    except OSError:
+        return None
+    return None
+
+
+def plugin_sources(session_id: str) -> list[dict]:
+    root = task_root(session_id)
+    path = evidence_file(root, "evidence/recordings/recordings.json")
+    if not path or path.stat().st_size > 4 * 1024 * 1024:
+        return []
+    try:
+        data = json.loads(path.read_text(encoding="utf-8"))
+        if data.get("version") != 1 or not isinstance(data.get("sources"), list):
+            return []
+        from apps.admin_console.services.media_service import MediaService
+
+        sources = []
+        for item in data["sources"]:
+            if not isinstance(item, dict) or not isinstance(item.get("id"), str):
+                continue
+            source = {**item, "segments": []}
+            for segment in item.get("segments", []):
+                if not isinstance(segment, dict):
+                    continue
+                video = evidence_file(root, str(segment.get("file") or ""))
+                if video is None or video.suffix.lower() != ".mp4":
+                    continue
+                duration = float(segment.get("duration", 0))
+                start = float(segment.get("offset_ms", 0)) / 1000
+                if (
+                    not math.isfinite(duration)
+                    or duration <= 0
+                    or not math.isfinite(start)
+                    or start < 0
+                ):
+                    continue
+                source["segments"].append(
+                    {
+                        **segment,
+                        "url": MediaService.path_to_video_url(video),
+                        "start": start,
+                        "duration": duration,
+                    }
+                )
+            sources.append(source)
+        return sources
+    except (OSError, ValueError, TypeError, KeyError):
+        return []
+
+
+def plugin_files(session_id: str) -> list[tuple[Path, str]]:
+    from apps.admin_console.services.exploration_evidence import demonstration_files
+
+    root = task_root(session_id)
+    files = dict(demonstration_files(session_id))
+    for source in plugin_sources(session_id):
+        for segment in source["segments"]:
+            path = evidence_file(root, segment["file"])
+            if path:
+                files[path] = "任务录屏/电脑/" + path.name
+    for directory in (root / "artifacts" / "plugins", root / "evidence"):
+        if not directory.is_dir():
+            continue
+        for path in directory.rglob("*"):
+            if path.suffix.lower() == ".tmp" or (
+                directory.name == "evidence" and path.suffix.lower() in {".mp4", ".mkv"}
+            ):
+                continue
+            relative = path.relative_to(root).as_posix()
+            checked = evidence_file(root, relative)
+            if checked:
+                files[checked] = (
+                    "任务文件/插件/" + path.relative_to(root / "artifacts" / "plugins").as_posix()
+                    if directory.name == "plugins"
+                    else "证据索引/" + path.relative_to(root / "evidence").as_posix()
+                )
+    return list(files.items())
+
+
+def export_index(
+    session_id: str, report: dict, android_videos: list, *, images=None, artifacts=None, extras=None
+) -> dict:
+    from apps.admin_console.services.report_export import session_artifacts, session_screenshots
+    from apps.admin_console.services.exploration_evidence import (
+        exploration_index,
+        demonstration_sources,
+    )
+
+    extras = plugin_files(session_id) if extras is None else extras
+    recordings = plugin_sources(session_id) + demonstration_sources(session_id)
+    root = task_root(session_id)
+    paths = {path.resolve(): name for path, name in extras}
+    for source in recordings:
+        for segment in source["segments"]:
+            segment["file"] = paths.get((root / segment["file"]).resolve())
+            segment.pop("url", None)
+    images = session_screenshots(report) if images is None else images
+    artifacts = session_artifacts(report) if artifacts is None else artifacts
+    files = [(path, "步骤截图/" + name) for path, name in images]
+    files += [(path, "任务录屏/安卓/" + name) for path, name in android_videos]
+    files += [(path, "任务文件/安卓/" + name) for path, name in artifacts]
+    files += extras
+    calls = []
+    ledger = evidence_file(root, "evidence/calls.jsonl")
+    if ledger:
+        with ledger.open(encoding="utf-8") as handle:
+            for line in handle:
+                try:
+                    call = json.loads(line)
+                    if not isinstance(call, dict):
+                        continue
+                    for attachment in call.get("attachments", []):
+                        relative = str(attachment.get("path") or "")
+                        attachment["file"] = paths.get((root / relative).resolve())
+                    calls.append(call)
+                except (ValueError, TypeError, AttributeError):
+                    continue  # A partial final line after a crash is not a valid receipt.
+    return {
+        "version": 1,
+        "session_id": session_id,
+        "exploration": exploration_index(session_id, paths),
+        "name": report["name"],
+        "status": report["status"],
+        "verdict": report["verdict"],
+        "steps": report.get("steps", []),
+        "calls": calls,
+        "recordings": recordings,
+        "android_videos": ["任务录屏/安卓/" + name for _, name in android_videos],
+        "image_references": {path.stem: "步骤截图/" + name for path, name in images},
+        "files": [{"path": name, "bytes": path.stat().st_size} for path, name in files],
+        "note": "工具执行完成不等于用例通过。录像保留真实时间与缺口；各来源截图不能相互替代。",
+    }

apps/admin_console/services/execution_config.py

@@ -0,0 +1,54 @@
+"""Credential-free execution settings persisted for repeatable favorites."""
+
+from apps.admin_console.services.task_import import TaskOptions
+from artemis.config import load_agent_config, parse_llm_config, verification_level_for_checker
+from artemis.config.settings import settings
+
+
+def execution_config(item: dict) -> dict:
+    agent = load_agent_config()
+    model = parse_llm_config().operator
+    overrides = item.get("runtime_overrides") or {}
+    profile = overrides.get("profile") or item.get("profile") or "flash"
+    options = {
+        "profile": profile,
+        "model": model.model,
+        "reasoning_effort": model.reasoning_effort or "default",
+        "request_timeout_seconds": model.timeout,
+        "max_retries": model.max_retries,
+        "retry_interval_seconds": settings.OPENAI_RETRY_INTERVAL_SECONDS,
+        "record_video": settings.ARTEMIS_RECORDING_ENABLED,
+        "recording": agent.recording.model_dump(),
+        "analyze_video": agent.video_analyzer.enable_ledger,
+        "timeout_seconds": settings.ARTEMIS_TASK_TIMEOUT_SECONDS,
+        "verification_level": verification_level_for_checker(agent.checker),
+        "explorer_mode": agent.explorer.resolve(profile=profile),
+        "observation_timeout_seconds": settings.ARTEMIS_OBSERVATION_TIMEOUT_SECONDS,
+        "observation_max_attempts": settings.ARTEMIS_OBSERVATION_MAX_ATTEMPTS,
+        "device_command_timeout_seconds": settings.ARTEMIS_DEVICE_COMMAND_TIMEOUT_SECONDS,
+        "action_failure_attempts": settings.ARTEMIS_ACTION_FAILURE_ATTEMPTS,
+        "worker_cleanup_timeout_seconds": settings.ARTEMIS_WORKER_CLEANUP_TIMEOUT_SECONDS,
+    }
+    for key in ("device_serial", "verification_level", "explorer_mode", "locked_app_package"):
+        if item.get(key) is not None:
+            options[key] = item[key]
+    options["recording"] = {**options["recording"], **(overrides.get("recording") or {})}
+    options.update({k: v for k, v in overrides.items() if k != "recording" and v is not None})
+    return {
+        "runtime_overrides": TaskOptions.model_validate(options).model_dump(exclude_none=True),
+        "expected_output": item.get("expected_output"),
+        "enable_outputter": item.get("enable_outputter")
+        if item.get("enable_outputter") is not None
+        else agent.outputter.enabled,
+        "app_path": item.get("app_path"),
+    }
+
+
+def persist_execution_config(item: dict) -> None:
+    from apps.admin_console.services.automation_store import store
+
+    snapshot = execution_config(item)
+    item["execution_config"] = snapshot
+    if job := store.job(str(item["session_id"])):
+        job["execution_config"] = snapshot
+        store.save_job(job)

apps/admin_console/services/exploration_deletion.py

@@ -0,0 +1,48 @@
+"""Delete an inactive conversation while keeping independently executed reruns."""
+
+import shutil
+from uuid import UUID
+
+from apps.admin_console.core.state import state
+from apps.admin_console.services import report_export
+from artemis.data_engine.storage import StorageManager
+from artemis.exploration.store import ExplorationConflict, ExplorationStore
+from artemis.runtime.process_probe import pid_is_alive
+from artemis.runtime.task_control import read_state
+
+
+def _check_runs(runs: list[dict]) -> None:
+    queued = {
+        item.get("session_id")
+        for item in state.queue_items
+        if item.get("status") in {"pending", "queued", "running"}
+    }
+    for run in runs:
+        run_id = run["id"]
+        if (
+            run_id in state.active_runs
+            or run_id in queued
+            or pid_is_alive(read_state(run_id).get("pid"))
+        ):
+            raise ExplorationConflict("探索工作进程尚未结束，请停止任务并稍后重试删除")
+
+
+def delete_exploration(repository: ExplorationStore, exploration_id: str) -> list[str]:
+    runs = repository.begin_delete(exploration_id, _check_runs)
+    # Validation sessions and batch copies remain ordinary Flash/Pro history.
+    deleted = [run["id"] for run in runs if run["kind"] == "exploration"]
+    storage = StorageManager(repository.database.path, report_export.TRACES_PATH)
+    with report_export.export_lock(*("session:" + run_id for run_id in deleted)):
+        for run_id in deleted:
+            safe_id = str(UUID(run_id))
+            storage.delete_session(UUID(safe_id))
+            repository.database.delete_history(safe_id)
+            for root in (
+                report_export.EXPORT_ROOT,
+                report_export.EXPORT_ROOT.parent / "mcp_exports",
+            ):
+                folder = (root / "sessions" / safe_id).resolve()
+                if folder.is_relative_to(root.resolve()) and folder.is_dir():
+                    shutil.rmtree(folder)
+        repository.finish_delete(exploration_id)
+    return deleted

apps/admin_console/services/exploration_evidence.py

@@ -0,0 +1,110 @@
+"""Read-only exploration provenance for reports, archives and source-aware video."""
+
+import json
+
+from artemis.exploration.store import explorations
+from artemis.exploration.demonstrations import Demonstrations
+from apps.admin_console.services.evidence_service import evidence_file, task_root
+
+
+def exploration_metadata(run_id: str) -> dict | None:
+    run = explorations.run(run_id)
+    if not run:
+        return None
+    return {
+        "id": run["exploration_id"],
+        "kind": run["kind"],
+        "case_version": run.get("case_version"),
+    }
+
+
+def demonstration_sources(run_id: str) -> list[dict]:
+    root = task_root(run_id)
+    directory = root / "evidence" / "demonstrations"
+    from apps.admin_console.services.media_service import MediaService
+
+    sources = []
+    for manifest in directory.glob("*/demonstration.json"):
+        if (
+            not evidence_file(root, manifest.relative_to(root).as_posix())
+            or manifest.stat().st_size > 4 * 1024 * 1024
+        ):
+            continue
+        try:
+            record = json.loads(manifest.read_text(encoding="utf-8"))
+            for track in record["media"]:
+                segments = []
+                for item in track["segments"]:
+                    path = evidence_file(root, item["file"])
+                    if path and path.suffix.lower() == ".mp4" and item["duration_ms"] > 0:
+                        segments.append(
+                            {
+                                **item,
+                                "start": item["offset_ms"] / 1000,
+                                "duration": item["duration_ms"] / 1000,
+                                "url": MediaService.path_to_video_url(path),
+                            }
+                        )
+                if segments:
+                    sources.append(
+                        {
+                            **track,
+                            "segments": segments,
+                            "label": "人工示范 · "
+                            + ("安卓" if track["kind"] == "android" else "电脑"),
+                            "status": "ready",
+                            "message": "人工操作录像，不等于自动复跑通过",
+                            "demonstration_id": record["id"],
+                        }
+                    )
+        except (OSError, ValueError, KeyError, TypeError):
+            continue
+    return sources
+
+
+def demonstration_files(run_id: str) -> list[tuple]:
+    root = task_root(run_id)
+    directory = root / "evidence" / "demonstrations"
+    files = []
+    for path in directory.rglob("*"):
+        if path.suffix.lower() not in {".mp4", ".mkv", ".webm"}:
+            continue
+        checked = evidence_file(root, path.relative_to(root).as_posix())
+        if checked:
+            files.append((checked, "任务录屏/人工示范/" + path.relative_to(directory).as_posix()))
+    return files
+
+
+def exploration_index(run_id: str, archive_paths: dict | None = None) -> dict | None:
+    run = explorations.run(run_id)
+    if not run:
+        return None
+    sid = run["exploration_id"]
+    session = explorations.get(sid)
+    messages = []
+    after = 0
+    while page := explorations.messages(sid, after=after):
+        messages.extend(page)
+        after = page[-1]["sequence"]
+    demos = Demonstrations(explorations).list(sid)
+    root = task_root(run_id)
+    for demo in demos:
+        for track in demo.get("media", []):
+            for segment in track.get("segments", []):
+                relative = segment.pop("file", "")
+                segment["trace_file"] = relative
+                segment["file"] = (
+                    (archive_paths or {}).get((root / relative).resolve())
+                    if demo["run_id"] == run_id
+                    else None
+                )
+    return {
+        "session": session,
+        "messages": messages,
+        "runs": explorations.runs(sid),
+        "cases": [
+            explorations.case(sid, version) for version in range(1, session["case_version"] + 1)
+        ],
+        "demonstrations": demos,
+        "note": "此导出保留会话、用例版本和各次运行的关联；只有当前任务的媒体文件在此目录中。",
+    }

apps/admin_console/services/exploration_listing.py

@@ -0,0 +1,45 @@
+"""Shared exploration metadata for queue cards and history pagination."""
+
+# Rank execution runs independently from validations: a rerun never replaces the
+# conversation card. Keep all raw sessions accessible for evidence inspection.
+EXPLORATION_RUNS_CTE = """
+exploration_runs AS (
+ SELECT r.id AS run_id,
+   json_extract(r.payload,'$.exploration_id') AS exploration_id,
+   json_extract(r.payload,'$.kind') AS exploration_kind,
+   json_extract(r.payload,'$.case_version') AS case_version,
+   json_extract(r.payload,'$.task_config.model') AS exploration_model,
+   json_extract(r.payload,'$.task_config.profile') AS exploration_profile,
+   json_extract(v.payload,'$.name') AS exploration_case_name,
+   CASE WHEN json_extract(r.payload,'$.kind')='exploration' AND
+     (json_extract(c.payload,'$.active_run_id') IS NULL
+     OR json_extract(c.payload,'$.active_run_id') = r.id)
+     THEN json_extract(c.payload,'$.status')
+     ELSE json_extract(r.payload,'$.status') END AS exploration_status,
+   FIRST_VALUE(r.id) OVER (
+     PARTITION BY json_extract(r.payload,'$.exploration_id'),json_extract(r.payload,'$.kind')
+     ORDER BY json_extract(r.payload,'$.created_at') DESC, r.id DESC
+   ) AS exploration_latest_run_id,
+   COUNT(*) OVER (
+     PARTITION BY json_extract(r.payload,'$.exploration_id'),json_extract(r.payload,'$.kind')
+   ) AS exploration_run_count
+ FROM automation_records r
+ LEFT JOIN automation_records c ON c.kind='exploration'
+   AND c.id=json_extract(r.payload,'$.exploration_id')
+ LEFT JOIN automation_records v ON v.kind='exploration_case'
+   AND v.id=json_extract(r.payload,'$.exploration_id') || '_' ||
+     printf('%010d',json_extract(r.payload,'$.case_version'))
+ WHERE r.kind='exploration_run'
+)
+"""
+
+
+def exploration_metadata(connection) -> dict[str, dict]:
+    rows = connection.execute("WITH " + EXPLORATION_RUNS_CTE + " SELECT * FROM exploration_runs")
+    result = {}
+    for row in rows:
+        item = {key: row[key] for key in row.keys() if key != "run_id"}
+        if item.get("exploration_kind") == "validation" and item.get("exploration_case_name"):
+            item["display_name"] = item["exploration_case_name"]
+        result[row["run_id"]] = item
+    return result

apps/admin_console/services/exploration_recovery.py

@@ -0,0 +1,95 @@
+"""Reconcile saved exploration metadata with the existing queue and worker owners."""
+
+import json
+import time
+
+from apps.admin_console.core.state import state
+from apps.admin_console.database.repositories.session_repository import SessionRepository
+from artemis.exploration.store import ExplorationConflict, ExplorationStore, explorations
+from artemis.runtime.process_probe import pid_is_alive
+from artemis.runtime.task_control import read_state
+from artemis.runtime.task_result import read_worker_outcome
+from artemis.runtime.case_result import read_case_result
+
+
+def reconcile_explorations(repository: ExplorationStore = explorations) -> None:
+    for kind in ("exploration_generation", "exploration_discussion"):
+        with repository._transaction() as connection:
+            for row in connection.execute(
+                "SELECT id, payload FROM automation_records WHERE kind=? AND json_extract(payload,'$.status')='running'",
+                (kind,),
+            ).fetchall():
+                job = json.loads(row[1])
+                if (
+                    not pid_is_alive(job.get("owner_pid"))
+                    or job.get("lease_until", 0) < time.time()
+                ):
+                    repository._put(
+                        connection,
+                        kind,
+                        row[0],
+                        {
+                            **job,
+                            "status": "failed",
+                            "lease_until": 0,
+                            "error": "生成进程已中断；已有结果保留，可重试。",
+                        },
+                    )
+    with repository.database.connection() as connection:
+        rows = connection.execute(
+            "SELECT payload FROM automation_records WHERE kind='exploration' "
+            "AND json_extract(payload,'$.active_run_id') IS NOT NULL"
+        ).fetchall()
+    history = SessionRepository(repository.database.path)
+    queued = {
+        str(item.get("session_id"))
+        for item in state.queue_items
+        if isinstance(item, dict) and item.get("status") in {"pending", "running", "queued"}
+    }
+    for row in rows:
+        session = json.loads(row[0])
+        run_id = session["active_run_id"]
+        run = repository.run(run_id)
+        if not run:
+            continue
+        outcome = read_worker_outcome(run_id)
+        if outcome:
+            repository.finish_run(run_id, {**outcome, "status": outcome["task_status"]})
+            continue
+        control = read_state(run_id)
+        if run["kind"] == "validation" and pid_is_alive(control.get("pid")):
+            observed = {"running": "running", "task_paused": "paused", "resuming": "queued"}.get(
+                control.get("status")
+            )
+            if observed and observed != run["status"]:
+                try:
+                    repository.set_run_state(run_id, observed, reason=control.get("pause_reason"))
+                except ExplorationConflict:
+                    continue  # The owner finalized while this tick read its control file.
+        if run_id in queued or run_id in state.active_runs:
+            continue
+        if control.get("pid") and pid_is_alive(control["pid"]):
+            continue
+        execution = history.get_session_by_id(run_id) or {}
+        if execution.get("pid") and history.process_is_alive(execution["pid"]):
+            continue
+        job = repository.database.job(run_id) or {}
+        terminal = execution.get("status") or job.get("terminal_status") or job.get("status")
+        if terminal in {"completed", "failed", "cancelled", "interrupted", "skipped"}:
+            repository.finish_run(
+                run_id,
+                {
+                    "status": terminal,
+                    "error": execution.get("error") or job.get("error") or "",
+                    "test_result": execution.get("test_result")
+                    or read_case_result(run_id, terminal),
+                },
+            )
+        elif time.time() - run["created_at"] > 300:
+            repository.finish_run(
+                run_id,
+                {
+                    "status": "interrupted",
+                    "error": "探索工作进程已结束；原证据已保留，可继续探索并重新观察当前设备。",
+                },
+            )

apps/admin_console/services/exploration_service.py

@@ -0,0 +1,334 @@
+"""Web and MCP use the same exploration admission and conversation service."""
+
+from artemis.config.agent import load_agent_config
+from artemis.config.exploration import ExplorationConfig
+from artemis.exploration.case_prompt import require_resolved_case
+from artemis.exploration.models import CreateExploration, SaveCaseRequest, ValidateCaseRequest
+from artemis.exploration.demonstrations import Demonstrations
+from artemis.exploration.store import ExplorationConflict, ExplorationStore, explorations
+from artemis.runtime.task_control import add_instruction, read_state, request_pause, request_resume
+from apps.admin_console.services.task_import import TaskOptions
+
+
+def exploration_case_prompt(case: dict) -> str:
+    return require_resolved_case(case)
+
+
+class ExplorationService:
+    def __init__(self, repository: ExplorationStore | None = None):
+        self.repository = repository if repository is not None else explorations
+
+    def settings(self) -> dict:
+        saved = self.repository.database.get("settings", "exploration") or {}
+        return ExplorationConfig.model_validate(
+            {
+                **load_agent_config().exploration.model_dump(),
+                **saved,
+            }
+        ).model_dump()
+
+    def save_settings(self, value: ExplorationConfig) -> dict:
+        return self.repository.database.put("settings", "exploration", value.model_dump())
+
+    def create(self, value: CreateExploration, *, source: str = "web") -> dict:
+        options = TaskOptions.model_validate(value.task_config).model_dump(exclude_none=True)
+        return self.repository.create(
+            value.goal,
+            value.config or ExplorationConfig.model_validate(self.settings()),
+            options,
+            value.request_id,
+            source=source,
+        )
+
+    def detail(self, session_id: str, *, after: int = 0) -> dict:
+        session = self.repository.get(session_id)
+        runs = self.repository.runs(session_id)
+        version = self.repository.case(session_id)
+        with self.repository.database.connection() as connection:
+            import json
+
+            generations = [
+                json.loads(row[0])
+                for row in connection.execute(
+                    "SELECT payload FROM automation_records WHERE kind='exploration_generation' "
+                    "AND json_extract(payload,'$.exploration_id')=? ORDER BY updated_at DESC LIMIT 20",
+                    (session_id,),
+                ).fetchall()
+            ]
+        from artemis.runtime.case_result import resolve_case_result
+
+        for run in runs:
+            outcome = run.get("outcome")
+            verdict = resolve_case_result(outcome, run["status"]) if outcome else None
+            run["test_result"] = verdict.model_dump() if verdict else None
+        return {
+            **session,
+            "messages": self.repository.messages(session_id, after=after),
+            "runs": runs,
+            "case": version,
+            "generations": [
+                {k: v for k, v in g.items() if k not in {"owner", "owner_pid", "lease_until"}}
+                for g in generations
+            ],
+            "demonstrations": Demonstrations(self.repository).list(session_id),
+            "validation_runs": [
+                r
+                for r in runs
+                if r["kind"] == "validation" and r["case_version"] == session["case_version"]
+            ],
+            "control": read_state(session["active_run_id"]) if session["active_run_id"] else {},
+        }
+
+    async def _admit(self, run: dict, *, source: str) -> dict:
+        # RunRequest performs the same readiness and idempotency checks as /api/run.
+        from apps.admin_console.routers.tasks import run_task
+        from apps.admin_console.schemas.task_schema import RunRequest
+        from fastapi import HTTPException
+
+        current = self.repository.run(run["id"])
+        if current and current.get("outcome") is not None:
+            return current
+        options = TaskOptions.model_validate(run["task_config"])
+        try:
+            result = await run_task(
+                RunRequest(
+                    goal=run["goal"],
+                    profile=options.profile or "pro",
+                    session_id=run["id"],
+                    ingress=source,
+                    conversation_id=run["exploration_id"],
+                    runtime_overrides=options,
+                )
+            )
+        except HTTPException as exc:
+            if exc.status_code < 500:
+                self.repository.finish_run(
+                    run["id"], {"status": "rejected", "error": str(exc.detail)}
+                )
+            raise
+        if result.get("status") in {"error", "rejected"}:
+            self.repository.finish_run(
+                run["id"],
+                {
+                    "status": "rejected",
+                    "error": result.get("error") or result.get("message") or "任务入队失败",
+                },
+            )
+        else:
+            self.repository.set_run_state(run["id"], "queued", expected_status="reserved")
+        return {"run": self.repository.run(run["id"]), "admission": result}
+
+    async def start(self, session_id: str, request_id: str, *, source: str = "web") -> dict:
+        session = self.repository.get(session_id)
+        config = ExplorationConfig.model_validate(session["config"])
+        options = {
+            **session["task_config"],
+            "profile": "pro",
+            "model": config.model,
+            "reasoning_effort": config.reasoning_effort,
+            "timeout_seconds": config.timeout_seconds,
+            "verification_level": "final",
+        }
+        run = self.repository.reserve_run(
+            session_id,
+            request_id,
+            kind="exploration",
+            goal=session["goal"],
+            task_config=options,
+        )
+        return await self._admit(run, source=source)
+
+    async def validate(
+        self, session_id: str, value: ValidateCaseRequest, *, source: str = "web"
+    ) -> dict:
+        case = self.repository.case(session_id, value.version)
+        if case is None:
+            raise KeyError("用例版本不存在")
+        session = self.repository.get(session_id)
+        # Device/plugin binding may be reused; exploration's expensive model is not.
+        options = {
+            k: v
+            for k, v in session["task_config"].items()
+            if k not in {"model", "reasoning_effort", "profile", "timeout_seconds"}
+        }
+        options.update(TaskOptions.model_validate(value.task_config).model_dump(exclude_none=True))
+        options["profile"] = value.profile
+        run = self.repository.reserve_run(
+            session_id,
+            value.request_id,
+            kind="validation",
+            goal=exploration_case_prompt(case),
+            task_config=options,
+            case_version=value.version,
+        )
+        return await self._admit(run, source=source)
+
+    def message(
+        self,
+        session_id: str,
+        text: str,
+        request_id: str,
+        *,
+        source: str = "web",
+        expected_run_id: str | None = None,
+    ) -> dict:
+        session = self.repository.get(session_id)
+        if expected_run_id and session["active_run_id"] != expected_run_id:
+            raise ExplorationConflict("所选执行已结束，指导未发送，请刷新任务状态")
+        question = session.get("pending_question")
+        message = self.repository.add_message(
+            session_id,
+            text,
+            request_id,
+            source=source,
+            metadata={"question_id": question, "run_id": session["active_run_id"]},
+        )
+        question = message.get("question_id")
+        run_id = message.get("run_id")
+        if run_id and run_id == session["active_run_id"]:
+            run = self.repository.run(run_id)
+            if run and run.get("outcome") is None:
+                add_instruction(
+                    run_id, message["text"], source=source, instruction_id=message["id"]
+                )
+                self.repository.answer_question(session_id, question)
+                if read_state(run_id).get("status") == "task_paused":
+                    request_resume(run_id)
+        return message
+
+    def pause(self, session_id: str) -> dict:
+        session = self.repository.get(session_id)
+        if session["status"] == "finalizing":
+            raise ExplorationConflict("设备操作已结束，正在提炼用例，无需暂停。")
+        if not session["active_run_id"]:
+            raise ExplorationConflict("当前没有正在执行的探索")
+        request_pause(session["active_run_id"], reason="user")
+        return {"status": "pausing", "id": session_id}
+
+    def stop(self, session_id: str) -> dict:
+        from apps.admin_console.services.task_queue_service import TaskQueueService
+
+        session = self.repository.get(session_id)
+        if run_id := session["active_run_id"]:
+            requested = TaskQueueService.stop_tasks(session_id=run_id)
+            return {"status": "stopping" if requested else "not_running", "id": session_id}
+        return {"status": session["status"], "id": session_id}
+
+    async def resume(self, session_id: str, request_id: str, *, source: str = "web") -> dict:
+        repo = self.repository
+        with repo._transaction() as connection:
+            session = repo._require(connection, session_id)
+
+            def prepare():
+                if session.get("pending_question"):
+                    raise ExplorationConflict("请先回答探索助手的问题，再继续执行。")
+                if run_id := session["active_run_id"]:
+                    if read_state(run_id).get("status") != "task_paused":
+                        raise ExplorationConflict("请等待探索安全暂停后再继续")
+                    request_resume(run_id)
+                    return {"status": "resuming", "id": session_id, "run_id": run_id}
+                return {"status": "start", "id": session_id}
+
+            result = repo._operation(connection, f"{session_id}:continue", request_id, {}, prepare)
+        if result["status"] == "resuming":
+            return result
+        return await self.start(session_id, request_id, source=source)
+
+    def save_case(self, session_id: str, value: SaveCaseRequest) -> dict:
+        return self.repository.save_case(
+            session_id, value.case, value.expected_version, value.request_id
+        )
+
+    def demonstrate(self, session_id: str, request_id: str, source: str | None = None) -> dict:
+        session = self.repository.get(session_id)
+        if session["status"] == "finalizing":
+            raise ExplorationConflict("正在提炼用例，请完成后继续探索再示范。")
+        source = source or (
+            "computer"
+            if session["task_config"].get("execution_target") == "computer"
+            else "android"
+        )
+        if source not in {"android", "computer", "both"}:
+            raise ValueError("无效的示范来源")
+        demo = Demonstrations(self.repository).request(session_id, source, request_id)
+        if demo["status"] == "requested":
+            request_pause(demo["run_id"], reason="demonstration")
+            if read_state(demo["run_id"]).get("status") == "task_paused":
+                request_resume(demo["run_id"])
+        return demo
+
+    def demonstration_command(
+        self, session_id: str, demo_id: str, command: str, request_id: str, *, seconds: int = 0
+    ) -> dict:
+        return Demonstrations(self.repository).command(
+            session_id, demo_id, command, request_id, seconds=seconds
+        )
+
+    async def generate_case(self, session_id: str, value) -> dict:
+        from artemis.exploration.synthesis import generate_case, _generation
+        from apps.admin_console.services.automation_service import spawn
+
+        generation = _generation(
+            self.repository,
+            session_id,
+            value.request_id,
+            value.source_run_id,
+            value.expected_version,
+        )
+        if generation["status"] == "completed":
+            return {"status": "completed", "id": generation["id"], "case": generation["case"]}
+
+        async def work():
+            try:
+                await generate_case(
+                    self.repository,
+                    session_id,
+                    value.request_id,
+                    value.source_run_id,
+                    value.expected_version,
+                    generation=generation,
+                )
+            except Exception:
+                import logging
+
+                logging.getLogger(__name__).exception("Exploration synthesis request failed")
+
+        spawn(work())
+        return {"status": "running", "id": generation["id"]}
+
+    def create_case_batch(self, session_id: str, value) -> dict:
+        from apps.admin_console.routers.automation import BatchRequest
+        from apps.admin_console.services.automation_service import create_batch
+        from artemis.exploration.store import _digest
+
+        case = self.repository.case(session_id, value.version)
+        if not case:
+            raise KeyError("用例版本不存在")
+        session = self.repository.get(session_id)
+        options = {
+            k: v
+            for k, v in session["task_config"].items()
+            if k not in {"model", "reasoning_effort", "profile", "timeout_seconds"}
+        }
+        options.update(TaskOptions.model_validate(value.task_config).model_dump(exclude_none=True))
+        options["profile"] = value.profile
+        payload = BatchRequest(
+            name=value.name,
+            profile=value.profile,
+            request_id="explore-" + _digest([session_id, value.request_id]),
+            cases=[
+                {
+                    "name": case["name"],
+                    "goal": exploration_case_prompt(case),
+                    "preconditions": case["preconditions"],
+                    "case_id": f"{session_id}-v{value.version}",
+                    "task_config": options,
+                    "category": "探索用例",
+                    "subcategory": f"版本{value.version}",
+                }
+            ],
+        ).model_dump()
+        return create_batch(payload)
+
+
+exploration_service = ExplorationService()

apps/admin_console/services/export_cache.py

@@ -0,0 +1,66 @@
+"""Per-object export coordination and content-aware artifact reuse."""
+
+from contextlib import contextmanager, ExitStack
+import hashlib
+import json
+from pathlib import Path
+import threading
+from weakref import WeakValueDictionary
+
+_guard = threading.Lock()
+_locks: WeakValueDictionary = WeakValueDictionary()
+TERMINAL = {"completed", "success", "failed", "cancelled", "interrupted", "skipped"}
+
+
+@contextmanager
+def export_lock(*keys: str):
+    # Retain strong references until every lock is released. Unrelated exports
+    # run concurrently; nested session exports can reacquire their own RLock.
+    with _guard:
+        locks = []
+        for key in sorted(set(keys)):
+            lock = _locks.get(key)
+            if lock is None:
+                lock = threading.RLock()
+                _locks[key] = lock
+            locks.append(lock)
+    with ExitStack() as stack:
+        for lock in locks:
+            stack.enter_context(lock)
+        yield
+
+
+def fingerprint(report: dict, files=(), extra=None) -> str:
+    metadata = []
+    for path, name in files:
+        stat = path.stat()
+        metadata.append((name, str(path), stat.st_size, stat.st_mtime_ns))
+    payload = json.dumps([2, report, metadata, extra], sort_keys=True, default=str)
+    return hashlib.sha256(payload.encode()).hexdigest()
+
+
+def reusable(path: Path, signature: str, status: str) -> bool:
+    if status not in TERMINAL or not path.is_file():
+        return False
+    try:
+        return path.with_suffix(path.suffix + ".fingerprint").read_text() == signature
+    except OSError:
+        return False
+
+
+def remember(path: Path, signature: str) -> None:
+    marker = path.with_suffix(path.suffix + ".fingerprint")
+    temporary = marker.with_suffix(".tmp")
+    temporary.write_text(signature, encoding="ascii")
+    temporary.replace(marker)
+
+
+def compression(path: Path) -> int:
+    import zipfile
+
+    return (
+        zipfile.ZIP_STORED
+        if path.suffix.lower()
+        in {".mp4", ".mkv", ".webm", ".docx", ".xlsx", ".png", ".jpg", ".jpeg", ".zip"}
+        else zipfile.ZIP_DEFLATED
+    )

apps/admin_console/services/favorites.py

@@ -0,0 +1,466 @@
+"""Independent favorite snapshots dispatched through the existing task services."""
+
+from __future__ import annotations
+
+from copy import deepcopy
+import asyncio
+import hashlib
+import json
+import time
+from typing import Literal
+from uuid import NAMESPACE_URL, uuid5
+from weakref import WeakValueDictionary
+
+from pydantic import BaseModel, ConfigDict, Field
+
+from apps.admin_console.database.repositories.session_repository import SessionRepository
+from apps.admin_console.services.automation_store import AutomationStore, identifier, store
+from apps.admin_console.services.execution_config import execution_config
+from apps.admin_console.services.model_service import model_service
+from apps.admin_console.services.task_import import TaskCase, TaskOptions, execution_title
+from artemis.config.exploration import ExplorationConfig
+from artemis.exploration.case_prompt import require_resolved_case
+from artemis.exploration.store import ExplorationStore
+
+
+class FavoriteCreate(BaseModel):
+    model_config = ConfigDict(extra="forbid")
+    source_kind: Literal["task", "batch", "exploration", "exploration_case"] = "task"
+    source_id: str = Field(pattern=r"^[A-Za-z0-9_-]{1,100}$")
+    version: int | None = Field(default=None, ge=1)
+    name: str | None = Field(default=None, min_length=1, max_length=200)
+
+
+class FavoriteUpdate(BaseModel):
+    model_config = ConfigDict(extra="forbid")
+    name: str | None = Field(default=None, min_length=1, max_length=200)
+    task_config: TaskOptions | None = None
+    case_configs: dict[int, TaskOptions] = Field(default_factory=dict)
+
+
+class FavoriteRun(FavoriteUpdate):
+    request_id: str = Field(pattern=r"^[A-Za-z0-9_-]{1,100}$")
+
+
+def _merge_options(target: dict, updates: dict) -> None:
+    if isinstance(updates.get("recording"), dict):
+        updates = {
+            **updates,
+            "recording": {**(target.get("recording") or {}), **updates["recording"]},
+        }
+    target.update(updates)
+
+
+class FavoriteService:
+    def __init__(self, storage: AutomationStore = store):
+        self.store = storage
+        self._run_locks: WeakValueDictionary = WeakValueDictionary()
+
+    def get(self, key: str) -> dict:
+        value = self.store.get("favorite", identifier(key))
+        if value is None:
+            raise KeyError("收藏不存在")
+        if value.get("config_version") != 1:
+            # Complete legacy snapshots once and preserve their list position.
+            # Re-read under the write lock so a concurrent edit cannot be lost.
+            with self.store.connection() as conn:
+                conn.execute("BEGIN IMMEDIATE")
+                row = conn.execute(
+                    "SELECT payload FROM automation_records WHERE kind='favorite' AND id=?", (key,)
+                ).fetchone()
+                if row is None:
+                    raise KeyError("收藏不存在")
+                value = self._complete(json.loads(row[0]))
+                conn.execute(
+                    "UPDATE automation_records SET payload=? WHERE kind='favorite' AND id=?",
+                    (json.dumps(value, ensure_ascii=False), key),
+                )
+                conn.commit()
+        return value
+
+    def list(self) -> list[dict]:
+        return [
+            self.get(item["id"]) if item.get("config_version") != 1 else item
+            for item in self.store.records("favorite", 10000)
+        ]
+
+    def _complete(self, favorite: dict) -> dict:
+        result = deepcopy(favorite)
+        if result.get("config_version") == 1:
+            return result
+        snapshot = result["snapshot"]
+        if result["kind"] == "batch":
+            for case in snapshot["cases"]:
+                case["task_config"] = execution_config(
+                    {
+                        "profile": snapshot.get("profile"),
+                        "device_serial": snapshot.get("device_serial"),
+                        "runtime_overrides": {
+                            "timeout_seconds": snapshot.get("timeout_seconds"),
+                            **(case.get("task_config") or {}),
+                        },
+                    }
+                )["runtime_overrides"]
+        elif result["kind"] == "exploration":
+            config = ExplorationConfig.model_validate(snapshot.get("config") or {}).model_dump()
+            snapshot["config"] = config
+            snapshot["task_config"] = execution_config(
+                {
+                    "runtime_overrides": {
+                        **(snapshot.get("task_config") or {}),
+                        **{
+                            key: config[key]
+                            for key in ("model", "reasoning_effort", "timeout_seconds")
+                        },
+                    }
+                }
+            )["runtime_overrides"]
+        else:
+            snapshot.update(execution_config(snapshot))
+        result["config_version"] = 1
+        return result
+
+    def _task(self, session_id: str, *, fallback: dict | None = None) -> tuple[dict, list[str]]:
+        job = self.store.job(session_id) or {}
+        if snapshot := job.get("execution_config"):
+            return execution_config(snapshot), []
+        row = SessionRepository(self.store.path).get_session_by_id(session_id) or {}
+        submission = job.get("submission") or fallback or {}
+        options = {
+            **(job.get("case") or {}).get("task_config", {}),
+            **(submission.get("runtime_overrides") or {}),
+        }
+        options = {key: value for key, value in options.items() if value is not None}
+        try:
+            device = row.get("device_info") or {}
+            device = json.loads(device) if isinstance(device, str) else device
+        except (ValueError, TypeError):
+            device = {}
+        device = device if isinstance(device, dict) else {}
+        profile = (
+            options.get("profile")
+            or submission.get("profile")
+            or model_service.resolve_session_profile(row)
+            or "flash"
+        )
+        options["profile"] = profile
+        serial = (
+            options.get("device_serial")
+            or submission.get("device_serial")
+            or job.get("device_serial")
+            or device.get("serial")
+        )
+        if serial:
+            options["device_serial"] = serial
+        if serial == "computer:local" or device.get("execution_target") == "computer":
+            options["execution_target"] = "computer"
+        # Historical diagnostics describe the actual model, unlike display metadata
+        # which may reflect today's defaults. Read only execution model roles.
+        with self.store.connection() as conn:
+            rows = conn.execute(
+                "SELECT payload FROM traces WHERE session_id=? AND type='llm_diagnostic' ORDER BY timestamp DESC",
+                (session_id,),
+            ).fetchall()
+        for trace in rows:
+            try:
+                data = json.loads(trace[0])
+            except (ValueError, TypeError):
+                continue
+            if str(data.get("role", "")).lower() in {
+                "operator",
+                "flashrunner",
+                "reactiverunner",
+                "planner",
+            }:
+                for key in ("model", "reasoning_effort"):
+                    if data.get(key):
+                        options.setdefault(key, data[key])
+                if options.get("model"):
+                    break
+        missing = [
+            key
+            for key in (
+                "model",
+                "reasoning_effort",
+                "record_video",
+                "recording",
+                "analyze_video",
+                "request_timeout_seconds",
+                "max_retries",
+                "retry_interval_seconds",
+                "timeout_seconds",
+            )
+            if key not in options
+        ]
+        snapshot = execution_config(
+            {**submission, "profile": profile, "runtime_overrides": options}
+        )
+        return snapshot, missing
+
+    def create(self, value: FavoriteCreate) -> dict:
+        token = f"{value.source_kind}:{value.source_id}:{value.version or ''}"
+        key = hashlib.sha256(token.encode()).hexdigest()[:32]
+        if self.store.get("favorite", key):
+            return self.get(key)
+        snapshot, missing, kind, version = {}, [], value.source_kind, None
+        source_id = value.source_id
+        repo = ExplorationStore(self.store)
+        if kind == "task":
+            run = repo.run(source_id)
+            if run and run["kind"] == "exploration":
+                source_id = run["exploration_id"]
+                kind = "exploration_case" if repo.case(source_id) else "exploration"
+        if kind in {"exploration", "exploration_case"}:
+            session = repo.get(source_id)
+            options = deepcopy(session["task_config"])
+            if kind == "exploration":
+                snapshot = {
+                    "goal": session["goal"],
+                    "config": session["config"],
+                    "task_config": options,
+                }
+                name = session["goal"][:120]
+            else:
+                case = repo.case(source_id, value.version)
+                if not case:
+                    raise KeyError("确认用例不存在")
+                version = case["version"]
+                validations = [
+                    r
+                    for r in repo.runs(source_id)
+                    if r["kind"] == "validation" and r.get("case_version") == version
+                ]
+                if validations:
+                    config, missing = self._task(
+                        validations[-1]["id"],
+                        fallback={"runtime_overrides": validations[-1]["task_config"]},
+                    )
+                else:
+                    options = {
+                        k: v
+                        for k, v in options.items()
+                        if k not in {"profile", "model", "reasoning_effort", "timeout_seconds"}
+                    }
+                    config = execution_config({"profile": "flash", "runtime_overrides": options})
+                    missing = ["首次用例复跑，默认 Flash"]
+                snapshot = {"goal": require_resolved_case(case), **config}
+                name = case["name"]
+        elif kind == "batch":
+            batch = self.store.get("batch", source_id)
+            if not batch or batch.get("status") == "deleting":
+                raise KeyError("批次不存在")
+            cases, seen = [], set()
+            for job in self.store.jobs(source_id):
+                position = job.get("case_position", job["position"])
+                if position in seen:
+                    continue
+                seen.add(position)
+                case = deepcopy(job["case"])
+                config, inherited = self._task(
+                    job["session_id"],
+                    fallback={
+                        "profile": batch["profile"],
+                        "device_serial": batch.get("device_serial"),
+                        "runtime_overrides": {
+                            "timeout_seconds": batch.get("timeout_seconds"),
+                            **case.get("task_config", {}),
+                        },
+                    },
+                )
+                case["task_config"] = config["runtime_overrides"]
+                cases.append(TaskCase.model_validate(case).model_dump(exclude_none=True))
+                missing.extend(inherited)
+            if not cases:
+                raise ValueError("批次没有可收藏的用例")
+            snapshot = {
+                key: batch[key]
+                for key in (
+                    "name",
+                    "formats",
+                    "profile",
+                    "device_serial",
+                    "timeout_seconds",
+                    "stop_on_failure",
+                    "repeat_count",
+                )
+                if key in batch
+            }
+            snapshot["cases"] = cases
+            name = batch["name"]
+        else:
+            row = SessionRepository(self.store.path).get_session_by_id(source_id) or {}
+            job = self.store.job(source_id) or {}
+            goal = (
+                (job.get("case") or {}).get("goal")
+                or row.get("initial_goal")
+                or (job.get("submission") or {}).get("goal")
+            )
+            if not goal:
+                raise KeyError("任务不存在或没有可复用的提示词")
+            config, missing = self._task(source_id)
+            snapshot = {"goal": goal, **config}
+            name = execution_title(job) or goal[:120]
+        favorite = {
+            "id": key,
+            "name": (value.name or name).strip(),
+            "kind": kind,
+            "source_kind": value.source_kind,
+            "source_id": value.source_id,
+            "case_version": version,
+            "snapshot": snapshot,
+            "inherited_fields": sorted(set(missing)),
+            "created_at": time.time(),
+            "updated_at": time.time(),
+        }
+        if not favorite["name"]:
+            raise ValueError("收藏名称不能为空")
+        return self.store.put("favorite", key, self._complete(favorite))
+
+    def updated(self, favorite: dict, value: FavoriteUpdate) -> dict:
+        result = self._complete(favorite)
+        if value.name is not None:
+            if not value.name.strip():
+                raise ValueError("收藏名称不能为空")
+            result["name"] = value.name.strip()
+        snapshot = result["snapshot"]
+        options = value.task_config.model_dump(exclude_none=True) if value.task_config else {}
+        if result["kind"] == "batch":
+            for index, case in enumerate(snapshot["cases"]):
+                if override := value.case_configs.get(index):
+                    _merge_options(case["task_config"], override.model_dump(exclude_none=True))
+                _merge_options(case["task_config"], options)
+            if set(value.case_configs) - set(range(len(snapshot["cases"]))):
+                raise ValueError("批次用例下标不存在")
+            if options.get("device_serial"):
+                snapshot["device_serial"] = options["device_serial"]
+        elif result["kind"] == "exploration":
+            config = snapshot["config"]
+            config.update(
+                {
+                    k: options[k]
+                    for k in ("model", "reasoning_effort", "timeout_seconds")
+                    if k in options
+                }
+            )
+            snapshot["config"] = ExplorationConfig.model_validate(config).model_dump()
+            _merge_options(snapshot["task_config"], options)
+        else:
+            _merge_options(snapshot["runtime_overrides"], options)
+        return result
+
+    def update(self, key: str, value: FavoriteUpdate) -> dict:
+        favorite = self.updated(self.get(key), value)
+        favorite["updated_at"] = time.time()
+        return self.store.put("favorite", key, favorite)
+
+    def delete(self, key: str) -> None:
+        with self.store.connection() as conn:
+            conn.execute(
+                "DELETE FROM automation_records WHERE kind='favorite' AND id=?", (identifier(key),)
+            )
+            conn.commit()
+
+    async def run(self, key: str, value: FavoriteRun) -> dict:
+        lock = self._run_locks.setdefault(value.request_id, asyncio.Lock())
+        async with lock:
+            return await self._run(key, value)
+
+    async def _run(self, key: str, value: FavoriteRun) -> dict:
+        key = identifier(key)
+        request = {"favorite_id": key, **value.model_dump(exclude_none=True)}
+        signature = hashlib.sha256(json.dumps(request, sort_keys=True).encode()).hexdigest()
+        # Reserve a durable snapshot before admission. Retries survive restarts,
+        # edits/deletions of the favorite, and lost HTTP responses.
+        with self.store.connection() as conn:
+            conn.execute("BEGIN IMMEDIATE")
+            row = conn.execute(
+                "SELECT payload FROM automation_records WHERE kind='favorite_run' AND id=?",
+                (value.request_id,),
+            ).fetchone()
+            record = json.loads(row[0]) if row else None
+            if record and record["signature"] != signature:
+                raise ValueError("request_id 已用于其他内容")
+            if record and record.get("result"):
+                return record["result"]
+            if record is None:
+                row = conn.execute(
+                    "SELECT payload FROM automation_records WHERE kind='favorite' AND id=?", (key,)
+                ).fetchone()
+                if row is None:
+                    raise KeyError("收藏不存在")
+                record = {
+                    "signature": signature,
+                    "favorite": self.updated(json.loads(row[0]), value),
+                }
+                conn.execute(
+                    "INSERT INTO automation_records VALUES (?,?,?,?)",
+                    (
+                        "favorite_run",
+                        value.request_id,
+                        json.dumps(record, ensure_ascii=False),
+                        time.time(),
+                    ),
+                )
+            conn.commit()
+        favorite = record["favorite"]
+        snapshot = favorite["snapshot"]
+        run_id = str(uuid5(NAMESPACE_URL, "artemis-favorite:" + value.request_id))
+        if favorite["kind"] == "batch":
+            from apps.admin_console.services.automation_service import create_batch, batch_action
+
+            batch = create_batch(
+                {**snapshot, "name": favorite["name"], "request_id": "favorite-" + run_id}
+            )
+            if batch["status"] == "draft":
+                batch = batch_action(batch["id"], "resume")
+            result = {"kind": "batch", "batch_id": batch["id"], "status": batch["status"]}
+        elif favorite["kind"] == "exploration":
+            from apps.admin_console.services.exploration_service import exploration_service
+            from artemis.exploration.models import CreateExploration
+
+            session = exploration_service.create(CreateExploration(request_id=run_id, **snapshot))
+            admission = await exploration_service.start(session["id"], run_id)
+            run = admission.get("run", admission)
+            rejection = admission.get("admission", {})
+            outcome = run.get("outcome") or {}
+            if (
+                rejection.get("status") in {"error", "rejected"}
+                or outcome.get("status") == "rejected"
+            ):
+                raise ValueError(
+                    rejection.get("error")
+                    or rejection.get("message")
+                    or outcome.get("error")
+                    or "探索未入队"
+                )
+            result = {
+                "kind": "exploration",
+                "session_id": run["id"],
+                "exploration_id": session["id"],
+                "admission": admission,
+            }
+        else:
+            from apps.admin_console.routers.tasks import run_task
+            from apps.admin_console.schemas.task_schema import RunRequest
+
+            admission = await run_task(
+                RunRequest(
+                    goal=snapshot["goal"],
+                    profile=snapshot["runtime_overrides"].get("profile", "flash"),
+                    runtime_overrides=TaskOptions.model_validate(snapshot["runtime_overrides"]),
+                    session_id=run_id,
+                    ingress="favorite",
+                    expected_output=snapshot.get("expected_output"),
+                    enable_outputter=snapshot.get("enable_outputter"),
+                    app_path=snapshot.get("app_path"),
+                )
+            )
+            if not admission.get("tasks"):
+                raise ValueError(admission.get("error") or admission.get("message") or "任务未入队")
+            result = {"kind": "task", "session_id": run_id, "admission": admission}
+        record["result"] = result
+        self.store.put("favorite_run", value.request_id, record)
+        return result
+
+
+favorites = FavoriteService()

apps/admin_console/services/ipc_service.py

@@ -20,9 +20,6 @@ from typing import Any
 
 logger = logging.getLogger(__name__)
 
-try:
-    from admin_console.core.state import state
-except ImportError:
 from apps.admin_console.core.state import state
 
 

apps/admin_console/services/model_settings_service.py

@@ -9,6 +9,7 @@ from urllib.parse import urlsplit
 
 from pydantic import BaseModel, Field, SecretStr, field_validator
 
+from artemis.config.agent import load_agent_config
 from artemis.config.paths import get_env_file
 from artemis.config.settings import settings
 
@@ -16,12 +17,19 @@ _write_lock = threading.Lock()
 
 
 class ModelSettingsUpdate(BaseModel):
-    base_url: str = "https://api.openai.com/v1"
+    base_url: str = "https://api-n-cd.playones.com/v1"
     model: str = Field(min_length=1, max_length=200)
     api_key: SecretStr | None = None  # Omitted/blank keeps the existing key.
     timeout_seconds: float = Field(default=300, gt=0, le=3600)
     max_retries: int = Field(default=5, ge=0, le=10)
     retry_interval_seconds: float = Field(default=10, ge=0, le=300)
+    rate_limit_retry_seconds: float | None = Field(default=None, ge=0, le=3600)
+    tool_fallback_model: str | None = Field(default=None, min_length=1, max_length=200)
+    tool_fallback_reasoning_effort: (
+        Literal["default", "none", "minimal", "low", "medium", "high", "xhigh", "max"] | None
+    ) = None
+    tool_fallback_max_retries: int | None = Field(default=None, ge=0, le=10)
+    tool_fallback_no_progress_attempts: int | None = Field(default=None, ge=1, le=10)
     reasoning_effort: Literal[
         "default", "none", "minimal", "low", "medium", "high", "xhigh", "max"
     ] = "high"
@@ -41,12 +49,14 @@ class ModelSettingsUpdate(BaseModel):
         ):
             raise ValueError("请输入有效的 HTTP(S) 接口根地址，不要包含密钥、查询参数或片段")
         if value.endswith(("/responses", "/chat/completions")):
-            raise ValueError("请输入接口根地址，例如 https://api.openai.com/v1")
+            raise ValueError("请输入接口根地址，例如 https://api-n-cd.playones.com/v1")
         return value
 
-    @field_validator("model")
+    @field_validator("model", "tool_fallback_model")
     @classmethod
-    def validate_model(cls, value: str) -> str:
+    def validate_model(cls, value: str | None) -> str | None:
+        if value is None:
+            return None
         value = value.strip()
         if not value or any(c in value for c in "\r\n\x00"):
             raise ValueError("模型名称不能为空或包含换行")
@@ -57,14 +67,24 @@ def get_model_settings() -> dict:
     return {
         "provider": "openai",
         "api_mode": "responses",
-        "base_url": settings.OPENAI_BASE_URL or "https://api.openai.com/v1",
+        "base_url": settings.OPENAI_BASE_URL or "https://api-n-cd.playones.com/v1",
         "model": settings.OPENAI_MODEL,
         "api_key_configured": bool(settings.OPENAI_API_KEY),
         "timeout_seconds": settings.OPENAI_TIMEOUT_SECONDS,
         "max_retries": settings.OPENAI_MAX_RETRIES,
         "retry_interval_seconds": settings.OPENAI_RETRY_INTERVAL_SECONDS,
+        "rate_limit_retry_seconds": settings.OPENAI_RATE_LIMIT_RETRY_SECONDS,
+        "tool_fallback_model": settings.OPENAI_TOOL_FALLBACK_MODEL,
+        "tool_fallback_reasoning_effort": settings.OPENAI_TOOL_FALLBACK_REASONING_EFFORT,
+        "tool_fallback_max_retries": settings.OPENAI_TOOL_FALLBACK_MAX_RETRIES,
+        "tool_fallback_no_progress_attempts": settings.ARTEMIS_TOOL_FALLBACK_MAX_NO_PROGRESS_ATTEMPTS,
         "reasoning_effort": settings.OPENAI_REASONING_EFFORT,
-        "model_presets": ["gpt-5.6-luna", "mimo-v2.6-flash"],
+        "model_presets": [
+            "gpt-6-luna",
+            "gpt-6.1-sol",
+            "gpt-6-astra",
+            "mimo-v2.6-pro",
+        ],
         "env_path": str(get_env_file()),
     }
 
@@ -80,6 +100,20 @@ def save_model_settings(update: ModelSettingsUpdate) -> dict:
         "OPENAI_RETRY_INTERVAL_SECONDS": str(update.retry_interval_seconds),
         "OPENAI_REASONING_EFFORT": update.reasoning_effort,
     }
+    optional_settings = {
+        "rate_limit_retry_seconds": ("OPENAI_RATE_LIMIT_RETRY_SECONDS", float),
+        "tool_fallback_model": ("OPENAI_TOOL_FALLBACK_MODEL", str),
+        "tool_fallback_reasoning_effort": ("OPENAI_TOOL_FALLBACK_REASONING_EFFORT", str),
+        "tool_fallback_max_retries": ("OPENAI_TOOL_FALLBACK_MAX_RETRIES", int),
+        "tool_fallback_no_progress_attempts": (
+            "ARTEMIS_TOOL_FALLBACK_MAX_NO_PROGRESS_ATTEMPTS",
+            int,
+        ),
+    }
+    for field_name, (env_name, expected_type) in optional_settings.items():
+        value = getattr(update, field_name)
+        if value is not None:
+            values[env_name] = str(value)
     if update.api_key and update.api_key.get_secret_value().strip():
         values["OPENAI_API_KEY"] = update.api_key.get_secret_value().strip()
     persist_environment(values)
@@ -89,6 +123,10 @@ def save_model_settings(update: ModelSettingsUpdate) -> dict:
     settings.OPENAI_MAX_RETRIES = update.max_retries
     settings.OPENAI_RETRY_INTERVAL_SECONDS = update.retry_interval_seconds
     settings.OPENAI_REASONING_EFFORT = update.reasoning_effort
+    for field_name, (env_name, expected_type) in optional_settings.items():
+        value = getattr(update, field_name)
+        if value is not None:
+            setattr(settings, env_name, expected_type(value))
     if "OPENAI_API_KEY" in values:
         settings.OPENAI_API_KEY = SecretStr(values["OPENAI_API_KEY"])
     from artemis.llm.router import ModelFactory
@@ -128,26 +166,79 @@ def persist_environment(values: dict[str, str]) -> None:
 
 
 class CapabilitySettings(BaseModel):
+    adb_auto_connect: bool = Field(default_factory=lambda: settings.ARTEMIS_ADB_AUTO_CONNECT)
+    adb_discovery_interval_seconds: float = Field(
+        default_factory=lambda: settings.ARTEMIS_ADB_DISCOVERY_INTERVAL_SECONDS, ge=1, le=300
+    )
+    adb_discovery_timeout_seconds: float = Field(
+        default_factory=lambda: settings.ARTEMIS_ADB_DISCOVERY_TIMEOUT_SECONDS, gt=0, le=30
+    )
     video_enabled: bool = False
+    recording_enabled: bool = False
+    recording: dict = Field(default_factory=dict)
     outputter_enabled: bool = True
     step_summarizer_enabled: bool = True
+    screenshot_limit_enabled: bool = True
+    screenshot_max_long_edge: int = 2048
+    task_timeout_seconds: int = Field(
+        default_factory=lambda: settings.ARTEMIS_TASK_TIMEOUT_SECONDS, ge=30, le=86400
+    )
+    observation_timeout_seconds: float = Field(
+        default_factory=lambda: settings.ARTEMIS_OBSERVATION_TIMEOUT_SECONDS, gt=0, le=300
+    )
+    observation_max_attempts: int = Field(
+        default_factory=lambda: settings.ARTEMIS_OBSERVATION_MAX_ATTEMPTS, ge=1, le=5
+    )
+    device_command_timeout_seconds: float = Field(
+        default_factory=lambda: settings.ARTEMIS_DEVICE_COMMAND_TIMEOUT_SECONDS, gt=0, le=60
+    )
+    action_failure_attempts: int = Field(
+        default_factory=lambda: settings.ARTEMIS_ACTION_FAILURE_ATTEMPTS, ge=1, le=10
+    )
+    worker_cleanup_timeout_seconds: float = Field(
+        default_factory=lambda: settings.ARTEMIS_WORKER_CLEANUP_TIMEOUT_SECONDS, gt=0, le=300
+    )
 
 
 def get_capability_settings() -> CapabilitySettings:
+    screenshot = load_agent_config().screenshot
     return CapabilitySettings(
         video_enabled=settings.ARTEMIS_VIDEO_ENABLED,
+        recording_enabled=settings.ARTEMIS_RECORDING_ENABLED,
+        recording=load_agent_config().recording.model_dump(),
         outputter_enabled=settings.ARTEMIS_OUTPUTTER_ENABLED,
         step_summarizer_enabled=settings.ARTEMIS_STEP_SUMMARIZER_ENABLED,
+        screenshot_limit_enabled=screenshot.enabled,
+        screenshot_max_long_edge=screenshot.max_long_edge,
     )
 
 
 def save_capability_settings(update: CapabilitySettings) -> CapabilitySettings:
+    from artemis.config.agent import RecordingConfig
+    import json
+
+    recording = RecordingConfig.model_validate(
+        {**load_agent_config().recording.model_dump(), **update.recording}
+    ).model_dump()
     values = {
+        "ARTEMIS_ADB_AUTO_CONNECT": update.adb_auto_connect,
+        "ARTEMIS_ADB_DISCOVERY_INTERVAL_SECONDS": update.adb_discovery_interval_seconds,
+        "ARTEMIS_ADB_DISCOVERY_TIMEOUT_SECONDS": update.adb_discovery_timeout_seconds,
         "ARTEMIS_VIDEO_ENABLED": update.video_enabled,
+        "ARTEMIS_RECORDING_ENABLED": update.recording_enabled,
         "ARTEMIS_OUTPUTTER_ENABLED": update.outputter_enabled,
         "ARTEMIS_STEP_SUMMARIZER_ENABLED": update.step_summarizer_enabled,
+        "ARTEMIS_SCREENSHOT_LIMIT_ENABLED": update.screenshot_limit_enabled,
+        "ARTEMIS_TASK_TIMEOUT_SECONDS": update.task_timeout_seconds,
+        "ARTEMIS_OBSERVATION_TIMEOUT_SECONDS": update.observation_timeout_seconds,
+        "ARTEMIS_OBSERVATION_MAX_ATTEMPTS": update.observation_max_attempts,
+        "ARTEMIS_DEVICE_COMMAND_TIMEOUT_SECONDS": update.device_command_timeout_seconds,
+        "ARTEMIS_ACTION_FAILURE_ATTEMPTS": update.action_failure_attempts,
+        "ARTEMIS_WORKER_CLEANUP_TIMEOUT_SECONDS": update.worker_cleanup_timeout_seconds,
     }
     persist_environment({key: str(value).lower() for key, value in values.items()})
+    persist_environment({"ARTEMIS_RECORDING_CONFIG": json.dumps(recording)})
+    settings.ARTEMIS_RECORDING_CONFIG = recording
     for key, value in values.items():
         setattr(settings, key, value)
     return get_capability_settings()

apps/admin_console/services/request_diagnostics.py

@@ -0,0 +1,112 @@
+"""Read/export model diagnostics from the task's existing trace database."""
+
+from datetime import UTC, datetime
+import json
+import time
+
+from apps.admin_console.database.connection import db_session
+
+
+def request_diagnostics(
+    session_id: str, *, db_path=None, limit: int | None = None, offset: int = 0
+) -> dict:
+    with db_session(db_path) as conn:
+        session = conn.execute(
+            "SELECT status FROM sessions WHERE session_id = ?", (session_id,)
+        ).fetchone()
+        sql = (
+            "SELECT payload FROM traces WHERE session_id = ? AND type = 'llm_diagnostic' "
+            "ORDER BY timestamp DESC"
+        )
+        params: tuple = (session_id,)
+        total = conn.execute(
+            "SELECT count(*) FROM traces WHERE session_id=? AND type='llm_diagnostic'",
+            params,
+        ).fetchone()[0]
+        if limit is not None:
+            sql += " LIMIT ? OFFSET ?"
+            params += (limit, offset)
+        rows = conn.execute(sql, params).fetchall()
+        review_rows = conn.execute(
+            "SELECT payload FROM traces WHERE session_id = ? "
+            "AND type = 'final_review_diagnostic' ORDER BY timestamp DESC"
+            + (" LIMIT ?" if limit is not None else ""),
+            (session_id, limit) if limit is not None else (session_id,),
+        ).fetchall()
+    requests = []
+    seen = set()
+    running = bool(session and session["status"] in {"running", "resuming"})
+    for row in rows:
+        try:
+            value = json.loads(row["payload"])
+        except (ValueError, TypeError):
+            continue
+        if not isinstance(value, dict):
+            continue
+        request_id = value.get("request_id")
+        value["active"] = (
+            running
+            and offset == 0
+            and request_id not in seen
+            and value.get("phase")
+            in {
+                "waiting_response",
+                "waiting_output",
+                "receiving_output",
+                "generating_tool_arguments",
+                "retry_wait",
+            }
+        )
+        seen.add(request_id)
+        value["timestamps_utc"] = {
+            field: datetime.fromtimestamp(value[field], UTC).isoformat()
+            for field in (
+                "started_at",
+                "headers_at",
+                "first_output_at",
+                "last_output_at",
+                "ended_at",
+                "retry_at",
+            )
+            if isinstance(value.get(field), (int, float))
+        }
+        requests.append(value)
+    final_reviews = []
+    for row in review_rows:
+        try:
+            value = json.loads(row["payload"])
+        except (ValueError, TypeError):
+            continue
+        if not isinstance(value, dict):
+            continue
+        value["active"] = (
+            running and offset == 0 and not final_reviews and value.get("status") == "running"
+        )
+        if value["active"] and isinstance(value.get("started_at"), (int, float)):
+            value["elapsed_seconds"] = round(max(0.0, time.time() - value["started_at"]), 3)
+            value["remaining_seconds"] = round(
+                max(0.0, value.get("total_seconds", 0.0) - value["elapsed_seconds"]), 3
+            )
+            value["budget_observed_at"] = time.time()
+        value["timestamps_utc"] = {
+            field: datetime.fromtimestamp(value[field], UTC).isoformat()
+            for field in ("started_at", "deadline_at", "ended_at")
+            if isinstance(value.get(field), (int, float))
+        }
+        final_reviews.append(value)
+    return {
+        "schema_version": 1,
+        "session_id": session_id,
+        "session_status": session["status"] if session else "unknown",
+        "collected_at": time.time(),
+        "requests": requests,
+        "final_reviews": final_reviews,
+        "total": total,
+        "offset": offset,
+        "next_offset": offset + limit if limit is not None and offset + limit < total else None,
+        "note": "时间戳为 Unix 秒；timestamps_utc 为 UTC。HTTP 200 不代表完整输出。usage=null 表示服务未报告用量。历史任务没有记录的字段无法补推。",
+    }
+
+
+def diagnostics_json(session_id: str) -> str:
+    return json.dumps(request_diagnostics(session_id), ensure_ascii=False, indent=2)

apps/admin_console/services/task_deletion.py

@@ -0,0 +1,40 @@
+"""Read the impact of deleting a task without deleting or stopping anything."""
+
+from uuid import UUID
+
+from artemis.exploration.store import ExplorationConflict, explorations
+from apps.admin_console.database.repositories.session_repository import session_repo
+from apps.admin_console.services.automation_store import store
+from apps.admin_console.services.exploration_deletion import _check_runs
+
+
+def deletion_preview(session_id: str) -> dict:
+    session_id = str(UUID(session_id))
+    run = explorations.run(session_id)
+    if run and run["kind"] == "exploration":
+        return exploration_deletion_preview(run["exploration_id"])
+    if not session_repo.get_session_by_id(session_id) and not store.job(session_id):
+        raise KeyError("任务不存在")
+    _check_runs([{"id": session_id}])
+    job = store.job(session_id) or {}
+    if job.get("status") in {"pending", "queued", "running", "task_paused", "resuming"}:
+        raise ExplorationConflict("任务仍在执行或排队，请先停止并等待结束")
+    return {
+        "session_ids": [session_id],
+        "preserved_session_ids": [],
+        "removes": ["任务历史", "报告", "截图、录像及附件"],
+    }
+
+
+def exploration_deletion_preview(exploration_id: str) -> dict:
+    session = explorations.get(exploration_id)
+    runs = explorations.runs(exploration_id)
+    _check_runs(runs)
+    if session.get("active_run_id") or session.get("active_demonstration_id"):
+        raise ExplorationConflict("探索尚有活动任务或示范，请先停止并等待结束")
+    return {
+        "exploration_id": exploration_id,
+        "session_ids": [run["id"] for run in runs if run["kind"] == "exploration"],
+        "preserved_session_ids": [run["id"] for run in runs if run["kind"] != "exploration"],
+        "removes": ["探索对话", "用例版本", "探索执行历史及证据"],
+    }

apps/admin_console/services/task_import.py

@@ -1,4 +1,4 @@
-"""Versioned Excel task template and bounded, non-executing import validation."""
+"""Validate the prompt-first batch workbook without executing its contents."""
 
 from io import BytesIO
 import json
@@ -6,33 +6,68 @@ import os
 import zipfile
 
 from openpyxl import Workbook, load_workbook
+from openpyxl.styles import Alignment, Font, PatternFill
 from openpyxl.utils import get_column_letter
-from openpyxl.styles import Font, PatternFill, Alignment
+from pydantic import BaseModel, ConfigDict, Field, field_serializer
-from pydantic import BaseModel, Field, ConfigDict
+from artemis.plugins.models import ExecutionTarget, PluginSelection
+from artemis.config.agent import RecordingConfig
 
-
+HEADERS = (
-HEADERS = {
+    "用例编号",
-    "任务名称": "name",
+    "检测项目",
-    "任务目标": "goal",
+    "测试子项",
-    "预期结果": "expected",
+    "测试用例",
-    "目标应用": "app",
+    "前置约束提取",
-    "前置条件": "preconditions",
+    "Artemis提示词",
-    "输入参数(JSON)": "parameters",
+    "任务配置(JSON)",
-    "标签": "tags",
+)
-}
 MAX_ROWS = int(os.getenv("ARTEMIS_IMPORT_MAX_ROWS", "1000"))
 MAX_BYTES = int(os.getenv("ARTEMIS_IMPORT_MAX_BYTES", "10485760"))
 
 
+class TaskOptions(BaseModel):
+    """Allowlisted per-case overrides; credentials are never accepted in a workbook."""
+
+    model_config = ConfigDict(extra="forbid")
+    profile: str | None = Field(default=None, pattern="^(flash|pro)$")
+    model: str | None = Field(default=None, min_length=1, max_length=200)
+    reasoning_effort: str | None = Field(
+        default=None, pattern="^(default|none|minimal|low|medium|high|xhigh|max)$"
+    )
+    device_serial: str | None = Field(default=None, min_length=1, max_length=200)
+    verification_level: str | None = Field(default=None, pattern="^(off|final|checkpoints|strict)$")
+    explorer_mode: str | None = Field(default=None, pattern="^(flash|pro|ultra)$")
+    locked_app_package: str | None = Field(default=None, min_length=1, max_length=200)
+    record_video: bool | None = None
+    recording: RecordingConfig | None = None
+    analyze_video: bool | None = None
+    request_timeout_seconds: float | None = Field(default=None, gt=0, le=3600)
+    max_retries: int | None = Field(default=None, ge=0, le=10)
+    retry_interval_seconds: float | None = Field(default=None, ge=0, le=300)
+    timeout_seconds: int | None = Field(default=None, ge=30, le=86400)
+    observation_timeout_seconds: float | None = Field(default=None, gt=0, le=300)
+    observation_max_attempts: int | None = Field(default=None, ge=1, le=5)
+    device_command_timeout_seconds: float | None = Field(default=None, gt=0, le=60)
+    action_failure_attempts: int | None = Field(default=None, ge=1, le=10)
+    worker_cleanup_timeout_seconds: float | None = Field(default=None, gt=0, le=300)
+    plugins: list[PluginSelection] | None = None
+    execution_target: ExecutionTarget | None = None
+
+    @field_serializer("recording", when_used="unless-none")
+    def serialize_recording(self, value: RecordingConfig) -> dict:
+        # A per-task partial override must not replace other saved recording defaults.
+        return value.model_dump(exclude_unset=True)
+
+
 class TaskCase(BaseModel):
     model_config = ConfigDict(extra="forbid")
+    case_id: str = Field(default="", max_length=200)
+    category: str = Field(default="", max_length=200)
+    subcategory: str = Field(default="", max_length=200)
     name: str = Field(min_length=1, max_length=200)
-    goal: str = Field(min_length=1, max_length=10000)
-    expected: str = Field(default="", max_length=10000)
-    app: str = Field(default="", max_length=200)
     preconditions: str = Field(default="", max_length=10000)
-    parameters: dict[str, str | int | float | bool] = Field(default_factory=dict)
+    goal: str = Field(min_length=1, max_length=10000)
-    tags: str = Field(default="", max_length=500)
+    task_config: TaskOptions = Field(default_factory=TaskOptions)
 
 
 def template() -> bytes:
@@ -40,27 +75,54 @@ def template() -> bytes:
     ws = wb.active
     assert ws is not None
     ws.title = "任务"
-    ws.append(list(HEADERS))
+    ws.append(HEADERS)
     ws.freeze_panes = "A2"
     ws.auto_filter.ref = "A1:G1"
     for cell in ws[1]:
         cell.font = Font(bold=True, color="FFFFFF")
         cell.fill = PatternFill("solid", fgColor="245C78")
-        ws.column_dimensions[get_column_letter(int(cell.column or 1))].width = 32
+        ws.column_dimensions[get_column_letter(int(cell.column or 1))].width = (
+            65 if cell.column in {6, 7} else 28
+        )
     help_sheet = wb.create_sheet("填写说明")
-    for line in [
+    help_sheet.append(
-        "模板版本：1",
+        ["模板版本：2。任务表每行一项，Artemis提示词必填；只把这一列作为 AI 任务正文。"]
-        "每行一个独立任务；任务名称、任务目标必填。",
+    )
-        "预期结果决定要验证的业务条件。未填写时只报告任务完成情况。",
+    help_sheet.append(["前置约束提取及其他描述列仅作查看、搜索和报告元数据，不追加到提示词。"])
-        "前置条件描述实际用例起点；不会统一重启或清空应用数据。",
+    help_sheet.append(["任务配置(JSON)可留空，空字段继承批次或全局配置；禁止密钥与未知字段。"])
-        '输入参数使用 JSON 对象，例如 {"关键词":"天气"}；目标可引用 {{关键词}}。',
+    help_sheet.append(["示例仅供参考，不会导入执行。请在“任务”页填写自己的用例。"])
-        "目标应用可填名称或包名；空白时由任务决定。",
+    help_sheet.append(HEADERS)
-        "不要填写 Excel 公式、宏或外部链接。导入仅预览，点击开始后执行。",
+    help_sheet.append(
-    ]:
+        [
-        help_sheet.append([line])
+            "示例-001",
-    help_sheet.column_dimensions["A"].width = 105
+            "示例模块",
+            "示例检查",
+            "检查示例页面",
+            "设备已经打开目标应用",
+            "观察当前页面并报告实际结果，不要假设检查已通过。",
+            json.dumps(
+                {
+                    "profile": "pro",
+                    "model": "gpt-6-luna",
+                    "reasoning_effort": "high",
+                    "verification_level": "final",
+                    "record_video": True,
+                    "analyze_video": False,
+                    "request_timeout_seconds": 300,
+                    "max_retries": 5,
+                    "retry_interval_seconds": 10,
+                    "timeout_seconds": 1800,
+                },
+                ensure_ascii=False,
+            ),
+        ]
+    )
+    for column in range(1, 8):
+        help_sheet.column_dimensions[get_column_letter(column)].width = 70 if column > 5 else 26
     for row in help_sheet:
-        row[0].alignment = Alignment(wrap_text=True)
+        for cell in row:
+            cell.alignment = Alignment(wrap_text=True, vertical="top")
+    help_sheet.row_dimensions[6].height = 110
     stream = BytesIO()
     wb.save(stream)
     return stream.getvalue()
@@ -90,7 +152,7 @@ def parse_workbook(data: bytes) -> dict:
         rows = iter(wb["任务"].iter_rows())
         headers = [str(cell.value or "").strip() for cell in next(rows, [])]
         if headers != list(HEADERS):
-            raise ValueError("列名或顺序与模板不一致，请重新下载模板")
+            raise ValueError("列名或顺序与新版模板不一致，请重新下载模板")
         tasks, errors = [], []
         for number, row in enumerate(rows, 2):
             if number > MAX_ROWS + 1:
@@ -100,28 +162,48 @@ def parse_workbook(data: bytes) -> dict:
             try:
                 if any(cell.data_type == "f" for cell in row):
                     raise ValueError("不允许公式，请粘贴为纯文本")
-                values = [str(cell.value or "").strip() for cell in row[: len(HEADERS)]]
+                values = [
-                item = dict(zip(HEADERS.values(), values))
+                    str(cell.value if cell.value is not None else "").strip() for cell in row[:7]
-                item["parameters"] = json.loads(item.get("parameters") or "{}")
+                ]
-                case = TaskCase.model_validate(item)
+                config = json.loads(values[6] or "{}")
+                if not isinstance(config, dict):
+                    raise ValueError("任务配置必须为 JSON 对象")
+                case = TaskCase.model_validate(
+                    {
+                        "case_id": values[0],
+                        "category": values[1],
+                        "subcategory": values[2],
+                        "name": values[3] or values[2] or values[0] or values[5][:120],
+                        "preconditions": values[4],
+                        "goal": values[5],
+                        "task_config": config,
+                    }
+                )
                 tasks.append({"row": number, **case.model_dump()})
             except (ValueError, TypeError) as exc:
                 errors.append({"row": number, "message": str(exc)[:500]})
         if not tasks and not errors:
             errors.append({"row": 2, "message": "没有任务，请填写模板后重新导入"})
-        return {"version": 1, "tasks": tasks, "errors": errors, "valid": bool(tasks) and not errors}
+        return {"version": 2, "tasks": tasks, "errors": errors, "valid": bool(tasks) and not errors}
     finally:
         wb.close()
 
 
 def case_goal(case: dict) -> str:
-    goal = case["goal"]
+    """Execution gets the prompt verbatim; metadata stays outside model input."""
-    for key, value in case.get("parameters", {}).items():
+    return case["goal"]
-        goal = goal.replace("{{" + key + "}}", str(value))
+
-    return (
+
-        f"任务名称：{case['name']}\n任务目标：{goal}\n目标应用：{case.get('app') or '按目标判断'}\n"
+def case_title(case: dict) -> str:
-        f"前置条件：{case.get('preconditions') or '根据实际用例判断合适起点'}\n"
+    """Human-facing title; never prepend metadata to the model prompt."""
-        "先观察手机并恢复到本用例需要的起始位置，记录准备过程。不要固定重启或清空数据。"
+    case_id = str(case.get("case_id") or "").strip()
-        "如果无法建立前置条件，明确报告前置条件未满足，不继续正文操作。\n"
+    name = str(case.get("name") or "").strip()
-        f"预期结果：{case.get('expected') or '仅确认任务完成情况，不宣称测试通过'}"
+    return " ".join(part for part in (case_id, name) if part)
-    )
+
+
+def execution_title(job: dict) -> str:
+    """Distinguish repeated executions without altering the original case or prompt."""
+    title = case_title(job.get("case") or {})
+    if job.get("repeat_count", 1) > 1:
+        title += f" · 第 {job['iteration']} / {job['repeat_count']} 轮"
+    return title

apps/admin_console/services/task_inspection.py

@@ -0,0 +1,147 @@
+"""Bounded, read-only task diagnostics shared by MCP and HTTP clients."""
+
+import os
+import re
+
+from apps.admin_console.database.repositories.session_repository import session_repo
+from apps.admin_console.services import evidence_service, report_export
+from apps.admin_console.services.automation_store import identifier, store
+from apps.admin_console.services.request_diagnostics import request_diagnostics
+
+
+_SECRET = re.compile(
+    r"(?i)^(api[_-]?key|access[_-]?token|refresh[_-]?token|authorization|password|secret|client_secret|token_key)$"
+)
+
+
+def public_value(value, *, max_chars: int = 60000):
+    """Remove common credential forms and bound text without inventing missing evidence."""
+    remaining = max_chars
+    remaining_nodes = 2000
+    truncated = False
+    secrets = [
+        v
+        for k, v in os.environ.items()
+        if len(v) >= 8 and re.search(r"(?i)(api_?key|password|secret|token)$", k)
+    ]
+
+    def clean(item, depth=0):
+        nonlocal remaining, remaining_nodes, truncated
+        remaining_nodes -= 1
+        if depth > 12:
+            truncated = True
+            return "[嵌套内容已截断]"
+        if isinstance(item, dict):
+            result = {}
+            for k, v in item.items():
+                if remaining_nodes <= 0 or len(result) >= 100:
+                    truncated = True
+                    break
+                key = clean(str(k), depth + 1)
+                result[key] = "[REDACTED]" if _SECRET.fullmatch(str(k)) else clean(v, depth + 1)
+            return result
+        if isinstance(item, list):
+            result = []
+            for v in item:
+                if remaining_nodes <= 0 or len(result) >= 100:
+                    truncated = True
+                    break
+                result.append(clean(v, depth + 1))
+            return result
+        if not isinstance(item, str):
+            return item
+        for secret in secrets:
+            item = item.replace(secret, "[REDACTED]")
+        item = re.sub(r"(?i)\bBearer\s+[^\s\"',;]+", "Bearer [REDACTED]", item)
+        item = re.sub(r"\bsk-[A-Za-z0-9_-]+", "[REDACTED]", item)
+        item = re.sub(r"(https?://)[^\s/@]+:[^\s/@]+@", r"\1[REDACTED]@", item)
+        item = re.sub(
+            r"""(?ix)((?:api[_-]?key|access[_-]?token|refresh[_-]?token|authorization|password|secret|token_key)["']?\s*[:=]\s*)("[^"]*"|'[^']*'|[^\s&,;]+)""",
+            r"\1[REDACTED]",
+            item,
+        )
+        allowance = min(8000, remaining)
+        if len(item) > allowance:
+            truncated = True
+            item = item[:allowance] + "…[已截断；按步骤或关键词缩小范围]"
+        remaining = max(0, remaining - min(len(item), allowance))
+        return item
+
+    result = clean(value)
+    return {**result, "text_truncated": truncated, "content_truncated": truncated}
+
+
+def page(items: list, offset: int, limit: int) -> dict:
+    return {
+        "items": items[offset : offset + limit],
+        "total": len(items),
+        "offset": offset,
+        "limit": limit,
+        "next_offset": offset + limit if offset + limit < len(items) else None,
+    }
+
+
+def inspect_task(session_id: str, view: str, *, offset: int = 0, limit: int = 20) -> dict:
+    identifier(session_id)
+    if offset < 0 or not 1 <= limit <= 100:
+        raise ValueError("offset 必须非负，limit 必须在 1–100 之间")
+    if not session_repo.get_session_by_id(session_id) and not store.job(session_id):
+        raise LookupError("任务记录不存在")
+    if view == "usage":
+        result = session_repo.get_session_usage(session_id)
+        result["cache_hit_rate"] = (
+            result["cached_tokens"] / result["prompt_tokens"] if result["prompt_tokens"] else None
+        )
+        result["note"] = (
+            "缓存率=已记录 cached_tokens / prompt_tokens；无用量不等于零消耗，未上报的请求无法补推。"
+        )
+    elif view == "requests":
+        result = request_diagnostics(session_id, limit=limit, offset=offset)
+    elif view == "logs":
+        root = evidence_service.task_root(session_id)
+        result = {"logs": []}
+        for name in ("stdout.log", "stderr.log"):
+            path = evidence_service.evidence_file(root, name)
+            if path is None:
+                result["logs"].append({"name": name, "available": False})
+                continue
+            with path.open("rb") as stream:
+                size = path.stat().st_size
+                start = max(0, size - 128 * 1024)
+                stream.seek(start)
+                raw = stream.read(128 * 1024)
+            lines = raw.decode("utf-8", errors="replace").splitlines()
+            if start and lines:
+                lines = lines[1:]
+            # Offset counts backward from the newest line within the bounded tail.
+            end = max(0, len(lines) - offset)
+            result["logs"].append(
+                {
+                    "name": name,
+                    "available": True,
+                    "bytes": size,
+                    "lines": lines[max(0, end - limit) : end],
+                    "offset": offset,
+                    "next_offset": offset + limit if end > limit else None,
+                    "older_bytes_omitted": start,
+                    "order": "oldest_to_newest",
+                }
+            )
+    elif view in {"report", "evidence"}:
+        report = report_export.session_report(session_id)
+        if view == "report":
+            result = {**report, "steps": page(report.get("steps", []), offset, limit)}
+        else:
+            index = evidence_service.export_index(
+                session_id, report, report_export.session_videos(report)
+            )
+            result = {
+                **index,
+                **{
+                    key: page(index.get(key, []), offset, limit)
+                    for key in ("steps", "calls", "recordings", "android_videos", "files")
+                },
+            }
+    else:
+        raise ValueError("view 需为 report / usage / requests / logs / evidence")
+    return public_value({"session_id": session_id, **result})

apps/admin_console/services/task_queue_service.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import asyncio
+import json
 from datetime import datetime
 import logging
 import os
@@ -24,12 +25,6 @@ import time
 from typing import Any
 import uuid
 
-try:
-    from admin_console.core.state import state
-    from admin_console.database.repositories.session_repository import session_repo
-    from admin_console.services import worker_process_io
-    from admin_console.services.media_service import media_service
-except ImportError:
 from apps.admin_console.core.state import state
 from apps.admin_console.database.repositories.session_repository import session_repo
 from apps.admin_console.services import worker_process_io
@@ -275,6 +270,11 @@ class TaskQueueService:
         ]
         for item in removed_items:
             DeviceExecutionLock.cancel_reservation(item.get("queue_ticket"))
+        from artemis.plugins.store import PluginStore
+
+        # A stop request may already have removed this row. The worker finalizer
+        # must still release the server-owned version reservation after exit.
+        PluginStore().release(str(session_id))
         state.queue_items = [
             t
             for t in state.queue_items
@@ -302,7 +302,7 @@ class TaskQueueService:
         normalized = current_status.lower().strip() if isinstance(current_status, str) else None
         if normalized == "success":
             return "completed", True
-        if normalized in {"completed", "failed", "cancelled"}:
+        if normalized in {"completed", "failed", "cancelled", "interrupted"}:
             return normalized, False
         return ("completed" if returncode == 0 else "failed"), True
 
@@ -388,10 +388,9 @@ class TaskQueueService:
     @classmethod
     def _dispatch_pending_tasks(cls) -> None:
         """Launch every pending task admissible under the current concurrency limit."""
+        cls._sync_control_states()
         limit = cls._concurrency_limit()
         state.prune_finished_runs()
-        if limit == 1 and state.is_running:
-            return
 
         # Dispatched-but-not-yet-spawned runs are only visible as queue items in
         # "running" state, so admission must count those too -- active_runs alone
@@ -410,12 +409,17 @@ class TaskQueueService:
                 and item.get("pid") not in active_pids
                 for item in in_flight
             )
-            capacity = limit - len(state.active_runs) - starting
+            capacity = (
+                limit - sum(not r.get("parked") for r in state.active_runs.values()) - starting
+            )
             if capacity <= 0:
                 return
         busy_devices = state.busy_device_ids | {
             cls._task_target(i).lock_key for i in in_flight if i.get("device_serial")
         }
+        busy_resources = {r for item in in_flight for r in item.get("plugin_resources", [])}
+        for owner in DeviceExecutionLock.get_active_owners().values():
+            busy_resources.update(owner.resource_ids)
         dispatched_any = False
         loop = asyncio.get_running_loop()
         for item in list(state.queue_items):
@@ -428,19 +432,27 @@ class TaskQueueService:
 
             device = item.get("device_serial")
             target = cls._task_target(item)
+            resources = set(item.get("plugin_resources", []))
+            if resources & busy_resources:
+                continue
             if limit == 0:
                 # A task without a resolved device may bind to any serial, so it
                 # only launches on an otherwise idle scheduler; the device lock
                 # then allocates freely without contending against active runs.
-                if device is None and (state.active_runs or in_flight or dispatched_any):
+                if device is None and (
+                    any(not r.get("parked") for r in state.active_runs.values())
+                    or in_flight
+                    or dispatched_any
+                ):
                     continue
                 if device is not None and target.lock_key in busy_devices:
                     continue
-            elif limit > 1 and device is not None and target.lock_key in busy_devices:
+            elif limit >= 1 and device is not None and target.lock_key in busy_devices:
                 # A second worker for this device would wait on its lock.
                 continue
 
             item["status"] = "running"
+            busy_resources.update(resources)
             dispatched_any = True
             if device is not None:
                 busy_devices.add(target.lock_key)
@@ -454,6 +466,42 @@ class TaskQueueService:
                 if capacity <= 0:
                     break
 
+    @classmethod
+    def _sync_control_states(cls) -> None:
+        from artemis.runtime.task_control import read_state
+        from apps.admin_console.services.automation_store import store
+
+        for item in state.queue_items:
+            if not isinstance(item, dict):
+                continue
+            sid = str(item.get("session_id") or "")
+            run = state.active_runs.get(sid)
+            if not run:
+                continue
+            status = read_state(sid).get("status")
+            if status not in {"running", "task_paused", "resuming"}:
+                continue
+            old = item.get("status")
+            if status == old:
+                continue
+            parked = status in {"task_paused", "resuming"}
+            now = time.time()
+            if parked and not run.get("parked"):
+                run["non_execution_started_at"] = now
+            elif not parked and run.get("parked"):
+                run["non_execution_seconds"] = (
+                    float(run.get("non_execution_seconds") or 0)
+                    + now
+                    - float(run.pop("non_execution_started_at", now))
+                )
+            run["parked"] = parked
+            item["status"] = status
+            job = store.job(sid)
+            if job and job.get("status") != "finalizing":
+                job["status"] = status
+                store.save_job(job)
+            cls._broadcast_event("task_status", {"session_id": sid, "status": status})
+
     @classmethod
     def _begin_task_run(
         cls,
@@ -531,7 +579,42 @@ class TaskQueueService:
             env["ARTEMIS_SESSION_ID"] = str(sess_id)
         env["ARTEMIS_TASK_INGRESS"] = str(task_item.get("ingress", "frontend"))
         env["ARTEMIS_TASK_WORKER"] = "1"
+        from artemis.exploration.store import explorations
+
+        exploration_run = explorations.run(str(sess_id)) if sess_id else None
+        env.pop("ARTEMIS_EXPLORATION_ID", None)
+        if exploration_run and exploration_run["kind"] == "exploration":
+            env["ARTEMIS_EXPLORATION_ID"] = exploration_run["exploration_id"]
+        overrides = task_item.get("runtime_overrides") or {}
+        if overrides.get("recording") is not None:
+            env["ARTEMIS_TASK_RECORDING_CONFIG"] = json.dumps(overrides["recording"])
+        else:
+            env.pop("ARTEMIS_TASK_RECORDING_CONFIG", None)
+        for option, variable in {
+            "model": "OPENAI_MODEL",
+            "reasoning_effort": "OPENAI_REASONING_EFFORT",
+            "request_timeout_seconds": "OPENAI_TIMEOUT_SECONDS",
+            "max_retries": "OPENAI_MAX_RETRIES",
+            "retry_interval_seconds": "OPENAI_RETRY_INTERVAL_SECONDS",
+            "record_video": "ARTEMIS_RECORDING_ENABLED",
+            "analyze_video": "ARTEMIS_VIDEO_ENABLED",
+            "observation_timeout_seconds": "ARTEMIS_OBSERVATION_TIMEOUT_SECONDS",
+            "observation_max_attempts": "ARTEMIS_OBSERVATION_MAX_ATTEMPTS",
+            "device_command_timeout_seconds": "ARTEMIS_DEVICE_COMMAND_TIMEOUT_SECONDS",
+            "action_failure_attempts": "ARTEMIS_ACTION_FAILURE_ATTEMPTS",
+        }.items():
+            if overrides.get(option) is not None:
+                env[variable] = (
+                    str(overrides[option]).lower()
+                    if isinstance(overrides[option], bool)
+                    else str(overrides[option])
+                )
         target.endpoint.apply_to_environment(env)
+        if "plugins" in overrides:
+            env["ARTEMIS_TASK_PLUGINS"] = json.dumps(overrides["plugins"])
+        else:
+            env.pop("ARTEMIS_TASK_PLUGINS", None)
+        env["ARTEMIS_EXECUTION_TARGET"] = overrides.get("execution_target") or "android"
         env[DeviceExecutionLock.LOCK_SCOPE_ENV] = target.lock_scope
         queue_ticket = task_item.get("queue_ticket")
         if queue_ticket:
@@ -562,7 +645,7 @@ class TaskQueueService:
         if app_path:
             cmd.extend(["--app-path", str(app_path)])
         device_serial = task_item.get("device_serial")
-        if device_serial:
+        if device_serial and env["ARTEMIS_EXECUTION_TARGET"] != "computer":
             cmd.extend(["--device-serial", str(device_serial)])
             env["ADB_DEVICE_SERIAL"] = str(device_serial)
         from apps.admin_console.services.automation_store import store
@@ -573,6 +656,83 @@ class TaskQueueService:
             env["RESULTS_OUTPUT_PATH"] = str(TRACES_PATH / str(sess_id) / "result.json")
         return cmd, env
 
+    @classmethod
+    async def _await_worker_with_budget(cls, proc, task_item: dict, run_key: str) -> int:
+        from artemis.runtime.task_limits import ExecutionBudget, task_timeout
+        from artemis.config.settings import settings
+        from artemis.runtime.task_result import save_failure_result, read_worker_outcome
+        from artemis.runtime.task_control import read_state
+        from apps.admin_console.services.automation_store import store
+
+        sid = str(task_item.get("session_id") or run_key)
+        job = store.job(sid)
+        batch = store.get("batch", job["batch_id"]) if job and job.get("batch_id") else None
+        limit = task_timeout(task_item, job, batch)
+        budget = ExecutionBudget(limit)
+        from artemis.exploration.store import explorations
+
+        exploration_run = explorations.run(sid)
+        is_exploration = bool(exploration_run and exploration_run["kind"] == "exploration")
+        exploration_finalized = False
+        cleanup_limit = (task_item.get("runtime_overrides") or {}).get(
+            "worker_cleanup_timeout_seconds"
+        ) or settings.ARTEMIS_WORKER_CLEANUP_TIMEOUT_SECONDS
+        finalized_at = None
+        waiter = asyncio.create_task(cls._wait_for_worker_process(proc))
+        try:
+            while True:
+                done, _ = await asyncio.wait({waiter}, timeout=0.25)
+                if waiter in done:
+                    return waiter.result()
+                outcome = read_worker_outcome(sid)
+                if outcome:
+                    if exploration_run and not exploration_finalized:
+                        explorations.finish_run(sid, {**outcome, "status": outcome["task_status"]})
+                        exploration_finalized = True
+                    finalized_at = finalized_at if finalized_at is not None else time.monotonic()
+                    if time.monotonic() - finalized_at >= cleanup_limit:
+                        logger.warning(
+                            "%s: task finalized but worker cleanup exceeded %s seconds; reclaiming process",
+                            sid,
+                            cleanup_limit,
+                        )
+                        await cls._terminate_worker_process(proc)
+                        return await waiter
+                    continue  # Final result is authoritative; only bound resource cleanup now.
+                cls._sync_control_states()
+                if is_exploration:
+                    # The worker owns its pausable budget so it survives a UI server restart.
+                    continue
+                run = state.active_runs.get(run_key, {})
+                remaining = budget.update(
+                    excluded=bool(run.get("parked")) or read_state(sid).get("status") == "queued"
+                )
+                if remaining > 0:
+                    continue
+                error = f"任务执行超时（{limit} 秒，不含暂停/排队时间）"
+                task_item.update(timed_out=True, error=error)
+                logger.error("%s: %s", sid, error)
+                try:
+                    latest = store.job(sid)
+                    if latest:
+                        latest.update(timed_out=True, error=error)
+                        store.save_job(latest)
+                except Exception:
+                    logger.exception(
+                        "Cannot persist timeout for %s; terminating worker anyway", sid
+                    )
+                await cls._terminate_worker_process(proc)
+                returncode = await waiter
+                try:
+                    await asyncio.to_thread(save_failure_result, sid, error, code="task_timeout")
+                except OSError:
+                    logger.exception("Cannot save timeout report for %s", sid)
+                return returncode
+        finally:
+            if not waiter.done():
+                waiter.cancel()
+            await asyncio.gather(waiter, return_exceptions=True)
+
     @classmethod
     def _register_worker_run(
         cls,
@@ -664,6 +824,16 @@ class TaskQueueService:
             returncode,
             manual_stop,
         )
+        from apps.admin_console.services.automation_store import store
+
+        if (store.job(str(sess_id)) or {}).get("timed_out"):
+            new_status, should_persist = "failed", True
+        if (
+            state.is_shutting_down
+            and not manual_stop
+            and current_status not in {"completed", "success", "failed", "interrupted"}
+        ):
+            new_status, should_persist = "interrupted", True
         if should_persist:
             if session_repo.update_session_status(sess_id, new_status, time.time()):
                 print(f"[QueueWorker] Updated session {sess_id} status to '{new_status}'")
@@ -714,6 +884,10 @@ class TaskQueueService:
         registered at recording start; remuxing that file is all that is
         needed to publish the full video.
         """
+        from apps.admin_console.services.evidence_service import task_root
+        from artemis.plugins.recording import recover_interrupted_recordings
+
+        await recover_interrupted_recordings(task_root(str(sess_id)))
         rec_info = session_repo.get_video_recording_for_session(sess_id)
         rec_status = (rec_info or {}).get("status")
         if rec_status == "ready":
@@ -859,6 +1033,9 @@ class TaskQueueService:
             if not isinstance(goal, str) or not goal.strip():
                 raise ValueError("Queued task must contain a non-empty string goal.")
             cls._begin_task_run(task_item, run_key, sess_id, goal, profile)
+            from apps.admin_console.services.execution_config import persist_execution_config
+
+            await asyncio.to_thread(persist_execution_config, task_item)
 
             target = cls._task_target(task_item)
             cmd, env = cls._build_worker_invocation(
@@ -880,7 +1057,7 @@ class TaskQueueService:
             await cls._terminate_if_cancelled_during_launch(run_key, sess_id, proc)
 
             # 3. Await subprocess completion
-            returncode = await cls._wait_for_worker_process(proc)
+            returncode = await cls._await_worker_with_budget(proc, task_item, run_key)
             print(f"[QueueWorker] Task [{sess_id}] exited with returncode {returncode}")
 
             manual_stop = run_key in state.manually_stopped_run_ids or bool(
@@ -914,8 +1091,34 @@ class TaskQueueService:
                         f"[QueueWorker] Could not persist failure status for [{sess_id}]"
                     )
         finally:
+            try:
                 await cls._finish_output_forwarder(output_task)
-            # 5. Clean up the finished task and release this run's scheduling slot
+                if sess_id:
+                    from artemis.exploration.store import explorations
+                    from artemis.runtime.task_result import read_worker_outcome
+
+                    exploration_run = explorations.run(str(sess_id))
+                    if exploration_run and exploration_run.get("outcome") is None:
+                        from artemis.runtime.case_result import read_case_result
+
+                        outcome = read_worker_outcome(str(sess_id)) or {}
+                        session = session_repo.get_session_by_id(str(sess_id)) or {}
+                        status = (
+                            outcome.get("task_status") or session.get("status") or "interrupted"
+                        )
+                        explorations.finish_run(
+                            str(sess_id),
+                            {
+                                **outcome,
+                                "status": status,
+                                "test_result": outcome.get("test_result")
+                                or read_case_result(str(sess_id), status),
+                            },
+                        )
+            except Exception:
+                logger.exception("Cannot finalize exploration metadata for %s", sess_id)
+            finally:
+                # Metadata/storage failure must never leak a scheduling slot.
                 cls._release_run_slot(sess_id, run_key, proc)
 
     @classmethod
@@ -1069,6 +1272,8 @@ class TaskQueueService:
         conversation_id: str | None = None,
         verification_level: str | None = None,
         explorer_mode: str | None = None,
+        runtime_overrides: dict[str, Any] | None = None,
+        display_name: str | None = None,
     ) -> dict[str, Any]:
         """Enqueues one or more goals and wakes up the background worker.
 
@@ -1076,6 +1281,12 @@ class TaskQueueService:
         forwarded to the worker as ``--verification-level`` / ``--explorer-pro-mode``;
         they are normalised here so the queue item and the CLI see one spelling.
         """
+        if runtime_overrides:
+            profile = runtime_overrides.get("profile") or profile
+            device_serial = runtime_overrides.get("device_serial") or device_serial
+            locked_app_package = runtime_overrides.get("locked_app_package") or locked_app_package
+            verification_level = runtime_overrides.get("verification_level") or verification_level
+            explorer_mode = runtime_overrides.get("explorer_mode") or explorer_mode
         verification_level = (
             str(verification_level).strip().lower() or None if verification_level else None
         )
@@ -1092,11 +1303,40 @@ class TaskQueueService:
         if duplicate_response is not None:
             return duplicate_response
 
-        rejection_response = await cls._reject_unavailable_device(device_serial)
+        execution_target = (runtime_overrides or {}).get("execution_target") or "android"
+        from pydantic import TypeAdapter
+        from artemis.plugins.models import PluginSelection
+        from artemis.plugins.store import PluginStore
+
+        choices = (runtime_overrides or {}).get("plugins")
+        try:
+            pinned = PluginStore().pin(
+                TypeAdapter(list[PluginSelection]).validate_python(choices)
+                if choices is not None
+                else None,
+                execution_target,
+            )
+        except ValueError:
+            return {
+                "status": "error",
+                "message": "插件未就绪、未授权或不支持所选任务目标，请检查插件配置",
+            }
+        runtime_overrides = {
+            **(runtime_overrides or {}),
+            "execution_target": execution_target,
+            "plugins": [p.model_dump() for p in pinned],
+        }
+        rejection_response = (
+            await cls._reject_unavailable_device(device_serial)
+            if execution_target != "computer"
+            else None
+        )
         if rejection_response is not None:
             return rejection_response
 
         single_session_id = session_id if (session_id and len(goals) == 1) else None
+        if execution_target == "computer":
+            device_serial = "computer:local"
         if not device_serial:
             # Device enumeration may block on ADB.
             from artemis.runtime import device_pool
@@ -1105,6 +1345,21 @@ class TaskQueueService:
                 device_serial = await device_pool.select_device_async()
             except Exception:
                 device_serial = None
+        from apps.admin_console.services.execution_config import execution_config
+
+        config_snapshot = execution_config(
+            {
+                "profile": profile,
+                "device_serial": device_serial,
+                "expected_output": expected_output,
+                "enable_outputter": enable_outputter,
+                "verification_level": verification_level,
+                "explorer_mode": explorer_mode,
+                "locked_app_package": locked_app_package,
+                "app_path": app_path,
+                "runtime_overrides": runtime_overrides,
+            }
+        )
         for i, goal in enumerate(goals):
             task_item = cls._create_queue_item(
                 goal,
@@ -1123,6 +1378,24 @@ class TaskQueueService:
                 verification_level=verification_level,
                 explorer_mode=explorer_mode,
             )
+            if runtime_overrides:
+                task_item["runtime_overrides"] = runtime_overrides
+            task_item["execution_config"] = config_snapshot
+            if pinned:
+                try:
+                    task_item["plugin_resources"] = PluginStore().reserve(
+                        task_item["session_id"], pinned, execution_target
+                    )
+                except (ValueError, OSError):
+                    DeviceExecutionLock.cancel_reservation(task_item.get("queue_ticket"))
+                    for previous in enqueued_tasks:
+                        cls._remove_task(previous["session_id"])
+                    return {
+                        "status": "error",
+                        "message": "插件配置在入队前发生变化，请检查后重新提交",
+                    }
+            if display_name:
+                task_item["display_name"] = display_name
             state.queue_items.append(task_item)
             enqueued_tasks.append(task_item)
             cls._broadcast_startup_progress(
@@ -1231,6 +1504,7 @@ class TaskQueueService:
         """Terminate all active device owners and clear pending queue submissions."""
         from apps.admin_console.services.automation_store import store
         from apps.admin_console.services.automation_service import on_terminal, spawn
+        from artemis.plugins.store import PluginStore
 
         for batch in store.records("batch"):
             if batch["status"] == "running":
@@ -1240,6 +1514,7 @@ class TaskQueueService:
         for item in state.queue_items:
             if isinstance(item, dict) and item.get("status") != "running":
                 DeviceExecutionLock.cancel_reservation(item.get("queue_ticket"))
+                PluginStore().release(str(item.get("session_id")))
                 if store.job(str(item.get("session_id"))):
                     spawn(on_terminal(item, "cancelled"))
         state.clear_queue()
@@ -1563,7 +1838,9 @@ class TaskQueueService:
         if stopped_item and stopped_item.get("status") != "running":
             from apps.admin_console.services.automation_store import store
             from apps.admin_console.services.automation_service import on_terminal, spawn
+            from artemis.plugins.store import PluginStore
 
+            PluginStore().release(str(stopped_session_id))
             if store.job(str(stopped_session_id)):
                 spawn(on_terminal(stopped_item, "cancelled"))
         if stopped_item and not reservation_cancelled:
@@ -1677,6 +1954,14 @@ class TaskQueueService:
         recordings published.
         """
         recovered = 0
+        from artemis.config import TRACES_PATH
+        from artemis.plugins.recording import recover_interrupted_recordings
+
+        # This hook runs in the startup worker thread, outside the server event loop.
+        for manifest in TRACES_PATH.glob("*/evidence/recordings/recordings.json"):
+            root = manifest.parent.parent.parent
+            if asyncio.run(recover_interrupted_recordings(root)):
+                recovered += 1
         try:
             pending = session_repo.get_unfinalized_video_recordings()
         except Exception as exc:

apps/showcase_ui/AGENTS.md

@@ -2,12 +2,14 @@
 
 ## Local Purpose
 
-本目录维护简体中文 Angular 网页，包括设备与任务操作、模型配置、报告、批量导入和历史检索。
+本目录维护简体中文 Angular 网页，包括 Flash/Pro/探索、人工示范、插件、精选任务、模型配置、批次、历史、报告与跨端证据。
 
 ## Inherits
 
 继承仓库根目录 AGENTS.md；本文件仅补充前端规则，当前用户指令优先。
 
+修改 UI 前阅读根目录 [DESIGN.md](../../DESIGN.md) 的视觉与交互约定；本应用没有独立的局部 DESIGN.md。SDK 的同名文件描述远程协议，不是网页设计来源。
+
 ## Local Commands
 
 在本目录执行，包管理器使用 npm 和现有 package-lock.json：
@@ -34,6 +36,11 @@ npm run build
 - 使用相邻的 .spec.ts 测试交互和请求数据；不要只测试实现细节。
 - 后端字段调整需同时核对类型、请求和响应；设备序列号、报告格式、批次编号等不能在前端猜测。
 - UI 有明显变化时在浏览器验证实际布局和关键操作，构建通过不能代替可读性检查。
+- 工作区底部输入框服务当前任务；探索建议只填入草稿，用户发送后再提交。用例编辑和配置按需展开，复跑固定保存版本。
+- 主模型、兜底模型及精选任务复用现有模型选择组件；收藏使用已保存的独立配置，设备离线时给出可操作的选择提示。
+- 星标仅占标题区域，后续状态行正常左对齐；人工示范来源与按钮保持等高、居中，窄屏可换行。执行、等待回答、示范准备/录制/分析、最终验收及未验证各有明确文字。
+- 电脑步骤使用电脑证据，录像缺口明确展示。播放器按真实来源和时间定位，跳转保留原播放/暂停状态。
+- 列表后台刷新保留选中项、滚动及折叠状态，不用反复插入的加载行挤动布局；实时连接沿用现有共享机制和页面可见性管理。
 
 ## Local Risks
 

apps/showcase_ui/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "artemis-showcase-ui",
-  "version": "1.0.0",
+  "version": "2.0.1",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "artemis-showcase-ui",
-      "version": "1.0.0",
+      "version": "2.0.1",
       "dependencies": {
         "@angular/cdk": "^22.1.5",
         "@angular/common": "^22.1.4",

apps/showcase_ui/package.json

@@ -1,6 +1,6 @@
 {
   "name": "artemis-showcase-ui",
-  "version": "1.0.0",
+  "version": "2.0.1",
   "scripts": {
     "ng": "ng",
     "start": "ng serve --proxy-config proxy.conf.json",

apps/showcase_ui/src/app/app.routes.ts

@@ -22,6 +22,7 @@ import { LegacyWorkspaceComponent } from './pages/legacy-workspace/legacy-worksp
 export const routes: Routes = [
   { path: '', component: HomeComponent },
   { path: 'workspace', component: WorkspaceComponent },
+  { path: 'plugins', loadComponent: () => import('./pages/plugins/plugins.component').then(m => m.PluginsComponent) },
   { path: 'check', component: LegacyWorkspaceComponent },
   { path: '**', redirectTo: '' }
 ];

apps/showcase_ui/src/app/components/agent-stream/agent-stream.component.html

@@ -44,7 +44,7 @@ limitations under the License.
       </button>
       <button #automationOrigin="cdkOverlayOrigin" cdkOverlayOrigin class="vscode-tab-btn" [class.active]="isAutomationOpen()" (click)="isAutomationOpen.set(!isAutomationOpen())" title="报告与批量" aria-label="报告与批量"><span class="material-symbols-outlined tab-icon">assignment</span><span>报告与批量</span></button>
     </div>
-    <ng-template cdkConnectedOverlay [cdkConnectedOverlayOrigin]="automationOrigin" [cdkConnectedOverlayOpen]="isAutomationOpen()" [cdkConnectedOverlayHasBackdrop]="true" cdkConnectedOverlayBackdropClass="cdk-overlay-transparent-backdrop" (backdropClick)="isAutomationOpen.set(false)" [cdkConnectedOverlayPositions]="dropdownPositions"><div class="vscode-dropdown-menu" style="width:min(560px,calc(100vw - 24px));max-height:85vh;overflow:auto;padding:20px"><div class="popover-title-header" style="justify-content:space-between;margin-bottom:6px"><span>报告与自动化</span><button type="button" (click)="isAutomationOpen.set(false)" aria-label="关闭报告与自动化" style="border:0;background:transparent;color:#64748b;cursor:pointer;padding:4px"><span class="material-symbols-outlined">close</span></button></div><app-automation-panel /></div></ng-template>
+    <ng-template cdkConnectedOverlay [cdkConnectedOverlayOrigin]="automationOrigin" [cdkConnectedOverlayOpen]="isAutomationOpen()" [cdkConnectedOverlayHasBackdrop]="true" cdkConnectedOverlayBackdropClass="cdk-overlay-transparent-backdrop" (backdropClick)="isAutomationOpen.set(false)" [cdkConnectedOverlayPositions]="dropdownPositions"><div class="vscode-dropdown-menu" style="width:min(560px,calc(100vw - 24px));max-height:85vh;overflow:auto;padding:20px"><div class="popover-title-header" style="justify-content:space-between;margin-bottom:6px"><span>报告与自动化</span><button type="button" (click)="isAutomationOpen.set(false)" aria-label="关闭报告与自动化" style="border:0;background:transparent;color:#64748b;cursor:pointer;padding:4px"><span class="material-symbols-outlined">close</span></button></div><app-automation-panel (queueRequested)="openBatchQueue()" /></div></ng-template>
 
     <!-- Dropdown Popover Panel for Task Queue -->
     <ng-template
@@ -61,6 +61,8 @@ limitations under the License.
           <span class="material-symbols-outlined">list_alt</span>
           <span>任务队列</span>
         </div>
+        <app-task-batch-filter />
+        @if (deleteError()) { <p class="item-error" role="alert">{{ deleteError() }}</p> }
         <div class="dropdown-section-label">
           <span class="material-symbols-outlined icon-bolt">bolt</span>
           <span>当前队列 ({{ activeQueue().length }})</span>
@@ -76,15 +78,19 @@ limitations under the License.
             class="dropdown-item-card"
             [class.selected]="agentService.currentSessionId() === session.session_id"
             [class.running]="getTaskStatus(session) === 'running'"
+            [ngClass]="getTaskStatus(session)"
             (click)="selectTask(session.session_id, $event)"
             >
             <div class="item-header">
+              <div class="task-status-group">
               <span class="task-badge" [ngClass]="getTaskStatus(session)">
                 @if (getTaskStatus(session) === 'running') {
                   <span class="status-dot"></span>
                 }
                 {{ getTaskStatus(session) | uiLabel }}
               </span>
+              <app-test-result-badge [result]="session.test_result" [executionStatus]="getTaskStatus(session)" [compact]="true" />
+              </div>
               <div class="item-header-actions">
                 @if (getDeviceSerial(session)) {
                   <span class="item-device" [title]="'Device: ' + getDeviceSerial(session)">
@@ -96,7 +102,7 @@ limitations under the License.
                   <span class="item-time-divider">·</span>
                 }
                 <span class="item-time">{{ session.start_time * 1000 | date:'shortTime' }}</span>
-                @if (getTaskStatus(session) === 'running' || getTaskStatus(session) === 'paused') {
+                @if (!agentService.canDeleteSession(session) && ['running', 'paused', 'task_paused', 'resuming'].includes(getTaskStatus(session))) {
                   <button
                     type="button"
                     class="btn-stop-card"
@@ -107,7 +113,7 @@ limitations under the License.
                     <span class="material-symbols-outlined">stop_circle</span>
                   </button>
                 }
-                @if (getTaskStatus(session) !== 'running' && getTaskStatus(session) !== 'paused') {
+                @if (agentService.canDeleteSession(session)) {
                   <button
                     type="button"
                     class="btn-delete-card"
@@ -120,12 +126,15 @@ limitations under the License.
                 }
               </div>
             </div>
-            <div class="item-goal" title="{{ session.initial_goal }}">
+            <div class="item-goal" [title]="session.initial_goal || ''">
-              {{ session.initial_goal || '未命名任务' }}
+              {{ session.display_name || session.initial_goal || '未命名任务' }}
             </div>
+            @if (session.error && getTaskStatus(session) === 'failed') {
+              <div class="item-error" [title]="session.error">{{session.error}}</div>
+            }
             <div class="item-footer">
-              <span class="item-id" [title]="session.session_id">ID: {{ session.session_id }}</span>
+              <app-task-id-copy [taskId]="session.session_id" />
-              <span class="item-date">{{ session.start_time * 1000 | date:'MM/dd/yy' }}</span>
+              <span class="item-date">{{ session.start_time * 1000 | date:'MM/dd/yy' }} · 已用 {{ formatTaskDuration(session.start_time, session.end_time) }}</span>
             </div>
           </div>
         }
@@ -174,6 +183,9 @@ limitations under the License.
         <!-- Rich Notes Content Viewer -->
         @if (noteKeys().length > 0) {
           <div class="dropdown-notes-content">
+            @if (agentService.selectedNoteKey() !== 'task_plan.md') {
+              <app-report-markdown [content]="currentNoteContent()" />
+            } @else {
             @if (parsedNote().title) {
               <h3 class="note-title">{{ parsedNote().title }}</h3>
             }
@@ -324,6 +336,7 @@ limitations under the License.
                 }
               </div>
             }
+            }
           </div>
         }
       </div>
@@ -331,7 +344,7 @@ limitations under the License.
   </header>
 
   <!-- Floating Top-Right Badges & Toolbar (Pinned overlay, hidden in empty/initial state) -->
-  @if (consolidatedBlocks().length > 0 || showPlanningLoader()) {
+  @if (consolidatedBlocks().length > 0 || showPlanningLoader() || agentService.currentSession()) {
     <div class="stream-floating-toolbar">
       <!-- Active Architecture Chip: click opens the run info popover -->
       @if (agentService.viewedModel(); as model) {
@@ -393,7 +406,7 @@ limitations under the License.
                 @if (runInfoContextPercent() !== null) {
                   <span class="run-info-value">
                     {{ runInfoContextPercent() }}%
-                    <span class="run-info-sub-inline">/ {{ formatRunTokens(runUsage()?.operator_context_window_tokens) }}</span>
+                    <span class="run-info-sub-inline">/ {{ formatRunTokens(runInfoContextWindow()) }}</span>
                   </span>
                 } @else {
                   <span class="run-info-value is-muted">—</span>
@@ -417,7 +430,9 @@ limitations under the License.
                 <span class="run-info-label">屏幕识别</span>
                 <span class="run-info-value">{{ runTuningLabel('explore', runInfoTuning()?.explorer_mode) }}</span>
               </div>
-            } @else {
+            } @else if (runInfoMode() === 'explore') {
+              <div class="run-info-note">探索支持对话指导与人工示范，模型和时限使用独立的探索设置。</div>
+            } @else if (runInfoMode() === 'flash') {
               <div class="run-info-note">Flash 无需额外设置。</div>
             }
           </div>
@@ -492,11 +507,14 @@ limitations under the License.
           </span>
         }
       </button>
+      <ng-content select="[task-guidance]"></ng-content>
     </div>
   }
 
   <!-- Event Logs Section -->
   <main class="stream-logs-content">
+    <ng-content select="[task-summary]"></ng-content>
+    <app-llm-request-status [sessionId]="agentService.currentSessionId()" [running]="agentService.isCurrentSessionRunning()" (activeChange)="hasModelRequest.set($event)" />
     @if (agentService.isSessionContentLoading()) {
       <div class="session-loading-state" aria-live="polite">
         <span class="material-symbols-outlined spin-animation">progress_activity</span>
@@ -504,7 +522,7 @@ limitations under the License.
       </div>
     }
 
-    @if (!agentService.isSessionContentLoading() && consolidatedBlocks().length === 0 && !showPlanningLoader() && startupWorkItems().length === 0) {
+    @if (!agentService.isSessionContentLoading() && phases().length === 0 && !showPlanningLoader() && !hasModelRequest() && startupWorkItems().length === 0) {
       <div class="empty-state">
         <div class="empty-icon"><span class="material-symbols-outlined">sensors</span></div>
         <h3>尚未选择执行记录</h3>
@@ -543,8 +561,21 @@ limitations under the License.
           </div>
         }
         @for (phase of phases(); track trackPhase($index, phase); let first = $first) {
+          @if(phase.demonstration; as demo) {
+            <article class="exploration-message demonstration"><strong>人工示范 · {{demo.source === 'android' ? '安卓' : demo.source === 'computer' ? '电脑' : '安卓 + 电脑'}} · {{demoLabel(demo.status)}}</strong>
+              @if(demo.analysis){<div class="markdown-content" [innerHTML]="renderMarkdown(demo.analysis.summary, phase.id)"></div>}
+              @if(demo.error){<p class="error">{{demo.error}}</p>}
+              @for(track of demo.media; track $index){@for(segment of track.segments; track segment.file){<details><summary>{{track.kind === 'android' ? '安卓' : '电脑'}}示范录像 · 片段 {{$index + 1}}</summary><video controls preload="none" [src]="demoMediaUrl(demo.run_id,segment.file)"></video><a [href]="demoMediaUrl(demo.run_id,segment.file)" download>下载录像</a></details>}}
+            </article>
+          } @else if(phase.conversation; as message) {
+            <article class="exploration-message" [class.user]="message.role === 'user'" [class.assistant]="message.role === 'assistant'" [class.question]="pendingQuestion() === message.id" [class.system]="message.role === 'system'">
+              <small><span class="material-symbols-outlined">{{message.role === 'user' ? 'person' : message.role === 'assistant' ? 'auto_awesome' : 'info'}}</span>{{message.role === 'user' ? '你的指导' : message.role === 'assistant' ? '探索助手' : '探索记录'}} · {{message.created_at * 1000 | date:'HH:mm:ss'}} @if(pendingQuestion() === message.id){<b>等待你回答</b>}</small>
+              <div class="markdown-content" [innerHTML]="renderMarkdown(message.text, phase.id)"></div>
+              @if(message.recommendation){<div class="recommendation">建议回复：{{message.recommendation}} @if(pendingQuestion() === message.id){<button (click)="useRecommendation.emit(message.recommendation)">填入回复</button>}</div>}
+            </article>
+          } @else {
           <div class="phase-container">
-            @if (hasVisibleBlocks(phase)) {
+            @if (hasVisibleBlocks(phase) && phase.blocks[0]?.data?.phase !== 'outcome') {
               <div
                 class="phase-header"
                 [class.first-phase]="first && startupWorkItems().length === 0"
@@ -568,10 +599,20 @@ limitations under the License.
                       >
                       <!-- Unified Step / Stream Body -->
                       <div class="step-body">
+                        @if (block.data?.extra_metadata?.injected_instruction; as guidance) {
+                          <div class="guidance-step-note">
+                            <div class="guidance-step-title"><span class="material-symbols-outlined">forum</span>用户指导 · 已进入第 {{block.data.step_number}} 步</div>
+                            <p>{{ guidance }}</p>
+                            <small>{{ hasRecordedDeviceAction(block.data.action_taken) ? '本步设备动作见下方；是否完成目标需以实际屏幕为准。' : '本步没有记录设备动作。' }}</small>
+                          </div>
+                        }
                         <!-- Check block (verifier lane): header, then the declared check items
                              and how each is judged; the chronological stream follows, and the
                              result rows close the block (see check-result below). -->
                         @if (isCheckerBlock(block)) {
+                          @if (block.data.phase === 'outcome' && block.data.test_result) {
+                            <app-test-result-card [result]="agentService.currentSession()?.test_result || block.data.test_result" [sessionId]="agentService.currentSessionId()" [executionStatus]="agentService.currentSession()?.status || ''" (reviewSaved)="agentService.applyCaseReview($event)" />
+                          } @else {
                           <div class="check-lane" [ngClass]="'tone-' + getCheckerTone(block)">
                             <div class="check-header">
                               <span class="stream-status-dot" [class.streaming]="isBlockRunning(block)"></span>
@@ -594,6 +635,7 @@ limitations under the License.
                             }
                           </div>
                           }
+                        }
                         <!-- Text, tool calls, and actions share one chronological timeline. -->
                         @if (getSortedStepEvents(block.data).length > 0) {
                           <div class="step-actions-container">
@@ -733,10 +775,8 @@ limitations under the License.
                                   <!-- Compact preview body (when collapsed, matches standard action card) -->
                                   @if (!isActionExpanded('action-' + block.id + '-' + itemIdx) && getReportStatusExplanation(item.data)) {
                                     <div class="action-card-body">
-                                      <div class="action-detail-row">
+                                      <app-report-markdown [content]="getReportStatusExplanation(item.data)" [compact]="true" />
-                                        <span class="detail-label">摘要：</span>
+                                      <button type="button" class="report-expand-button" (click)="toggleActionExpanded('action-' + block.id + '-' + itemIdx, $event)">查看完整报告</button>
-                                        <span class="detail-value">"{{ getReportStatusExplanation(item.data) }}"</span>
-                                      </div>
                                     </div>
                                   }
                                   <!-- Detailed expanded view (when clicked, matches standard action card structure) -->
@@ -746,7 +786,7 @@ limitations under the License.
                                         @if (getReportStatusExplanation(item.data); as expText) {
                                           <div class="grid-item full-width">
                                             <span class="grid-label">报告与发现</span>
-                                            <span class="grid-value report-markdown-body" [innerHTML]="renderMarkdown(expText, 'report-' + block.id + '-' + itemIdx)"></span>
+                                            <app-report-markdown [content]="expText" />
                                           </div>
                                         }
                                       </div>
@@ -789,6 +829,7 @@ limitations under the License.
                                   <div class="action-card-header" (click)="toggleActionExpanded('action-' + block.id + '-' + itemIdx, $event)">
                                     <span class="material-symbols-outlined action-icon" [class.is-failed]="isActionFailed(item.data, block.data)">{{ getActionIcon(item.data) }}</span>
                                     <span class="action-title">{{ getActionTitle(item.data) }}</span>
+                                    <button class="evidence-source" [title]="evidenceSourceLabel(item.data) + ' · 查看对应录像'" (click)="openEvidenceVideo(item.data, block.data, $event)">{{ evidenceSourceLabel(item.data) }}</button>
                                     @if (isActionFailed(item.data, block.data)) {
                                       <span class="action-status-badge is-failed">
                                         <span class="material-symbols-outlined status-icon">error</span>
@@ -1030,7 +1071,7 @@ limitations under the License.
                                   </div>
                                 }
                                 <!-- Generic Tool Calls (Text format for note tools, video analyzer, and other sub-agents) -->
-                                @if (shouldShowTool(item.data, block.data) && !isDeviceActionTool(item.data)) {
+                                @if (shouldShowTool(item.data, block.data) && !isEvidenceTool(item.data)) {
                                   <div class="step-generic-tool"
                                     [class.is-running]="isToolLineRunning(item, block)"
                                     [class.is-note-tool]="isNoteTool(item.data)"
@@ -1071,9 +1112,11 @@ limitations under the License.
                                     </div>
                                   </div>
                                 }
-                                <!-- Action Card format (For Device Actions & ADB Commands) -->
+                                <!-- Device actions and plugin calls share evidence cards with explicit sources. -->
-                                @if (shouldShowTool(item.data, block.data) && isDeviceActionTool(item.data)) {
+                                @if (shouldShowTool(item.data, block.data) && isEvidenceTool(item.data)) {
                                   <div class="step-android-action-card tool-action-card"
+                                    [class.plugin-tool-card]="getPluginToolView(item.data)"
+                                    [class.is-auxiliary]="getPluginToolView(item.data)?.auxiliary"
                                     [class.is-running]="isItemRunning(item, block)"
                                     [class.has-no-details]="isAdbCommandTool(item.data) ? !getAdbCommandLine(item.data) : (!getToolTargetText(item.data) && !getToolInputText(item.data) && !getToolCoords(item.data) && !getToolAnalysisText(item.data) && !getToolGenericDetails(item.data) && getToolExtraParams(item.data).length === 0)"
                                     [class.is-failed]="isToolFailed(item.data)"
@@ -1082,6 +1125,10 @@ limitations under the License.
                                     <div class="action-card-header" (click)="toggleActionExpanded('tool-' + block.id + '-' + itemIdx, $event)">
                                       <span class="material-symbols-outlined action-icon" [class.is-failed]="isToolFailed(item.data)">{{ getToolIcon(item.data) }}</span>
                                       <span class="action-title">{{ getToolTitle(item.data) }}</span>
+                                      <button class="evidence-source" [title]="evidenceSourceLabel(item.data) + ' · 查看对应证据'" (click)="openEvidenceVideo(item.data, block.data, $event)">{{ evidenceSourceLabel(item.data) }}</button>
+                                      @if (getPluginToolView(item.data); as plugin) {
+                                        <span class="plugin-call-kind">{{ plugin.kind }} · {{ plugin.status }}</span>
+                                      }
                                       @if (isToolFailed(item.data)) {
                                         <span class="action-status-badge is-failed">
                                           <span class="material-symbols-outlined status-icon">error</span>
@@ -1094,7 +1141,7 @@ limitations under the License.
                                       </span>
                                     </div>
                                     <!-- Compact preview body (when collapsed) -->
-                                    @if (!isActionExpanded('tool-' + block.id + '-' + itemIdx) && (isAdbCommandTool(item.data) ? getAdbCommandLine(item.data) : (getToolTargetText(item.data) || getToolInputText(item.data) || getToolCoords(item.data) || getToolAnalysisText(item.data) || getToolGenericDetails(item.data)))) {
+                                    @if (!getPluginToolView(item.data) && !isActionExpanded('tool-' + block.id + '-' + itemIdx) && (isAdbCommandTool(item.data) ? getAdbCommandLine(item.data) : (getToolTargetText(item.data) || getToolInputText(item.data) || getToolCoords(item.data) || getToolAnalysisText(item.data) || getToolGenericDetails(item.data)))) {
                                       <div class="action-card-body">
                                         <!-- ADB command details -->
                                         @if (isAdbCommandTool(item.data)) {
@@ -1176,7 +1223,21 @@ limitations under the License.
                                               </div>
                                             }
                                           }
-                                          @if (!isAdbCommandTool(item.data)) {
+                                          @if (getPluginToolView(item.data); as plugin) {
+                                            @for (param of plugin.args; track param.key) {
+                                              <div class="grid-item">
+                                                <span class="grid-label">{{ param.key }}</span>
+                                                <span class="grid-value code-pill">{{ param.value }}</span>
+                                              </div>
+                                            }
+                                            @if (plugin.result) {
+                                              <div class="grid-item full-width">
+                                                <span class="grid-label">实际结果</span>
+                                                <span class="grid-value plugin-result">{{ plugin.result }}</span>
+                                              </div>
+                                            }
+                                          }
+                                          @if (!isAdbCommandTool(item.data) && !getPluginToolView(item.data)) {
                                             @if (getToolTargetText(item.data)) {
                                               <div class="grid-item">
                                                 <span class="grid-label">目标</span>
@@ -1231,7 +1292,7 @@ limitations under the License.
                                             }
                                             @if (getStepPostImageUrl(block.data, item.data)) {
                                               <div class="screenshot-column">
-                                                <span class="screenshot-label">操作后</span>
+                                                <span class="screenshot-label">{{ getPluginToolView(item.data)?.auxiliary ? '观察截图' : '操作后' }}</span>
                                                 <div class="image-wrapper" (click)="openImageModal(getStepPostImageUrl(block.data, item.data)!, null, $event)">
                                                   <div class="img-relative-box">
                                                     <img [src]="getStepPostImageUrl(block.data, item.data)" alt="操作后截图" />
@@ -1256,7 +1317,7 @@ limitations under the License.
                           </div>
                         }
                         <!-- Check block result: one row per check item, keyed by its number above -->
-                        @if (hasCheckerResult(block)) {
+                        @if (hasCheckerResult(block) && !(block.data.phase === 'outcome' && block.data.test_result)) {
                           <div class="check-result" [ngClass]="'tone-' + getCheckerTone(block)">
                             @for (verdict of getCheckerVerdicts(block); track $index) {
                               <div class="check-verdict" [ngClass]="verdict.status">
@@ -1283,6 +1344,24 @@ limitations under the License.
             }
           </div>
           }
+        }
+        @if (!conversationMessages().length && awaitingGuidance().length) {
+          <div class="guidance-queue" aria-label="待处理用户指导">
+            @for (item of awaitingGuidance(); track item.id) {
+              <div class="guidance-queue-item">
+                <strong>用户指导 · {{ guidanceStatusLabel(item) }}</strong>
+                <p>{{ item.instruction }}</p>
+              </div>
+            }
+          </div>
+        }
+        @if (terminalFailureReason(); as reason) {
+          <div class="session-failure-summary" role="status">
+            <strong>任务执行失败 · 原因</strong>
+            <p>{{ reason }}</p>
+            <button type="button" (click)="isAutomationOpen.set(true)">查看完整报告</button>
+          </div>
+        }
         <!-- Outputter Post-Execution Synthesis Card (Rendered ONCE at the end of stream) -->
         @if (outputterReport(); as outReport) {
           <div class="outputter-synthesis-stream-card">
@@ -1299,62 +1378,7 @@ limitations under the License.
               </div>
             </div>
             <div class="synthesis-card-preview">
-              @if (parsedOutputReport(); as parsedOut) {
+              <app-report-markdown [content]="outReport" />
-                <div class="parsed-note-container">
-                  @if (parsedOut.title) {
-                    <h3 class="parsed-note-title">{{ parsedOut.title }}</h3>
-                  }
-                  @if (parsedOut.milestones.length > 0) {
-                    <div class="parsed-milestones">
-                      @for (m of parsedOut.milestones; track trackMilestone($index, m)) {
-                        <div class="milestone-item" [class]="'status-' + m.type">
-                          <span class="material-symbols-outlined status-icon">
-                            {{ m.type === 'checked' ? 'check_circle' : (m.type === 'progress' ? 'pending' : 'radio_button_unchecked') }}
-                          </span>
-                          <span class="milestone-text">
-                            @for (seg of m.segments; track seg) {
-                              @if (seg.bold) {
-                                <strong>{{ seg.text }}</strong>
-                              }
-                              @if (seg.code) {
-                                <code>{{ seg.text }}</code>
-                              }
-                              @if (!seg.bold && !seg.code) {
-                                <span>{{ seg.text }}</span>
-                              }
-                            }
-                          </span>
-                        </div>
-                      }
-                    </div>
-                  }
-                  @if (parsedOut.otherLines.length > 0) {
-                    <div class="parsed-other-lines">
-                      @for (line of parsedOut.otherLines; track trackMarkdownLine($index, line)) {
-                        <p class="other-line" [class]="'line-' + line.type">
-                          @if (line.type === 'verify') {
-                            <span class="verify-badge" [class.assert]="line.checkKind === 'assert'">{{ line.checkKind || 'verify' }}{{ line.atEnd ? '@end' : '' }}</span>
-                          }
-                          @if (line.type === 'finding') {
-                            <span class="finding-badge">finding</span>
-                          }
-                          @for (seg of line.segments; track seg) {
-                            @if (seg.bold) {
-                              <strong>{{ seg.text }}</strong>
-                            }
-                            @if (seg.code) {
-                              <code>{{ seg.text }}</code>
-                            }
-                            @if (!seg.bold && !seg.code) {
-                              <span>{{ seg.text }}</span>
-                            }
-                          }
-                        </p>
-                      }
-                    </div>
-                  }
-                </div>
-              }
             </div>
           </div>
         }
@@ -1369,6 +1393,7 @@ limitations under the License.
         }
       </div>
     }
+    <ng-content select="[task-result]"></ng-content>
   </main>
 
   <!-- Fullscreen Image Zoom Lightbox Modal -->

apps/showcase_ui/src/app/components/agent-stream/agent-stream.component.scss

@@ -1921,6 +1921,8 @@ $stream-row-height: 25px;
                 }
 
                &.task-report-card {
+                  width: min(680px, 100%);
+                  box-sizing: border-box;
                   background: #f0fdf4;
                   border: 1px solid #bbf7d0;
 
@@ -1967,12 +1969,13 @@ $stream-row-height: 25px;
                   }
 
                   .report-markdown-body {
-                    font-size: 12px;
+                    font-size: 14px;
-                    line-height: 1.5;
+                    line-height: 1.7;
                     color: #1f2937;
                     font-weight: 400;
                     user-select: text;
-                    word-break: break-word;
+                    word-break: normal;
+                    overflow-wrap: anywhere;
 
                     p {
                       margin: 3px 0;
@@ -2233,6 +2236,14 @@ $stream-row-height: 25px;
                         color: #1e293b;
                         word-break: break-all;
 
+                        &.report-markdown-body {
+                          font-size: 14px;
+                          font-weight: 400;
+                          line-height: 1.7;
+                          word-break: normal;
+                          overflow-wrap: anywhere;
+                        }
+
                        &.code-pill {
                          font-family: 'Roboto Mono', 'Fira Code', Consolas, monospace;
                          font-size: 11px;
@@ -3673,21 +3684,33 @@ $stream-row-height: 25px;
     &.running {
       border-left: 3.5px solid #2563eb;
     }
+    &.paused, &.task_paused { border-left: 3.5px solid #d97706; background: #fffbeb; }
+    &.resuming, &.queued { border-left: 3.5px solid #7c3aed; }
+    &.failed { border-left: 3.5px solid #dc2626; background: #fff7f7; }
+    &.completed { border-left: 3.5px solid #16a34a; }
+    &.cancelled, &.interrupted { border-left: 3.5px solid #ea580c; }
+    .item-error { color: #b91c1c; font-size: 11px; line-height: 1.4; overflow-wrap: anywhere; }
 
     .item-header {
       display: flex;
       align-items: center;
-      justify-content: space-between;
+      justify-content: flex-start;
+      flex-wrap: wrap;
       gap: 8px;
       min-width: 0;
 
+      .task-status-group { display: inline-flex; align-items: center; gap: 6px; flex-shrink: 0; }
+
       .task-badge {
         flex-shrink: 0;
         display: inline-flex;
         align-items: center;
-        font-size: 10.5px;
+        box-sizing: border-box;
+        height: 20px;
+        line-height: 14px;
+        font-size: 11px;
         font-weight: 700;
-        padding: 2px 5px;
+        padding: 0 5px;
         border-radius: 4px;
         letter-spacing: 0.05em;
         border: 1px solid #e2e8f0;
@@ -3718,12 +3741,24 @@ $stream-row-height: 25px;
           border-color: #fdba74;
           &::before { background-color: #f97316; }
         }
+        &.task_paused {
+          background: #ffedd5;
+          color: #9a3412;
+          border-color: #fb923c;
+          &::before { background-color: #ea580c; }
+        }
+        &.resuming, &.queued {
+          background: #f3e8ff;
+          color: #6b21a8;
+          border-color: #c4b5fd;
+          &::before { background-color: #7c3aed; }
+        }
         &.completed {
-          background: #fafafa;
+          background: #dcfce7;
-          color: #1f1f1f;
+          color: #166534;
-          border-color: #e2e8f0;
+          border-color: #86efac;
           &::before {
-            background-color: #1f1f1f;
+            background-color: #16a34a;
           }
         }
         &.pending {
@@ -3735,21 +3770,22 @@ $stream-row-height: 25px;
           }
         }
         &.failed {
-          background: #fafafa;
+          background: #fee2e2;
-          color: #ef4444;
+          color: #991b1b;
-          border-color: #fee2e2;
+          border-color: #fca5a5;
           &::before {
             background-color: #ef4444;
           }
         }
         &.cancelled {
-          background: #fafafa;
+          background: #ffedd5;
-          color: #d97706;
+          color: #9a3412;
-          border-color: #fef3c7;
+          border-color: #fdba74;
           &::before {
-            background-color: #f59e0b;
+            background-color: #ea580c;
           }
         }
+        &.interrupted { background: #fef3c7; color: #92400e; border-color: #fbbf24; &::before { background-color: #d97706; } }
 
         .status-dot {
           display: none;
@@ -3757,6 +3793,7 @@ $stream-row-height: 25px;
       }
 
       .item-header-actions {
+        margin-left: auto;
         display: flex;
         align-items: center;
         gap: 4px;
@@ -3971,3 +4008,84 @@ $stream-row-height: 25px;
     }
   }
 }
+.guidance-step-note, .guidance-queue-item {
+  padding: 10px 13px;
+  border: 1px solid #bfdbfe;
+  border-radius: 10px;
+  background: #eff6ff;
+  color: #1e3a8a;
+  line-height: 1.6;
+  overflow-wrap: anywhere;
+  .guidance-step-title { display: flex; align-items: center; gap: 6px; font-weight: 650; }
+  .material-symbols-outlined { font-size: 17px; }
+  p { margin: 5px 0; white-space: pre-wrap; color: #1e293b; }
+  small { color: #475569; }
+}
+.guidance-queue { display: grid; gap: 8px; padding: 12px 16px; }
+.guidance-queue-item strong { font-size: 12px; }
+.session-failure-summary {
+  margin: 12px 16px 28px;
+  padding: 12px 14px;
+  border: 1px solid #fecdd3;
+  border-radius: 10px;
+  background: #fff7f8;
+  color: #881337;
+  strong { font-size: 13px; }
+  p { margin: 6px 0 10px; line-height: 1.6; white-space: pre-wrap; overflow-wrap: anywhere; }
+  button { border: 0; background: transparent; padding: 0; color: #be123c; font: inherit; font-weight: 600; cursor: pointer; }
+}
+.report-expand-button {
+  align-self: flex-start;
+  border: 1px solid #cbd5e1;
+  border-radius: 6px;
+  padding: 4px 10px;
+  background: #fff;
+  color: #334155;
+  font: inherit;
+  font-size: 12px;
+  cursor: pointer;
+  &:hover { background: #f1f5f9; }
+  &:focus-visible { outline: 2px solid #3b82f6; outline-offset: 2px; }
+}
+:host ::ng-deep .report-markdown-body {
+  display: block;
+  width: 100%;
+  min-width: 0;
+  font-size: 13px;
+  font-weight: 400;
+  line-height: 1.7;
+  overflow-wrap: anywhere;
+  word-break: normal;
+  > div { margin: 3px 0; }
+  h1, h2, h3 { margin: 12px 0 6px; font-size: 14px; color: #0f172a; }
+  ul, ol { padding-left: 22px; margin: 8px 0; }
+  li { margin: 4px 0; }
+  pre { overflow: auto; white-space: pre-wrap; padding: 10px; border-radius: 7px; background: #e2e8f0; }
+  table { width: 100%; border-collapse: collapse; }
+  th, td { padding: 7px 9px; border: 1px solid #cbd5e1; text-align: left; vertical-align: top; }
+  th { background: #dbeafe; }
+  .md-para-gap { height: 8px; }
+}
+
+.evidence-source { max-width: 180px; overflow: hidden; text-overflow: ellipsis; white-space: nowrap; padding: 2px 7px; border: 1px solid #cbd5e1; border-radius: 6px; background: #f1f5f9; color: #475569; font-size: 11px; cursor: pointer; }
+
+.plugin-tool-card {
+  .action-card-header { flex-wrap: wrap; }
+  .plugin-call-kind { font-size: 10px; color: #64748b; }
+  .plugin-result { white-space: pre-wrap; overflow-wrap: anywhere; }
+  &.is-auxiliary:not(.is-expanded):not(.is-failed):not(.is-running) {
+    background: #f8fafc;
+    border-color: #e2e8f0;
+    box-shadow: none;
+  }
+}
+
+.stream-logs-content {padding-bottom:24px}
+.exploration-message {padding:12px 16px;border:1px solid #e2e8f0;border-radius:12px;margin:14px 0;background:#fff;color:#334155;font-size:13px;line-height:1.7}
+.exploration-message.user {background:#eff6ff;border-color:#dbeafe;margin-left:28px}.exploration-message.system {color:#64748b;background:#f8fafc;font-size:12px}
+.exploration-message small {color:#64748b;font-size:11px}.exploration-message p {white-space:pre-wrap;overflow-wrap:anywhere;margin:6px 0}.exploration-message button {font:inherit;padding:4px 10px;border:1px solid #cbd5e1;border-radius:6px;background:#fff;color:#334155;cursor:pointer}
+
+.exploration-message video {width:100%;max-height:300px;background:#0f172a}.exploration-message summary {cursor:pointer}.exploration-message .error {color:#b91c1c}
+.exploration-message.assistant {border-left:3px solid #818cf8;background:#fafaff}
+.exploration-message.question {border:1px solid #f4c97a;border-left:3px solid #f59e0b;background:#fffbeb}
+.exploration-message small {display:flex;align-items:center;gap:6px;font-size:11px;margin-bottom:8px}.exploration-message small .material-symbols-outlined {font-size:16px;color:#6366f1}.exploration-message small b {margin-left:auto;color:#92400e;background:#fef3c7;border-radius:12px;padding:2px 8px}

apps/showcase_ui/src/app/components/agent-stream/agent-stream.component.ts

@@ -1,4 +1,9 @@
+import { ExplorationMessage, Demonstration } from '../../services/exploration.service';
+import { mergeConversation } from '../../utils/exploration-timeline.util';
+import { TaskIdCopyComponent } from '../task-id-copy/task-id-copy.component';
 import { TaskHistoryComponent } from '../task-history/task-history.component';
+import { LlmRequestStatusComponent } from '../llm-request-status/llm-request-status.component';
+import { ReportMarkdownComponent } from '../report-markdown/report-markdown.component';
 import { AutomationPanelComponent } from '../automation-panel/automation-panel.component';
 import { UiLabelPipe } from '../../utils/ui-label.pipe';
 /**
@@ -17,18 +22,23 @@ import { UiLabelPipe } from '../../utils/ui-label.pipe';
  * limitations under the License.
  */
 
-import { Component, ChangeDetectionStrategy, NgZone, signal, computed, effect, inject, untracked, DestroyRef, AfterViewInit } from '@angular/core';
+import { Component, ChangeDetectionStrategy, NgZone, signal, computed, effect, inject, untracked, DestroyRef, AfterViewInit, input, output } from '@angular/core';
 import { CommonModule } from '@angular/common';
 import { FormsModule } from '@angular/forms';
 import { RouterLink } from '@angular/router';
 import { HttpClient } from '@angular/common/http';
-import { AgentService, StartupProgressEvent } from '../../services/agent.service';
+import { AgentService, GuidanceEvent, StartupProgressEvent } from '../../services/agent.service';
 import { Session, ModelInfo, SessionUsage } from '../../core/models/session.model';
+import { TestResultBadgeComponent, TEST_RESULT_LABELS } from '../test-result-badge/test-result-badge.component';
+import { TestResultCardComponent } from '../test-result-badge/test-result-card.component';
+import { TaskBatchFilterComponent } from '../task-batch-filter/task-batch-filter.component';
 import { MarkdownSegment, MarkdownLine, NoteMilestone, ParsedNote } from '../../core/models/markdown.model';
 import { OverlayModule, ConnectedPosition } from '@angular/cdk/overlay';
 import { StepBlock, PhaseBlock, StepEvent, ActionParam, CheckerResult, StreamResetNotice, DEFAULT_STREAM_RESET_MESSAGE } from '../../core/models/stream.model';
 import {
   contextPercent,
+  contextWindow,
+  runMode,
   formatCompactTokens,
   formatElapsed,
   parseRunTuning,
@@ -158,6 +168,7 @@ import {
   extractCheckerResult,
   renderMarkdownToHtml
 } from '../../utils/markdown-parser.util';
+import { getPluginToolView } from '../../utils/plugin-tool-view.util';
 
 import {
   getActionObject,
@@ -178,7 +189,8 @@ import {
   getStepPostImageUrl,
   isReportStatusAction,
   getReportStatusValue,
-  getReportStatusExplanation
+  getReportStatusExplanation,
+  actionEvidenceSource
 } from '../../utils/action-formatter.util';
 
 import {
@@ -236,16 +248,28 @@ export type { MarkdownSegment, MarkdownLine, NoteMilestone, ParsedNote, StreamRe
 @Component({
   selector: 'app-agent-stream',
   standalone: true,
-  imports: [UiLabelPipe, CommonModule, FormsModule, OverlayModule, TaskHistoryComponent, AutomationPanelComponent],
+  imports: [ReportMarkdownComponent, LlmRequestStatusComponent, TaskIdCopyComponent, UiLabelPipe, CommonModule, FormsModule, OverlayModule, TaskHistoryComponent, AutomationPanelComponent, TestResultBadgeComponent, TestResultCardComponent, TaskBatchFilterComponent],
   templateUrl: './agent-stream.component.html',
   styleUrl: './agent-stream.component.scss',
   changeDetection: ChangeDetectionStrategy.OnPush
 })
 export class AgentStreamComponent implements AfterViewInit {
+  conversationMessages = input<ExplorationMessage[]>([]);
+  demonstrations = input<Demonstration[]>([]);
+  pendingQuestion = input('');
+  demoMediaUrl(runId: string, path: string) { return '/videos/' + [runId, ...path.split('/')].map(encodeURIComponent).join('/'); }
+  demoLabel(status: string) { return ({requested:'准备录制',recording:'录制中',analyzing:'分析中',completed:'已分析',failed:'分析失败',cancelled:'已取消',timed_out:'到时已保存'} as Record<string,string>)[status] || status; }
+  useRecommendation = output<string>();
+
   public agentService = inject(AgentService);
   private http = inject(HttpClient);
   private destroyRef = inject(DestroyRef);
   private zone = inject(NgZone);
+  public terminalFailureReason = signal('');
+  private selectedTerminalStatus = computed(() => {
+    const id = this.agentService.currentSessionId();
+    return this.agentService.sessions().find(session => session.session_id === id)?.status || '';
+  });
 
   // Auto-scroll state tracking
   public isUserAtBottom = true;
@@ -270,6 +294,7 @@ export class AgentStreamComponent implements AfterViewInit {
   public isTaskDropdownOpen = signal<boolean>(false);
   public isNotesDropdownOpen = signal<boolean>(false);
   public isAutomationOpen = signal(false);
+  public deleteError = signal('');
 
   // Run info popover (opened from the Flash / Pro chip): elapsed time, token
   // usage, executor context share and the Pro tuning the run started with.
@@ -277,7 +302,9 @@ export class AgentStreamComponent implements AfterViewInit {
   public runUsage = signal<SessionUsage | null>(null);
   private runInfoClock = signal<number>(Date.now());
 
-  public runInfoIsPro = computed(() => this.getModelClass(this.agentService.viewedModel()?.name) === 'is-pro');
+  public runInfoMode = computed(() => runMode(this.agentService.currentSession(), this.agentService.viewedModel()));
+  public runInfoIsPro = computed(() => this.runInfoMode() === 'pro');
+  public runInfoContextWindow = computed(() => contextWindow(this.runUsage()?.operator_context_window_tokens));
 
   public runInfoElapsed = computed(() => {
     const session = this.agentService.currentSession();
@@ -372,6 +399,7 @@ export class AgentStreamComponent implements AfterViewInit {
   public expandedActionCards = signal<Set<string>>(new Set<string>());
   public collapsedActionCards = signal<Set<string>>(new Set<string>());
   public retryClock = signal<number>(Date.now());
+  public hasModelRequest = signal(false);
 
   // Performance caches for parameter and event extractions
   private actionParamsCache = new WeakMap<any, ActionParam[]>();
@@ -382,13 +410,13 @@ export class AgentStreamComponent implements AfterViewInit {
   public activeQueue = computed(() => {
     const list = this.agentService.sessions().filter((s) => {
       const status = this.getTaskStatus(s);
-      return status === 'running' || status === 'paused' || status === 'pending';
+      return ['running', 'paused', 'task_paused', 'resuming', 'pending', 'queued'].includes(status) && this.agentService.matchesBatchFilter(s) && this.agentService.isQueueRepresentative(s);
     });
     return list.sort((a, b) => {
       const statusA = this.getTaskStatus(a);
       const statusB = this.getTaskStatus(b);
-      const isRunA = statusA === 'running' || statusA === 'paused';
+      const isRunA = ['running', 'paused', 'task_paused', 'resuming'].includes(statusA);
-      const isRunB = statusB === 'running' || statusB === 'paused';
+      const isRunB = ['running', 'paused', 'task_paused', 'resuming'].includes(statusB);
       if (isRunA && !isRunB) return -1;
       if (!isRunA && isRunB) return 1;
       return (a.start_time || 0) - (b.start_time || 0);
@@ -398,7 +426,7 @@ export class AgentStreamComponent implements AfterViewInit {
   public historyTasks = computed(() => {
     return this.agentService.sessions().filter((s) => {
       const status = this.getTaskStatus(s);
-      return status === 'completed' || status === 'failed' || status === 'cancelled';
+      return ['completed', 'failed', 'cancelled', 'interrupted', 'skipped'].includes(status);
     });
   });
 
@@ -419,13 +447,6 @@ export class AgentStreamComponent implements AfterViewInit {
     return notes['output.md'] || null;
   });
 
-  // Memoized so the Task Report card is not re-parsed (and its DOM rebuilt)
-  // on every change-detection pass while streams are typing.
-  public parsedOutputReport = computed<ParsedNote | null>(() => {
-    const report = this.outputterReport();
-    return report ? parseNote(report) : null;
-  });
-
   public parsedNote = computed<ParsedNote>(() => {
     return this.getParsedNote(this.currentNoteContent());
   });
@@ -446,6 +467,7 @@ export class AgentStreamComponent implements AfterViewInit {
   });
 
   public showPlanningLoader = computed(() => {
+    if (this.hasModelRequest()) return false;
     const isRunning = this.agentService.agentStatus() === 'running';
     const currentSessionId = this.agentService.currentSessionId();
     const runningSessionId = this.agentService.runningSessionId();
@@ -473,9 +495,26 @@ export class AgentStreamComponent implements AfterViewInit {
     return consolidateLogsToBlocks(this.filteredLogs());
   });
 
-  public phases = computed<PhaseBlock[]>(() => {
+  public awaitingGuidance = computed<GuidanceEvent[]>(() => {
+    const steps = this.consolidatedBlocks().filter(block => block.type === 'step');
+    return (this.agentService.guidanceEvents?.() || []).filter(event => !steps.some(block =>
+      new Date(block.timestamp).getTime() >= event.timestamp * 1000
+      && String(block.data?.extra_metadata?.injected_instruction || '').includes(event.instruction)
+    ));
+  });
+
+  public guidanceStatusLabel(event: GuidanceEvent): string {
+    const terminal = ['completed', 'failed', 'cancelled', 'interrupted']
+      .includes(this.agentService.currentSession()?.status || '');
+    if (event.status === 'pending') {
+      return terminal ? '任务结束前未进入模型' : '已接收，正在打断当前模型步骤';
+    }
+    return terminal ? '模型已读取，未留下后续动作记录' : '模型已读取，等待动作记录';
+  }
+
+  public phases = computed(() => {
     const currentSession = this.agentService.sessions().find(s => s.session_id === this.agentService.currentSessionId());
-    return groupBlocksToPhases(this.consolidatedBlocks(), currentSession?.start_time || 0);
+    return mergeConversation(groupBlocksToPhases(this.consolidatedBlocks(), currentSession?.start_time || 0), this.conversationMessages(), this.demonstrations());
   });
 
   public startupWorkItems = computed(() => {
@@ -540,6 +579,23 @@ export class AgentStreamComponent implements AfterViewInit {
   }
 
   constructor() {
+    const conversationRevision = computed(() => JSON.stringify([
+      this.conversationMessages().map(message => message.id),
+      this.demonstrations().map(demo => [demo.id,demo.status])
+    ]));
+    effect(() => { conversationRevision(); untracked(() => this.scheduleAutoScroll(false)); });
+
+    effect((onCleanup) => {
+      const id = this.agentService.currentSessionId();
+      const status = this.selectedTerminalStatus();
+      this.terminalFailureReason.set('');
+      if (!id || !['failed', 'interrupted'].includes(status || '')) return;
+      const request = this.http.get<{error?: string}>(`/api/automation/reports/${encodeURIComponent(id)}`).subscribe({
+        next: report => this.terminalFailureReason.set(report.error || '任务未完成；请查看已记录的执行步骤。'),
+        error: () => this.terminalFailureReason.set('任务未完成，暂时无法读取异常详情。')
+      });
+      onCleanup(() => request.unsubscribe());
+    });
     // Run info popover: fetch usage when opened or when the viewed session
     // changes, and keep refreshing while that session is still running.
     effect((onCleanup) => {
@@ -818,6 +874,12 @@ export class AgentStreamComponent implements AfterViewInit {
   }
 
   // Top Nav Action Methods
+  public openBatchQueue(): void {
+    this.isAutomationOpen.set(false);
+    this.isNotesDropdownOpen.set(false);
+    this.isTaskDropdownOpen.set(true);
+  }
+
   public toggleTaskDropdown(event: Event): void {
     event.stopPropagation();
     this.isTaskDropdownOpen.update(v => !v);
@@ -878,14 +940,17 @@ export class AgentStreamComponent implements AfterViewInit {
     return tuningLabel(kind, id);
   }
 
-  public getTaskStatus(session: Session): 'running' | 'paused' | 'completed' | 'pending' | 'failed' | 'cancelled' {
+  public getTaskStatus(session: Session): string {
+    const exploration = this.agentService.explorationQueueStatus(session);
+    if (exploration) return exploration as 'running' | 'paused' | 'completed' | 'pending' | 'failed' | 'cancelled';
+
     if (session.status) {
       const s = session.status.toLowerCase();
-      if (s === 'completed' || s === 'success' || s === 'failed' || s === 'cancelled') {
+      if (['completed', 'success', 'failed', 'cancelled', 'interrupted', 'skipped'].includes(s)) {
-        return (s === 'success' ? 'completed' : s) as any;
+        return s === 'success' ? 'completed' : s;
       }
-      if (s === 'running' || s === 'paused' || s === 'pending') {
+      if (['running', 'paused', 'task_paused', 'resuming', 'pending', 'queued'].includes(s)) {
-        return s as any;
+        return s;
       }
     }
     if (session.session_id === this.agentService.runningSessionId() && (this.agentService.agentStatus() === 'running' || this.agentService.agentStatus() === 'paused')) {
@@ -933,12 +998,13 @@ export class AgentStreamComponent implements AfterViewInit {
 
   public deleteTask(sessionId: string, event?: Event): void {
     if (event) event.stopPropagation();
-    if (!confirm("确定删除此任务？此操作无法撤销。")) {
+    if (!confirm(this.agentService.deletionPrompt(sessionId))) {
       return;
     }
+    this.deleteError.set('');
     this.agentService.deleteSession(sessionId).subscribe({
       error: (err) => {
-        console.error(`删除会话失败：${sessionId}:`, err);
+        this.deleteError.set(err.error?.detail || "删除失败，请稍后重试");
       }
     });
   }
@@ -1326,6 +1392,19 @@ export class AgentStreamComponent implements AfterViewInit {
     return isAndroidAction(action);
   }
 
+  public formatTaskDuration(start?: number, end?: number): string {
+    if (!start || !Number.isFinite(start)) return '—';
+    const seconds = Math.max(0, Math.floor((end || Date.now() / 1000) - start));
+    const hours = Math.floor(seconds / 3600);
+    const minutes = Math.floor((seconds % 3600) / 60);
+    return hours ? `${hours}时${minutes}分` : minutes ? `${minutes}分${seconds % 60}秒` : `${seconds}秒`;
+  }
+
+  public hasRecordedDeviceAction(action: any): boolean {
+    const actions = Array.isArray(action) ? action : action ? [action] : [];
+    return actions.some(item => this.isAndroidAction(item) && !this.isReportStatusAction(item));
+  }
+
   public getActionIcon(action: any): string {
     return getActionIcon(action);
   }
@@ -1556,6 +1635,17 @@ export class AgentStreamComponent implements AfterViewInit {
     }
   }
 
+  public evidenceSourceLabel(item: any): string { return actionEvidenceSource(item).label; }
+
+  public openEvidenceVideo(item: any, step: any, event: Event): void {
+    event.stopPropagation();
+    const started = this.agentService.currentSession()?.start_time;
+    const value = item?.timestamp || item?.start_time || step?.timestamp;
+    const timestamp = typeof value === 'number' ? value : Date.parse(value) / 1000;
+    if (!started || !Number.isFinite(timestamp)) return;
+    this.agentService.openVideoPlayer(undefined, undefined, undefined, Math.max(0, timestamp - started), undefined, actionEvidenceSource(item).id);
+  }
+
   public getScreenRecordingButtonTitle(): string {
     if (this.agentService.isCurrentSessionRunning()) {
       return this.agentService.currentSession()?.recording_status === 'recording' ? "任务正在运行（录屏中）" : "任务正在运行";
@@ -1581,6 +1671,12 @@ export class AgentStreamComponent implements AfterViewInit {
     return isDeviceActionTool(tool);
   }
 
+  public getPluginToolView = getPluginToolView;
+
+  public isEvidenceTool(tool: any): boolean {
+    return isDeviceActionTool(tool) || getPluginToolView(tool) !== null;
+  }
+
   public isFailureAnalyzerActionTool(tool: any): boolean {
     return isDeviceActionTool(tool);
   }
@@ -1707,7 +1803,9 @@ export class AgentStreamComponent implements AfterViewInit {
   /** Follow-ups not already visible in the verdict rows (the findings text repeats them). */
   public getCheckerNotes(block: any): string[] {
     const d = block?.data || {};
-    if (d.phase === 'outcome') return Array.isArray(d.last_findings) ? d.last_findings : [];
+    if (d.phase === 'outcome') {
+      return [d.test_result?.reason, ...(d.test_result?.evidence || []), ...(Array.isArray(d.last_findings) ? d.last_findings : [])].filter(Boolean);
+    }
     const notes: string[] = [];
     if (Array.isArray(d.unmet_subgoals)) {
       for (const text of d.unmet_subgoals) notes.push(`尚未完成：${text}`);
@@ -1749,7 +1847,9 @@ export class AgentStreamComponent implements AfterViewInit {
   public getCheckerTone(block: any): string {
     const d = block?.data || {};
     if (d.phase === 'outcome') {
-      return d.task_status === 'completed' ? 'passed' : (d.task_status === 'blocked' ? 'failed' : 'inconclusive');
+      if (d.test_result) return d.test_result.status === 'blocked' ? 'inconclusive' : d.test_result.status;
+      if (d.tests?.failed) return 'failed';
+      return d.tests?.passed && !d.tests?.unchecked && !d.tests?.inconclusive ? 'passed' : 'inconclusive';
     }
     if (d.isCompleted === false) return 'running';
     if (d.status === 'superseded' || d.status === 'unchecked') return 'muted';
@@ -1764,12 +1864,12 @@ export class AgentStreamComponent implements AfterViewInit {
     const d = block?.data || {};
     if (d.phase === 'outcome') {
       const t = d.tests || {};
-      const label = d.task_status === 'completed' ? "目标已完成"
+      const label = d.test_result ? TEST_RESULT_LABELS[d.test_result.status as keyof typeof TEST_RESULT_LABELS]
-        : (d.task_status === 'blocked' ? "受阻" : "部分完成");
+        : (d.task_status === 'completed' ? "执行完成" : (d.task_status === 'blocked' ? "执行受阻" : "部分完成"));
       const counts = [
-        [t.passed, 'passed'],
+        [t.passed, '项通过'],
-        [t.failed, 'failed'],
+        [t.failed, '项不通过'],
-        [t.inconclusive, 'inconclusive'],
+        [t.inconclusive, '项无法确认'],
         [t.unchecked, "未检查"]
       ]
         .filter(([n]) => Number(n) > 0)
@@ -1812,6 +1912,7 @@ export class AgentStreamComponent implements AfterViewInit {
   public hasVisibleContent(block: any): boolean {
     if (!block) return false;
     if (block.type === 'checker') return true;
+    if (block.data?.extra_metadata?.injected_instruction) return true;
     const hasNative = Boolean(this.getNativeThinking(block));
     const hasRaw = Boolean(this.getRawThinking(block));
     const hasReset = this.hasBlockResets(block, false) || this.hasBlockResets(block, true);
@@ -1972,11 +2073,12 @@ export class AgentStreamComponent implements AfterViewInit {
   }
 
   public getToolIcon(tool: any): string {
-    return getToolIcon(tool);
+    const plugin = getPluginToolView(tool);
+    return plugin ? (plugin.auxiliary ? 'visibility' : 'extension') : getToolIcon(tool);
   }
 
   public getToolTitle(tool: any): string {
-    return getToolTitle(tool);
+    return getPluginToolView(tool)?.title || getToolTitle(tool);
   }
 
   public getToolTargetText(tool: any): string {
@@ -2152,6 +2254,7 @@ export class AgentStreamComponent implements AfterViewInit {
   public getModelIcon(name: string | undefined | null): string {
     if (!name) return 'bolt';
     const lower = name.toLowerCase();
+    if (lower === '探索') return 'explore';
     if (lower.includes('pro')) return 'diamond';
     if (lower.includes('flash')) return 'bolt';
     return 'smart_toy';
@@ -2167,6 +2270,7 @@ export class AgentStreamComponent implements AfterViewInit {
 
   public getArchitectureTooltip(model?: ModelInfo | null): string {
     if (!model) return "执行模式：ARTEMIS Flash（快速响应）";
+    if (model.architecture === 'explore') return `探索模式（可交流与人工示范） · LLM: ${model.id}`;
     const name = this.getModelDisplayName(model.name);
     const isPro = name.toLowerCase().includes('pro');
     const archDesc = isPro

apps/showcase_ui/src/app/components/automation-panel/automation-panel.component.html

@@ -5,6 +5,7 @@
 @if (error()) { <p class="error" role="alert">{{error()}}</p> }
 @if (notice()) { <p class="notice" role="status">{{notice()}}</p> }
 @if (tab() === 'reports') {
+  @if(batch(); as b){<button type="button" class="back-to-batch" (click)="tab.set('batches')">返回批次：{{b.name}}</button>}
   <section><h3>默认报告格式</h3><p>任务结束后自动生成。批次报告包同时包含汇总和每个任务的报告。</p>
     <div class="format-options"><label><input type="checkbox" [(ngModel)]="word"> Word 文档</label>
     <label><input type="checkbox" [(ngModel)]="excel"> Excel 表格</label></div>
@@ -12,42 +13,100 @@
   </section>
   <section><h3>当前任务报告</h3>
     @if (report(); as r) {
-      <strong>{{r.name}}</strong><p>{{label(r.status)}} · {{r.verdict}}</p>
+      <div class="report-overview">
-      <div class="actions"><a [href]="'/api/automation/reports/'+r.id+'/docx'">下载 Word</a><a [href]="'/api/automation/reports/'+r.id+'/xlsx'">下载 Excel</a></div>
+        <div class="report-heading"><app-favorite-button [sourceId]="r.id" /><strong class="report-title">{{r.name}}</strong></div>
+        <div class="report-badges"><span [class.failed]="r.status === 'failed'">{{label(r.status)}}</span><app-test-result-badge [result]="r.test_result" [executionStatus]="r.status" /></div>
+        @if(r.test_result; as result){<app-test-result-card [result]="result" [sessionId]="r.id" [executionStatus]="r.status" (reviewSaved)="reviewSaved($event)" />}
+        @if (r.error) { <div class="report-error" role="status"><strong>异常原因</strong><p>{{r.error}}</p></div> }
+      </div>
+      <div class="report-meta"><div><strong>任务目标与预期</strong><p>{{r.goal || '未记录'}}</p></div>@if(r.expected){<div><strong>预期结果</strong><p>{{r.expected}}</p></div>}</div>
+      <div class="actions"><a [href]="'/api/automation/reports/'+r.id+'/docx'">下载 Word</a><a [href]="'/api/automation/reports/'+r.id+'/xlsx'">下载 Excel</a>
+        <a [href]="'/api/automation/reports/'+r.id+'/screenshots'">步骤截图 ZIP</a>
+        <a [href]="'/api/automation/reports/'+r.id+'/video'">分段录屏 ZIP</a>
+        <a [href]="'/api/automation/reports/'+r.id+'/all'">下载全部资料</a></div>
       <p class="muted">报告包含目标、预期、实际结论、步骤与截图证据。执行中的下载为当前快照。</p>
-      <pre class="preview">{{r.content}}</pre>
+      <div class="report-content"><h3>结论与发现</h3><app-report-markdown class="report-body" [content]="r.content" /></div>
     } @else { <p>从任务队列或历史记录中选择任务，即可下载报告。</p> }
+    @if (liveTask() && agent.currentSessionId()) {
+      <div class="task-guidance">
+        <h3>任务指导 · {{label(taskStatus())}}</h3>
+          <p>{{ taskStatus() === 'paused' ? '模型请求已暂停；提交指导后请在执行日志中恢复模型调用。' : '运行中提交会立即打断模型思考并重新观察设备；正在执行的设备动作会先安全结束。' }}</p>
+        <label for="task-guidance-text">补充信息或调整方向</label>
+        <textarea id="task-guidance-text" [(ngModel)]="guidance" maxlength="10000" rows="3" placeholder="描述已观察到的情况和下一步方向"></textarea>
+        <div class="actions">
+          @if(taskStatus() === 'running'){<button [disabled]="busy()" (click)="pauseTask(agent.currentSessionId()!)">暂停当前任务</button>}
+           <button [disabled]="busy() || !guidance.trim()" (click)="guideTask(agent.currentSessionId()!)">提交指导{{taskStatus() === 'task_paused' ? '并恢复' : ''}}</button>
+           @if(taskStatus() === 'task_paused'){<button [disabled]="busy()" (click)="continueTask(agent.currentSessionId()!)">直接继续</button>}
+        </div>
+      </div>
+    }
   </section>
 }
 @if (tab() !== 'reports') {
-  <section class="device"><label>执行设备<select [(ngModel)]="device"><option value="">请选择设备</option>@for(d of system.connectedDevices(); track d.serial){<option [value]="d.serial">{{d.serial}}</option>}</select></label></section>
+  <app-task-plugins (changed)="taskPlugins = $event" /><section class="device"><label>执行设备<select [(ngModel)]="device"><option value="">请选择设备</option>@for(d of system.connectedDevices(); track d.serial){<option [value]="d.serial">{{d.serial}}</option>}</select></label></section>
 }
 @if (tab() === 'batches') {
-  <section><h3>导入任务</h3><p>每行一个用例，填写名称与目标。预期结果、应用、前置条件、参数和标签可选。</p>
+  <section><h3>导入任务</h3><p>每行一个用例，仅「Artemis提示词」作为任务正文；「任务配置(JSON)」可设置每项模型、模式、超时和录屏。「前置约束提取」仅供查看与搜索。示例在模板的「填写说明」页。</p>
     <div class="actions"><a href="/api/automation/template">下载 Excel 模板</a><label class="upload">导入 Excel<input type="file" accept=".xlsx" [disabled]="busy()" (change)="upload($event)"></label></div>
     @if(preview(); as p) {
-      <p>有效任务 {{p.tasks.length}} 条 · 错误 {{p.errors.length}} 条</p>
+      <div class="selection-toolbar"><span>已选 <strong>{{selectedCount()}}</strong> / {{p.tasks.length}} 个用例 · 错误 {{p.errors.length}} 条</span><button type="button" [disabled]="busy() || !p.tasks.length" (click)="toggleAllRows()">{{selectedCount() === p.tasks.length ? '全不选' : '全选'}}</button></div>
       @for(e of p.errors; track e.row){<p class="error">第 {{e.row}} 行：{{e.message}}</p>}
-      <div class="table-wrap"><table><thead><tr><th>行</th><th>名称</th><th>目标</th></tr></thead><tbody>@for(t of p.tasks; track t.row){<tr><td>{{t.row}}</td><td>{{t.name}}</td><td>{{t.goal}}</td></tr>}</tbody></table></div>
+      <div class="case-preview-list" aria-label="选择导入用例">
+        @for(t of p.tasks; track t.row){
+          <div class="case-preview" [class.unselected]="!selectedRows().has(t.row)">
+            <span class="selection-column"><input type="checkbox" [checked]="selectedRows().has(t.row)" [disabled]="busy()" (change)="toggleRow(t.row)" [attr.aria-label]="'导入用例：' + (t.case_id ? t.case_id + ' ' : '') + t.name"></span>
+            <div class="case-preview-content"><strong>{{t.case_id ? t.case_id + ' ' : ''}}{{t.name}}</strong><small>第 {{t.row}} 行</small>
+              <p class="case-goal" [title]="t.goal">{{t.goal}}</p>
+              @if(caseConfig(t.task_config); as config){<details><summary>自定义参数</summary><pre>{{config}}</pre></details>}@else{<small>使用批次 / 全局参数</small>}
+            </div>
+          </div>
+        }
+      </div>
       <label>批次名称<input [(ngModel)]="name" maxlength="200"></label>
       <div class="fields"><label>执行模式<select [(ngModel)]="profile"><option value="flash">Flash</option><option value="pro">Pro</option></select></label><label>单任务时限（秒）<input type="number" min="30" max="86400" [(ngModel)]="timeout"></label></div>
+      <label for="batch-repeat-count">整批循环次数<input id="batch-repeat-count" type="number" min="1" [max]="maxExecutions" step="1" [(ngModel)]="repeatCount"></label>
+      <p class="execution-summary">{{selectedCount()}} 个用例 × {{repeatCount || 0}} 轮，共 {{selectedCount() * (repeatCount || 0)}} 次执行</p>
+      <p class="muted">每轮按表格顺序执行所选用例，分别保存结果。未勾选用例不会导入。</p>
+      @if(!validRepetitions()){<p class="error">循环次数须为正整数，执行总数不超过 {{maxExecutions}}。</p>}
       <label><input type="checkbox" [(ngModel)]="stopOnFailure"> 失败即停止后续任务（默认继续）</label>
       <label><input type="checkbox" [(ngModel)]="word"> Word</label><label><input type="checkbox" [(ngModel)]="excel"> Excel</label>
-      <button class="primary" [disabled]="busy() || !p.valid || !device || !name.trim() || !formats().length" (click)="startBatch()">确认导入并开始执行</button>
+      <button class="primary" [disabled]="busy() || !p.valid || !selectedCount() || !validRepetitions() || !name.trim() || !formats().length" (click)="startBatch()">确认导入并开始执行</button>
     }
   </section>
   <section><h3>批次记录</h3>
-    @for(b of batches(); track b.id){<button class="record" [class.selected]="batch()?.id === b.id" (click)="selectBatch(b.id)"><strong>{{b.name}}</strong><span>{{label(b.status)}} · {{b.created_at*1000 | date:'MM-dd HH:mm'}}</span></button>}
+    @for(b of batches(); track b.id){
+      <div class="batch-record">
+        <app-favorite-button sourceKind="batch" [sourceId]="b.id" />
+        <button class="record" [class.selected]="batch()?.id === b.id" (click)="selectBatch(b.id)"><strong>{{b.name}}</strong><span>{{label(b.status)}} · {{b.created_at*1000 | date:'MM-dd HH:mm'}} · {{b.repeat_count || 1}} 轮@if(b.total_executions){ · {{b.total_executions}} 次执行}</span></button>
+        <button class="delete-batch" type="button" [disabled]="busy() || b.status === 'running'" (click)="deleteBatch(b)" [attr.aria-label]="'删除批次：' + b.name" [title]="b.status === 'running' ? '执行中的批次不能删除' : '删除批次及其任务、报告和文件'"><span class="material-symbols-outlined">delete_outline</span></button>
+      </div>
+    }
     @if(!batches().length){<p>暂无批次。</p>}
   </section>
   @if(batch(); as b){<section><h3>{{b.name}}</h3><p>{{label(b.status)}} · {{b.reason}}</p>
+     <p class="execution-summary">{{b.case_count || b.jobs?.length || 0}} 个用例 · {{b.repeat_count || 1}} 轮 · 共 {{b.total_executions || b.jobs?.length || 0}} 次执行</p>
      <div class="actions">
+      <button type="button" (click)="showBatchQueue(b.id)">在任务队列查看此批次</button>
+      @if(agent.currentSessionId()){<button type="button" (click)="tab.set('reports')">查看当前任务报告</button>}
       @if(b.status === 'running'){<button [disabled]="busy()" (click)="batchAction('pause')">暂停后续任务</button>}
       @if(b.status === 'draft' || b.status === 'paused'){<button [disabled]="busy()" (click)="batchAction('resume')">{{b.status === 'draft' ? '开始执行' : '继续未开始任务'}}</button>}
-      @if(b.status !== 'running'){<button [disabled]="busy()" (click)="batchAction('retry_failed')">复制失败项为新批次</button>}
+      @if(b.status !== 'running' && b.status !== 'deleting'){<button [disabled]="busy()" (click)="batchAction('retry_failed')">复制失败项为新批次</button>}
-      <a [href]="'/api/automation/batches/'+b.id+'/archive'">下载全部报告 ZIP</a>
+       <button type="button" [disabled]="archiveBusy()" (click)="downloadBatchArchive(b.id)">{{archiveBusy() && archiveBatchId() === b.id ? '正在准备报告包…' : '下载批次报告 ZIP'}}</button>
      </div><p class="muted">服务重启后等待手动继续。失败项不会自动重试；暂停不打断当前任务。</p>
+     @if(archiveBatchId() === b.id && archiveProgress(); as progress){
+       <div class="archive-progress" role="status" aria-live="polite" [class.archive-failed]="progress.status === 'failed'">
+         <div class="archive-progress-heading">
+           @if(progress.status === 'preparing'){<span class="archive-spinner" aria-hidden="true"></span>}
+           <strong>{{progress.stage}}</strong>
+         </div>
+         @if(progress.status === 'preparing'){
+           <div class="archive-track" role="progressbar" [attr.aria-valuenow]="progress.total ? progress.completed : null" [attr.aria-valuemin]="progress.total ? 0 : null" [attr.aria-valuemax]="progress.total || null" [attr.aria-label]="'批次报告打包进度：' + progress.stage"><span [style.width.%]="progress.total ? progress.completed / progress.total * 100 : 8"></span></div>
+           <p>已处理 {{progress.completed}} / {{progress.total}} 个任务 · 已用 {{duration(progress.elapsed_seconds)}}@if(progress.eta_seconds !== null){ · 预计还需约 {{duration(progress.eta_seconds)}}}</p>
+         } @else if(progress.status === 'failed'){<p>{{progress.error}}。可点击下载按钮重试。</p>}
+           @else if(progress.status === 'ready'){<p>打包完成，下载已开始。</p>}
+       </div>
+     }
     @if(b.report_error){<p class="error">报告生成失败：{{b.report_error}}</p>}
-    @for(j of b.jobs; track j.session_id){<div class="job"><button (click)="agent.selectSession(j.session_id,true)">{{j.position}}. {{j.case.name}}</button><span>{{label(j.status)}}</span>@if(j.error){<small>{{j.error}}</small>}</div>}
+      @for(j of b.jobs; track j.session_id){<div class="job" [class.selected]="agent.currentSessionId() === j.session_id"><app-favorite-button [sourceId]="j.session_id" /><button (click)="openBatchTask(j.session_id)" [title]="j.case.goal" [attr.aria-pressed]="agent.currentSessionId() === j.session_id">{{j.case.case_id ? j.case.case_id + ' ' : ''}}{{j.case.name}}@if(j.repeat_count > 1){ · 第 {{j.iteration}} / {{j.repeat_count}} 轮}</button><span>{{label(j.status)}}</span><app-test-result-badge [result]="j.test_result" [executionStatus]="j.status" [compact]="true" />@if(j.error){<small>{{j.error}}</small>}</div>}
   </section>}
 }

apps/showcase_ui/src/app/components/automation-panel/automation-panel.component.scss

@@ -1,4 +1,6 @@
 :host { display:block;min-width:0;color:#1e293b;font-size:13px;line-height:1.65; }
+.back-to-batch{margin-bottom:12px;max-width:100%;text-align:left;overflow-wrap:anywhere}
+.job.selected{background:#eff6ff;border-radius:8px;padding-left:8px;padding-right:8px}.job.selected button{color:#1d4ed8;border-color:#93c5fd;background:#fff}
 .tabs {display:flex;gap:4px;padding:4px;margin-bottom:18px;background:#f1f5f9;border:1px solid #e2e8f0;border-radius:20px;
  button {flex:1;padding:7px 8px;border:0;border-radius:16px;color:#475569;background:transparent;white-space:nowrap;
  &.selected {color:#2563eb;background:#fff;box-shadow:0 1px 4px #0f172a12;} }
@@ -16,7 +18,43 @@ input:not([type=checkbox]):not([type=file]),select,textarea{box-sizing:border-bo
 input[type=checkbox]{width:15px;height:15px;accent-color:#2563eb;vertical-align:-3px;margin-right:6px;}
 .upload{position:relative;margin:0;color:#fff;background:#2563eb;border-color:#2563eb;}.upload:hover{background:#1d4ed8;border-color:#1d4ed8;}.upload input{position:absolute;inset:0;width:100%;height:100%;opacity:0;cursor:pointer;}
 .device{background:#f8fafc;padding:12px 16px;}.device label{margin:0;}
+.selection-toolbar{display:flex;align-items:center;justify-content:space-between;gap:8px;margin:12px 0;color:#475569;}
+.selection-column{flex:0 0 18px;text-align:center;}.selection-column input{margin:0;}
+.case-preview-list{max-height:320px;overflow:auto;border:1px solid #e2e8f0;border-radius:8px;margin-bottom:14px;}
+.case-preview{display:flex;align-items:flex-start;gap:10px;padding:12px;border-bottom:1px solid #e2e8f0;&:last-child{border-bottom:0;}&.unselected{background:#f8fafc;color:#64748b;}}
+.case-preview-content{min-width:0;flex:1;strong{display:block;overflow-wrap:anywhere;}small{display:block;}details{margin:5px 0 0;}summary{padding:0;font-size:12px;color:#2563eb;}pre{max-height:160px;overflow:auto;white-space:pre-wrap;overflow-wrap:anywhere;background:#f8fafc;border-radius:6px;padding:8px;font-size:11px;}}
+.case-goal{display:-webkit-box;-webkit-box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;white-space:pre-wrap;margin:5px 0;font-size:12px;}
+.execution-summary{padding:9px 12px;border-radius:8px;background:#eff6ff;color:#1d4ed8;font-weight:500;}
 .record{display:flex;width:100%;text-align:left;align-items:flex-start;flex-direction:column;gap:4px;margin:8px 0 0;padding:12px;overflow-wrap:anywhere;}.record span{color:#64748b;font-size:12px;font-weight:400;}.record.selected{border-color:#93c5fd;background:#eff6ff;}.record.selected strong{color:#1d4ed8;}
+.batch-record{position:relative;display:flex;gap:8px;align-items:stretch;margin-top:8px;}.batch-record>app-favorite-button{position:absolute;left:9px;top:9px;z-index:1}.batch-record .record{padding-left:44px;flex:1;min-width:0;width:auto;margin:0;}.batch-record .delete-batch{flex:0 0 40px;min-width:40px;padding:0;color:#b91c1c;}.batch-record .delete-batch:hover:not(:disabled){background:#fff1f2;border-color:#fda4af;}.batch-record .delete-batch .material-symbols-outlined{font-size:19px;}
+.archive-progress{margin:12px 0;padding:11px 13px;border:1px solid #bfdbfe;border-radius:10px;background:#f8fbff;color:#1e40af;}
+.archive-progress-heading{display:flex;align-items:center;gap:8px;}
+.archive-progress p{margin:7px 0 0;font-size:12px;color:#475569;}
+.archive-progress.archive-failed{border-color:#fecdd3;background:#fff1f2;color:#be123c;}
+.archive-track{height:7px;margin-top:10px;border-radius:999px;background:#dbeafe;overflow:hidden;}
+.archive-track span{display:block;height:100%;border-radius:inherit;background:#3b82f6;transition:width .3s ease;}
+.archive-spinner{width:13px;height:13px;border:2px solid #bfdbfe;border-top-color:#2563eb;border-radius:50%;animation:archive-spin .8s linear infinite;}
+@keyframes archive-spin{to{transform:rotate(360deg);}}
 .code{font-family:Consolas,monospace;font-size:12px;line-height:1.6;resize:vertical;background:#f8fafc;}.table-wrap{max-height:260px;overflow:auto;border:1px solid #e2e8f0;border-radius:8px;margin-bottom:14px;}table{width:100%;border-collapse:collapse;}th,td{padding:9px 10px;border-bottom:1px solid #e2e8f0;text-align:left;overflow-wrap:anywhere;}th{background:#f8fafc;color:#475569;font-weight:500;}
 .preview{white-space:pre-wrap;max-height:400px;overflow:auto;font:inherit;color:#334155;background:#f8fafc;border-radius:8px;padding:12px;border:1px solid #e2e8f0;}
 .job{display:flex;flex-wrap:wrap;gap:8px;align-items:center;padding:10px 0;border-bottom:1px solid #f1f5f9;}.job:last-child{border-bottom:0;}.job button{flex:1;justify-content:flex-start;text-align:left;min-width:150px;}.job span{font-size:12px;background:#f1f5f9;color:#475569;border-radius:5px;padding:3px 7px;}.job small{width:100%;}details{margin:14px 0;}summary{color:#475569;cursor:pointer;padding:8px 0;}
+.task-guidance{margin-top:16px;padding:14px;border:1px solid #bfdbfe;border-radius:10px;background:#f8fbff;}
+.task-guidance label{margin:8px 0 4px;}
+.report-overview{padding:14px;border:1px solid #dbeafe;border-radius:10px;background:#f8fbff;}
+.report-heading{display:flex;align-items:flex-start;gap:8px;}.report-heading .report-title{flex:1;min-width:0;}
+.report-title{display:block;font-size:14px;line-height:1.5;color:#0f172a;overflow-wrap:anywhere;}
+.report-badges{display:flex;gap:7px;flex-wrap:wrap;margin-top:9px;span{padding:3px 8px;border:1px solid #bfdbfe;border-radius:20px;background:#eff6ff;color:#1e40af;font-size:12px;&.failed{background:#fff1f2;border-color:#fecdd3;color:#be123c;}}}
+.report-error{margin-top:11px;padding:10px 12px;background:#fff1f2;border-left:3px solid #e11d48;border-radius:6px;color:#9f1239;strong{font-size:12px;}p{margin:4px 0 0;color:#9f1239;white-space:pre-wrap;}}
+.report-meta{display:grid;grid-template-columns:repeat(auto-fit,minmax(190px,1fr));gap:10px;margin:12px 0;div{min-width:0;padding:10px 12px;border:1px solid #e2e8f0;border-radius:8px;background:#fff;}strong{font-size:12px;color:#475569;}p{margin:5px 0 0;white-space:pre-wrap;color:#1e293b;}}
+.report-content{margin-top:14px;padding:14px;border:1px solid #e2e8f0;border-radius:10px;background:#fff;h3{margin-bottom:12px;}}
+.report-body{overflow-x:auto;overflow-wrap:anywhere;color:#1e293b;line-height:1.75;}
+:host ::ng-deep .report-body, :host ::ng-deep .report-markdown-body {
+  h1,h2,h3,h4{margin:18px 0 8px;color:#0f172a;line-height:1.4;}h1{font-size:18px;}h2{font-size:16px;}h3,h4{font-size:14px;}
+  > div{margin:3px 0;}
+  ul,ol{margin:8px 0;padding-left:24px;}li{margin:4px 0;}
+  blockquote{margin:10px 0;padding:8px 12px;border-left:3px solid #93c5fd;background:#f8fafc;}
+  pre{max-width:100%;overflow:auto;padding:10px 12px;background:#f1f5f9;border-radius:6px;white-space:pre-wrap;}
+  code{font-family:Consolas,monospace;background:#f1f5f9;padding:1px 3px;border-radius:3px;}
+  table{width:100%;border-collapse:collapse;font-size:12px;}th,td{padding:7px 9px;border:1px solid #e2e8f0;vertical-align:top;}th{background:#eff6ff;color:#1e3a8a;}
+  .md-para-gap{height:8px;}
+}

apps/showcase_ui/src/app/components/automation-panel/automation-panel.component.spec.ts

@@ -0,0 +1,214 @@
+import { signal } from '@angular/core';
+import { HttpClient } from '@angular/common/http';
+import { TestBed } from '@angular/core/testing';
+import { of } from 'rxjs';
+
+import { AgentService } from '../../services/agent.service';
+import { SystemService } from '../../services/system.service';
+import { AutomationPanelComponent } from './automation-panel.component';
+
+describe('AutomationPanelComponent batch deletion', () => {
+  const agent = {
+    currentSessionId: signal<string | null>(null),
+    runningSessionId: signal<string | null>(null),
+    agentStatus: signal('idle'),
+    sessions: signal<any[]>([]),
+    resultRevision: signal(0), batchFilter: signal(''), activeTab: signal('automation'),
+    selectSession: jasmine.createSpy('selectSession'), applyCaseReview: jasmine.createSpy('applyCaseReview')
+  };
+  const http = jasmine.createSpyObj<HttpClient>('HttpClient', ['get', 'post', 'delete']);
+
+  beforeEach(async () => {
+    (http.get.and as any).callFake((url: string) => of(url.endsWith('/preferences')
+      ? { formats: ['docx', 'xlsx'], batch_timeout_seconds: 1800 } : []));
+    http.delete.and.returnValue(of({ status: 'deleted' }));
+    http.post.and.returnValue(of({ status: 'ready', stage: '报告包已就绪', completed: 0, total: 0, elapsed_seconds: 0, eta_seconds: null, error: null }));
+    http.delete.calls.reset();
+    http.post.calls.reset();
+    agent.currentSessionId.set(null); agent.activeTab.set('automation'); agent.batchFilter.set(''); agent.selectSession.calls.reset();
+    await TestBed.configureTestingModule({
+      imports: [AutomationPanelComponent],
+      providers: [
+        { provide: AgentService, useValue: agent },
+        { provide: SystemService, useValue: { connectedDevices: signal([]) } },
+        { provide: HttpClient, useValue: http }
+      ]
+    }).compileComponents();
+  });
+
+  it('selects all imported cases by default, toggles all and submits only checked cases with rounds', async () => {
+    const tasks = [
+      { row: 2, case_id: '1.1', name: '打开设置', goal: '观察设置', task_config: { model: 'mimo-v2.6-pro' } },
+      { row: 3, case_id: '1.2', name: '观察首页', goal: '观察首页', task_config: {} }
+    ];
+    (http.post.and as any).callFake((url: string) => of(url.endsWith('/preview')
+      ? { valid: true, tasks, errors: [] } : url.endsWith('/batches')
+        ? { id: 'selected-batch', status: 'draft' } : { id: 'selected-batch', status: 'running' }));
+    const fixture = TestBed.createComponent(AutomationPanelComponent);
+    const component = fixture.componentInstance;
+    component.tab.set('batches');
+    component.upload({ target: { files: [new File(['test'], '用例.xlsx')] } } as unknown as Event);
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    fixture.detectChanges();
+    const checkboxes = () => [...fixture.nativeElement.querySelectorAll('.selection-column input')] as HTMLInputElement[];
+    expect(checkboxes().every(c => c.checked)).toBeTrue();
+    const toggle = fixture.nativeElement.querySelector('.selection-toolbar button') as HTMLButtonElement;
+    expect(toggle.textContent).toContain('全不选');
+    toggle.click(); fixture.detectChanges();
+    expect(component.selectedCount()).toBe(0);
+    expect((fixture.nativeElement.querySelector('.primary') as HTMLButtonElement).disabled).toBeTrue();
+    expect(toggle.textContent).toContain('全选');
+    toggle.click(); fixture.detectChanges();
+    expect(component.selectedCount()).toBe(2);
+    checkboxes()[1].click();
+    component.repeatCount = 3; component.device = 'test-device';
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.execution-summary').textContent).toContain('共 3 次执行');
+    component.startBatch();
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    const request = http.post.calls.allArgs().find(args => args[0] === '/api/automation/batches');
+    expect(request?.[1]).toEqual(jasmine.objectContaining({
+      cases: [{ case_id: '1.1', name: '打开设置', goal: '观察设置', task_config: { model: 'mimo-v2.6-pro' } }],
+      repeat_count: 3, device_serial: 'test-device'
+    }));
+  });
+
+  it('rejects empty selections and invalid or excessive loop counts without creating a batch', async () => {
+    const fixture = TestBed.createComponent(AutomationPanelComponent);
+    const component = fixture.componentInstance;
+    component.preview.set({ valid: true, tasks: [{ row: 2, name: '检查', goal: '观察' }], errors: [] });
+    component.device = 'test-device';
+    component.startBatch();
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    expect(component.error()).toContain('至少勾选');
+    component.selectedRows.set(new Set([2]));
+    for (const count of [0, 1.5, 10001]) {
+      component.repeatCount = count; component.startBatch();
+      await new Promise<void>(resolve => setTimeout(resolve, 0));
+      expect(component.error()).toContain('循环次数');
+    }
+    expect(http.post).not.toHaveBeenCalled();
+  });
+
+  it('confirms deletion and removes the batch from the visible list', async () => {
+    const fixture = TestBed.createComponent(AutomationPanelComponent);
+    fixture.componentInstance.tab.set('batches');
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    fixture.componentInstance.batches.set([
+      { id: 'batch-1', name: '批量冒烟', status: 'completed', created_at: 100 }
+    ]);
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('[aria-label="删除批次：批量冒烟"]')).not.toBeNull();
+
+    spyOn(window, 'confirm').and.returnValue(true);
+    const button = fixture.nativeElement.querySelector('.delete-batch') as HTMLButtonElement;
+    button.click();
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    fixture.detectChanges();
+    expect(http.delete).toHaveBeenCalledWith('/api/automation/batches/batch-1');
+    expect(fixture.componentInstance.batches()).toEqual([]);
+  });
+
+  it('keeps a running batch protected from the delete control', async () => {
+    const fixture = TestBed.createComponent(AutomationPanelComponent);
+    fixture.componentInstance.tab.set('batches');
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    fixture.componentInstance.batches.set([
+      { id: 'batch-2', name: '进行中', status: 'running', created_at: 100 }
+    ]);
+    fixture.detectChanges();
+    expect((fixture.nativeElement.querySelector('.delete-batch') as HTMLButtonElement).disabled).toBeTrue();
+  });
+
+  it('查看同批次用例时保留批次列表，可直接筛选该批次队列', () => {
+    const fixture = TestBed.createComponent(AutomationPanelComponent);
+    const component = fixture.componentInstance;
+    component.tab.set('batches');
+    component.batch.set({id:'batch-1',name:'冒烟批次',status:'completed',jobs:[
+      {session_id:'task-a',status:'completed',case:{name:'用例 A',goal:'观察设置'}},
+      {session_id:'task-b',status:'completed',case:{name:'用例 B',goal:'观察主页'}}
+    ]});
+    fixture.detectChanges();
+    const jobs: HTMLButtonElement[] = [...fixture.nativeElement.querySelectorAll('.job > button')];
+    jobs[0].click(); jobs[1].click();
+    expect(agent.selectSession.calls.allArgs()).toEqual([['task-a',true],['task-b',true]]);
+    expect(component.tab()).toBe('batches');
+    expect(fixture.nativeElement.querySelectorAll('.job').length).toBe(2);
+    const queueRequested = jasmine.createSpy('queueRequested');
+    component.queueRequested.subscribe(queueRequested);
+    component.showBatchQueue('batch-1');
+    expect(agent.batchFilter()).toBe('batch-1'); expect(agent.activeTab()).toBe('tasks');
+    expect(queueRequested).toHaveBeenCalled();
+    fixture.destroy();
+  });
+
+  it('shows the failure reason and renders Chinese report structure with line breaks', () => {
+    const fixture = TestBed.createComponent(AutomationPanelComponent);
+    fixture.detectChanges();
+    fixture.componentInstance.report.set({
+      id: 'task-1', name: '系统设置检查', status: 'failed', verdict: '未验证',
+      error: '模型连续 3 次未提供可执行工具调用', goal: '查看设置', expected: '',
+      content: '## 结果\n第一行\n第二行\n\n| 项目 | 状态 |\n| --- | --- |\n| 设置 | 未验证 |'
+    });
+    fixture.detectChanges();
+    const root = fixture.nativeElement as HTMLElement;
+    expect(root.querySelector('.report-error')?.textContent).toContain('连续 3 次');
+    expect(root.querySelector('.report-body h2')?.textContent).toBe('结果');
+    expect(root.querySelector('.report-body')?.textContent).toContain('第一行');
+    expect(root.querySelector('.report-body')?.textContent).toContain('第二行');
+    expect(root.querySelector('.report-body table td')?.textContent).toBe('设置');
+  });
+
+  it('prepares a batch archive before starting the browser download', async () => {
+    const fixture = TestBed.createComponent(AutomationPanelComponent);
+    const component = fixture.componentInstance;
+    component.tab.set('batches');
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    component.batch.set({ id: 'batch-3', name: '批量测试', status: 'completed', jobs: [] });
+    const click = spyOn(HTMLAnchorElement.prototype, 'click');
+    fixture.detectChanges();
+    (fixture.nativeElement.querySelector('.actions button:last-child') as HTMLButtonElement).click();
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    fixture.detectChanges();
+    expect(http.post).toHaveBeenCalledWith('/api/automation/batches/batch-3/archive/prepare', {});
+    expect(click).toHaveBeenCalled();
+    expect(fixture.nativeElement.querySelector('.archive-progress')?.textContent).toContain('下载已开始');
+  });
+
+  it('shows archive errors and lets the user retry', async () => {
+    http.post.and.returnValue(of({ status: 'failed', stage: '打包失败', completed: 1, total: 2, elapsed_seconds: 3, eta_seconds: null, error: '截图文件无法读取' }));
+    const fixture = TestBed.createComponent(AutomationPanelComponent);
+    const component = fixture.componentInstance;
+    component.tab.set('batches');
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    component.batch.set({ id: 'batch-4', name: '批量测试', status: 'completed', jobs: [] });
+    fixture.detectChanges();
+    (fixture.nativeElement.querySelector('.actions button:last-child') as HTMLButtonElement).click();
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.archive-progress')?.textContent).toContain('截图文件无法读取');
+    expect((fixture.nativeElement.querySelector('.actions button:last-child') as HTMLButtonElement).disabled).toBeFalse();
+  });
+
+  it('shows live archive progress and estimated remaining time while preparing', async () => {
+    http.post.and.returnValue(of({ status: 'preparing', stage: '正在打包任务 2/4', completed: 1, total: 4, elapsed_seconds: 12, eta_seconds: 36, error: null }));
+    (http.get.and as any).callFake((url: string) => of(url.endsWith('/status')
+      ? { status: 'ready', stage: '报告包已就绪', completed: 4, total: 4, elapsed_seconds: 13, eta_seconds: null, error: null }
+      : url.endsWith('/preferences') ? { formats: ['docx', 'xlsx'], batch_timeout_seconds: 1800 } : []));
+    const fixture = TestBed.createComponent(AutomationPanelComponent);
+    const component = fixture.componentInstance;
+    component.tab.set('batches');
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    component.batch.set({ id: 'batch-5', name: '批量测试', status: 'completed', jobs: [] });
+    spyOn(HTMLAnchorElement.prototype, 'click');
+    fixture.detectChanges();
+    (fixture.nativeElement.querySelector('.actions button:last-child') as HTMLButtonElement).click();
+    await new Promise<void>(resolve => setTimeout(resolve, 0));
+    fixture.detectChanges();
+    const progress = fixture.nativeElement.querySelector('.archive-progress') as HTMLElement;
+    expect(progress.textContent).toContain('已处理 1 / 4 个任务');
+    expect(progress.textContent).toContain('预计还需约 36 秒');
+    expect(progress.querySelector('.archive-spinner')).not.toBeNull();
+    fixture.destroy();
+  });
+});

apps/showcase_ui/src/app/components/automation-panel/automation-panel.component.ts

@@ -1,35 +1,102 @@
-import { Component, DestroyRef, effect, inject, signal } from '@angular/core';
+import { Component, DestroyRef, computed, effect, inject, output, signal } from '@angular/core';
+import { FavoriteButtonComponent } from '../favorite-button/favorite-button.component';
+import { ReportMarkdownComponent } from '../report-markdown/report-markdown.component';
 import { CommonModule } from '@angular/common';
 import { FormsModule } from '@angular/forms';
 import { HttpClient } from '@angular/common/http';
 import { firstValueFrom } from 'rxjs';
 import { AgentService } from '../../services/agent.service';
 import { SystemService } from '../../services/system.service';
+import { TestResultBadgeComponent } from '../test-result-badge/test-result-badge.component';
+import { TestResultCardComponent } from '../test-result-badge/test-result-card.component';
+import { CaseReviewResponse } from '../../core/models/session.model';
+import { TaskPluginsComponent, TaskPluginOptions } from '../task-plugins/task-plugins.component';
+
+interface ArchiveProgress {
+  status: 'idle' | 'preparing' | 'ready' | 'failed';
+  stage: string;
+  completed: number;
+  total: number;
+  elapsed_seconds: number;
+  eta_seconds: number | null;
+  error: string | null;
+}
 
 @Component({
-  selector: 'app-automation-panel', standalone: true, imports: [CommonModule, FormsModule],
+  selector: 'app-automation-panel', standalone: true, imports: [FavoriteButtonComponent, ReportMarkdownComponent, CommonModule, FormsModule, TestResultBadgeComponent, TestResultCardComponent, TaskPluginsComponent],
   templateUrl: './automation-panel.component.html', styleUrl: './automation-panel.component.scss'
 })
 export class AutomationPanelComponent {
+  taskPlugins: TaskPluginOptions = {};
+  queueRequested = output<void>();
   agent = inject(AgentService); system = inject(SystemService); http = inject(HttpClient);
   tab = signal('reports'); error = signal(''); notice = signal(''); busy = signal(false);
   batches = signal<any[]>([]); preview = signal<any>(null);
+  selectedRows = signal<Set<number>>(new Set());
+  selectedCount = computed(() => this.selectedRows().size);
   batch = signal<any>(null); report = signal<any>(null);
+  archiveProgress = signal<ArchiveProgress | null>(null);
+  archiveBatchId = signal(''); archiveBusy = signal(false);
+  guidance = '';
+  taskStatus = computed(() => {
+    const sid = this.agent.currentSessionId();
+    if (!sid) return '';
+    return this.agent.sessions().find(s => s.session_id === sid)?.status
+      || (this.agent.runningSessionId() === sid ? this.agent.agentStatus() : '') || '';
+  });
+  liveTask = computed(() => ['running', 'paused', 'task_paused', 'resuming'].includes(this.taskStatus()));
   word = true; excel = true; name = ''; device = ''; profile = 'flash'; stopOnFailure = false; timeout = 1800;
+  repeatCount = 1; maxExecutions = 10000;
   private destroyed = false;
   constructor() {
     const destroy = inject(DestroyRef);
-    this.http.get<any>('/api/automation/preferences').subscribe(p => { this.word = p.formats.includes('docx'); this.excel = p.formats.includes('xlsx'); this.timeout = p.batch_timeout_seconds || 1800; });
+    this.http.get<any>('/api/automation/preferences').subscribe(p => { this.word = p.formats.includes('docx'); this.excel = p.formats.includes('xlsx'); this.timeout = p.batch_timeout_seconds || 1800; this.maxExecutions = p.batch_max_executions || 10000; });
     void this.refresh();
     const timer = setInterval(() => void this.refresh(), 5000);
     destroy.onDestroy(() => { this.destroyed = true; clearInterval(timer); });
     effect(() => {
-      const sid = this.agent.currentSessionId(); this.report.set(null);
+      this.agent.resultRevision();
+      const sid = this.agent.currentSessionId(); this.report.set(null); this.guidance = '';
       if (sid) this.http.get<any>(`/api/automation/reports/${encodeURIComponent(sid)}`).subscribe({ next: r => { if (this.agent.currentSessionId() === sid) this.report.set(r); }, error: () => {} });
     });
   }
   formats() { return [ ...(this.word ? ['docx'] : []), ...(this.excel ? ['xlsx'] : []) ]; }
-  label(s: string) { return ({draft:'待开始',running:'执行中',paused:'已暂停',completed:'执行完成',failed:'失败',cancelled:'已取消',pending:'待执行',queued:'排队中',interrupted:'中断',skipped:'跳过',finalizing:'正在生成报告'} as Record<string,string>)[s] || s; }
+  label(s: string) { return ({draft:'待开始',running:'执行中',paused:'已暂停',task_paused:'等待指导',resuming:'等待设备恢复',completed:'执行完成',failed:'失败',cancelled:'已取消',pending:'待执行',queued:'排队中',interrupted:'中断',skipped:'跳过',finalizing:'正在生成报告',deleting:'删除中，可重试清理'} as Record<string,string>)[s] || s; }
+  reviewSaved(review: CaseReviewResponse) { this.agent.applyCaseReview(review); if (this.report()?.id === review.id) this.report.update(value => ({...value, ...review})); }
+  openBatchTask(id: string) { this.agent.selectSession(id, true); }
+  showBatchQueue(id: string) { this.agent.batchFilter.set(id); this.agent.activeTab.set('tasks'); this.queueRequested.emit(); }
+  duration(seconds: number) { const minutes = Math.floor(seconds / 60); return minutes ? `${minutes} 分 ${seconds % 60} 秒` : `${seconds} 秒`; }
+  downloadBatchArchive(id: string) { void this.prepareBatchArchive(id); }
+  private async prepareBatchArchive(id: string) {
+    if (this.archiveBusy()) return;
+    this.archiveBusy.set(true); this.archiveBatchId.set(id);
+    this.archiveProgress.set({ status: 'preparing', stage: '正在请求打包', completed: 0, total: 0, elapsed_seconds: 0, eta_seconds: null, error: null });
+    const base = `/api/automation/batches/${encodeURIComponent(id)}/archive`;
+    try {
+      let progress = await firstValueFrom(this.http.post<ArchiveProgress>(`${base}/prepare`, {}));
+      while (!this.destroyed && progress.status === 'preparing') {
+        this.archiveProgress.set(progress);
+        await new Promise<void>(resolve => setTimeout(resolve, 1000));
+        if (this.destroyed) return;
+        progress = await firstValueFrom(this.http.get<ArchiveProgress>(`${base}/status`));
+      }
+      if (this.destroyed) return;
+      this.archiveProgress.set(progress);
+      if (progress.status === 'failed') throw Error(progress.error || '批次报告打包失败');
+      if (progress.status !== 'ready') throw Error('打包任务已中断，请重试');
+      const link = document.createElement('a');
+      link.href = base;
+      link.download = `批次报告-${id.slice(0, 8)}.zip`;
+      document.body.appendChild(link); link.click(); link.remove();
+      this.notice.set('报告包已就绪，正在开始下载');
+    } catch (e: any) {
+      const detail = e.error?.detail;
+      this.archiveProgress.set({
+        ...(this.archiveProgress() || { completed: 0, total: 0, elapsed_seconds: 0, eta_seconds: null }),
+        status: 'failed', stage: '打包失败', error: typeof detail === 'string' ? detail : e.message || '下载失败'
+      });
+    } finally { this.archiveBusy.set(false); }
+  }
   async action(fn: () => Promise<void>) {
     if (this.busy()) return;
     this.busy.set(true); this.error.set(''); this.notice.set('');
@@ -50,17 +117,60 @@ export class AutomationPanelComponent {
   savePreferences() { void this.action(async () => { if (!this.formats().length) throw Error('请至少选择一种报告格式'); await firstValueFrom(this.http.put('/api/automation/preferences',{formats:this.formats()})); this.notice.set('已保存，后续任务和批次使用此默认格式'); }); }
   upload(event: Event) {
     const file = (event.target as HTMLInputElement).files?.[0]; if (!file) return;
-    void this.action(async () => { this.preview.set(null); const data = new FormData(); data.append('file',file); this.preview.set(await firstValueFrom(this.http.post('/api/automation/import/preview',data))); this.name = file.name.replace(/\.xlsx$/i,''); });
+    void this.action(async () => { this.preview.set(null); this.selectedRows.set(new Set()); const data = new FormData(); data.append('file',file); const preview: any = await firstValueFrom(this.http.post('/api/automation/import/preview',data)); this.preview.set(preview); this.selectedRows.set(new Set<number>(preview.tasks.map((t: any) => t.row))); this.repeatCount = 1; this.name = file.name.replace(/\.xlsx$/i,''); });
+  }
+  toggleRow(row: number) {
+    const selected = new Set(this.selectedRows());
+    if (selected.has(row)) selected.delete(row); else selected.add(row);
+    this.selectedRows.set(selected);
+  }
+  toggleAllRows() {
+    const rows: number[] = (this.preview()?.tasks || []).map((t: any) => t.row);
+    this.selectedRows.set(this.selectedCount() === rows.length ? new Set() : new Set(rows));
+  }
+  validRepetitions() {
+    return Number.isInteger(this.repeatCount) && this.repeatCount >= 1
+      && this.selectedCount() * this.repeatCount <= this.maxExecutions;
+  }
+  caseConfig(config: Record<string, unknown> | undefined) {
+    const overrides = Object.fromEntries(Object.entries(config || {}).filter(([, value]) => value !== null && value !== undefined));
+    return Object.keys(overrides).length ? JSON.stringify(overrides, null, 2) : '';
   }
   startBatch() { void this.action(async () => {
     if (!this.preview()?.valid) throw Error('请先导入并修正所有错误');
-    if (!this.device) throw Error('请选择执行设备');
+    if (!this.selectedCount()) throw Error('请至少勾选一个用例');
-    const cases = this.preview().tasks.map(({row, ...c}: any) => c);
+    if (!this.validRepetitions()) throw Error(`循环次数必须为正整数，执行总数不能超过 ${this.maxExecutions}`);
-    const created: any = await firstValueFrom(this.http.post('/api/automation/batches',{name:this.name.trim(),cases,device_serial:this.device,profile:this.profile,stop_on_failure:this.stopOnFailure,timeout_seconds:this.timeout,formats:this.formats()}));
+    const cases = this.preview().tasks.filter((t: any) => this.selectedRows().has(t.row)).map(({row, ...c}: any) => ({...c, task_config: {...this.taskPlugins, ...Object.fromEntries(Object.entries(c.task_config || {}).filter(([,v]) => v !== null && v !== undefined))}}));
+    if (!this.device && cases.some((c: any) => c.task_config.execution_target !== 'computer' && !c.task_config.device_serial)) throw Error('请选择执行设备');
+    const created: any = await firstValueFrom(this.http.post('/api/automation/batches',{name:this.name.trim(),cases,repeat_count:this.repeatCount,device_serial:this.device || null,profile:this.profile,stop_on_failure:this.stopOnFailure,timeout_seconds:this.timeout,formats:this.formats()}));
     this.batch.set(created);
     await firstValueFrom(this.http.post(`/api/automation/batches/${created.id}/resume`,{}));
     this.preview.set(null); this.notice.set('批次已开始，可暂停或查看每项结果'); await this.refresh();
   }); }
   selectBatch(id: string) { void this.action(async () => { this.batch.set(await firstValueFrom(this.http.get(`/api/automation/batches/${id}`))); }); }
+  deleteBatch(item: { id: string; name: string }) {
+    if (!window.confirm(`删除批次“${item.name}”及其全部任务记录、报告和文件？此操作无法撤销。`)) return;
+    void this.action(async () => {
+      await firstValueFrom(this.http.delete(`/api/automation/batches/${encodeURIComponent(item.id)}`));
+      if (this.batch()?.id === item.id) this.batch.set(null);
+      await this.refresh();
+      this.agent.fetchSessions();
+      this.notice.set('批次及关联任务、报告和文件已删除');
+    });
+  }
   batchAction(action: string) { void this.action(async () => { this.batch.set(await firstValueFrom(this.http.post(`/api/automation/batches/${this.batch().id}/${action}`,{}))); await this.refresh(); }); }
+  pauseTask(id: string) { void this.action(async () => { await firstValueFrom(this.http.post(`/api/tasks/${encodeURIComponent(id)}/pause`,{})); this.notice.set('已请求暂停：当前步骤结束后释放设备'); await this.refresh(); }); }
+  guideTask(id: string) { void this.action(async () => {
+    if (!this.guidance.trim()) throw Error('请先填写指导内容');
+    const status = this.taskStatus();
+    await firstValueFrom(this.http.post(`/api/tasks/${encodeURIComponent(id)}/guide`,{instruction:this.guidance.trim()}));
+    this.guidance = '';
+    this.agent.fetchGuidance(id);
+    if (status === 'paused') this.notice.set('模型请求暂停，请在执行日志中恢复模型调用');
+    await this.refresh();
+  }); }
+  continueTask(id: string) { void this.action(async () => {
+    await firstValueFrom(this.http.post(`/api/tasks/${encodeURIComponent(id)}/continue`,{}));
+    this.notice.set('任务已申请恢复；设备忙时等待排队'); await this.refresh();
+  }); }
 }

apps/showcase_ui/src/app/components/chat-interface/chat-interface.component.html

@@ -58,6 +58,8 @@ limitations under the License.
     @if (agentService.activeTab() === 'automation') { <app-automation-panel /> }
     <!-- Tasks Tab Content -->
     @if (agentService.activeTab() === 'tasks') {
+      <app-task-batch-filter />
+      @if (errorMessage()) { <p class="error-message" role="alert">{{ errorMessage() }}</p> }
       <!-- Active Queue Section -->
       <div class="queue-section-header">
         <span class="material-symbols-outlined section-icon">bolt</span> 当前队列 ({{ activeQueue().length }})
@@ -76,12 +78,17 @@ limitations under the License.
           (click)="selectTask(session.session_id)"
           >
           <div class="task-card-header">
+            <div class="task-status-group">
             <span class="task-badge" [ngClass]="getTaskStatus(session)">
               @if (getTaskStatus(session) === 'running') {
                 <span class="status-dot"></span>
               }
-              {{ getTaskStatus(session) | uiLabel }}
+              {{ agentService.explorationStatusLabel(session) || (getTaskStatus(session) | uiLabel) }}
             </span>
+            @if (session.exploration_kind !== 'exploration') { <app-test-result-badge [result]="session.test_result" [executionStatus]="getTaskStatus(session)" [compact]="true" /> }
+            <app-task-mode-badge [session]="session" />
+            @if (session.exploration_kind === 'validation') { <span class="task-origin-badge">用例复跑 · v{{session.case_version}}</span> }
+            </div>
             <div class="task-card-header-actions">
               @if (getDeviceSerial(session)) {
                 <span class="task-device" [title]="'Device: ' + getDeviceSerial(session)">
@@ -93,7 +100,7 @@ limitations under the License.
                 <span class="task-time-divider">·</span>
               }
               <span class="task-time">{{ session.start_time * 1000 | date:'shortTime' }}</span>
-              @if (getTaskStatus(session) === 'running' || getTaskStatus(session) === 'paused') {
+              @if (!agentService.canDeleteSession(session) && (getTaskStatus(session) === 'running' || getTaskStatus(session) === 'paused')) {
                 <button
                   type="button"
                   class="btn-stop-card"
@@ -104,7 +111,7 @@ limitations under the License.
                   <span class="material-symbols-outlined">stop_circle</span>
                 </button>
               }
-              @if (getTaskStatus(session) !== 'running' && getTaskStatus(session) !== 'paused') {
+              @if (agentService.canDeleteSession(session)) {
                 <button
                   type="button"
                   class="btn-delete-card"
@@ -118,11 +125,14 @@ limitations under the License.
               }
             </div>
           </div>
+          <div class="task-title-row">
+            <app-favorite-button [sourceId]="session.session_id" />
             <div class="task-goal" title="{{ session.initial_goal }}">
-            {{ session.initial_goal || '未命名任务' }}
+              {{ session.display_name || session.initial_goal || '未命名任务' }}
+            </div>
           </div>
           <div class="task-card-footer">
-            <span class="task-id" [title]="session.session_id">ID: {{ session.session_id }}</span>
+            <app-task-id-copy [taskId]="session.session_id" />
             <span class="task-date">{{ session.start_time * 1000 | date:'MM/dd/yy' }}</span>
           </div>
         </div>

apps/showcase_ui/src/app/components/chat-interface/chat-interface.component.scss

@@ -329,18 +329,25 @@ $text-muted: #a1a1aa;
 
       .task-card-header {
         display: flex;
-        justify-content: space-between;
+        justify-content: flex-start;
+        flex-wrap: wrap;
         align-items: center;
         gap: 8px;
         min-width: 0;
 
+        .task-status-group { display: inline-flex; align-items: center; gap: 6px; flex-wrap: wrap; min-width: 0; }
+        .task-origin-badge { display: inline-flex; align-items: center; box-sizing: border-box; height: 20px; padding: 0 6px; border: 1px solid #e2e8f0; border-radius: 5px; background: #f8fafc; color: #64748b; font-size: 11px; line-height: 14px; white-space: nowrap; }
+
         .task-badge {
           flex-shrink: 0;
           display: inline-flex;
           align-items: center;
-          font-size: 10.5px;
+          box-sizing: border-box;
+          height: 20px;
+          line-height: 14px;
+          font-size: 11px;
           font-weight: 700;
-          padding: 2px 5px;
+          padding: 0 5px;
           border-radius: 4px;
           letter-spacing: 0.05em;
           border: 1px solid #e2e8f0;
@@ -410,6 +417,7 @@ $text-muted: #a1a1aa;
         }
 
         .task-card-header-actions {
+          margin-left: auto;
           display: flex;
           align-items: center;
           gap: 4px;
@@ -551,6 +559,16 @@ $text-muted: #a1a1aa;
         }
       }
 
+      .task-title-row {
+        display: flex;
+        align-items: flex-start;
+        gap: 5px;
+        min-width: 0;
+
+        app-favorite-button { margin-top: -5px; }
+        .task-goal { flex: 1; min-width: 0; }
+      }
+
       .task-goal {
         font-size: 12.5px;
         font-weight: 500;

apps/showcase_ui/src/app/components/chat-interface/chat-interface.component.ts

@@ -1,3 +1,5 @@
+import { TaskIdCopyComponent } from '../task-id-copy/task-id-copy.component';
+import { FavoriteButtonComponent } from '../favorite-button/favorite-button.component';
 import { AutomationPanelComponent } from '../automation-panel/automation-panel.component';
 import { TaskHistoryComponent } from '../task-history/task-history.component';
 import { UiLabelPipe } from '../../utils/ui-label.pipe';
@@ -22,6 +24,9 @@ import { CommonModule } from '@angular/common';
 import { FormsModule } from '@angular/forms';
 import { AgentService } from '../../services/agent.service';
 import { Session } from '../../core/models/session.model';
+import { TestResultBadgeComponent } from '../test-result-badge/test-result-badge.component';
+import { TaskModeBadgeComponent } from '../task-mode-badge/task-mode-badge.component';
+import { TaskBatchFilterComponent } from '../task-batch-filter/task-batch-filter.component';
 import { MarkdownSegment, MarkdownLine, NoteMilestone, ParsedNote } from '../../core/models/markdown.model';
 import { parseNote, parseNoteLines } from '../../utils/markdown-parser.util';
 
@@ -30,7 +35,7 @@ export type { MarkdownSegment, MarkdownLine, NoteMilestone, ParsedNote };
 @Component({
   selector: 'app-chat-interface',
   standalone: true,
-  imports: [UiLabelPipe, CommonModule, FormsModule, AutomationPanelComponent, TaskHistoryComponent],
+  imports: [FavoriteButtonComponent, TaskIdCopyComponent, UiLabelPipe, CommonModule, FormsModule, AutomationPanelComponent, TaskHistoryComponent, TestResultBadgeComponent, TaskModeBadgeComponent, TaskBatchFilterComponent],
   templateUrl: './chat-interface.component.html',
   styleUrl: './chat-interface.component.scss',
   changeDetection: ChangeDetectionStrategy.OnPush
@@ -53,7 +58,7 @@ export class ChatInterfaceComponent {
   public activeQueue = computed(() => {
     const list = this.agentService.sessions().filter((s) => {
       const status = this.getTaskStatus(s);
-      return status === 'running' || status === 'paused' || status === 'pending';
+      return ['running', 'paused', 'task_paused', 'resuming', 'pending', 'queued'].includes(status) && this.agentService.matchesBatchFilter(s) && this.agentService.isQueueRepresentative(s);
     });
     return list.sort((a, b) => {
       const statusA = this.getTaskStatus(a);
@@ -158,7 +163,7 @@ export class ChatInterfaceComponent {
    */
   public deleteTask(sessionId: string, event: MouseEvent): void {
     event.stopPropagation();
-    if (!confirm("确定删除此任务？此操作无法撤销。")) {
+    if (!confirm(this.agentService.deletionPrompt(sessionId))) {
       return;
     }
     this.isSubmitting.set(true);
@@ -179,6 +184,9 @@ export class ChatInterfaceComponent {
    * Determine the current task execution status
    */
   public getTaskStatus(session: Session): 'running' | 'paused' | 'completed' | 'pending' | 'failed' | 'cancelled' {
+    const exploration = this.agentService.explorationQueueStatus(session);
+    if (exploration) return exploration as 'running' | 'paused' | 'completed' | 'pending' | 'failed' | 'cancelled';
+
     if (session.status) {
       const s = session.status.toLowerCase();
       if (s === 'completed' || s === 'success' || s === 'failed' || s === 'cancelled') {

apps/showcase_ui/src/app/components/exploration-panel/exploration-panel.component.html

@@ -0,0 +1,57 @@
+@if(session(); as s) {
+  <app-agent-stream [conversationMessages]="messages()" [demonstrations]="s.demonstrations" [pendingQuestion]="s.pending_question || ''" (useRecommendation)="fillRecommendation($event)">
+    <section task-summary class="session-summary">
+      <div class="row"><strong><span class="material-symbols-outlined">explore</span> {{activeValidation() ? '用例复跑中' : label(s.status)}}</strong><small>{{s.config.model}} / {{s.config.reasoning_effort}} · 续接 {{s.continuation_count}} 次</small></div>
+      <div class="exploration-goal"><app-favorite-button sourceKind="exploration" [sourceId]="s.id" label="收藏为重新探索" /><p>{{s.goal}}</p></div>
+      @if(s.pause_reason === 'execution_budget'){<p class="feedback">本段执行到时，进度已保存。可继续探索，不计为测试失败。</p>}
+      @if(s.pause_reason === 'device_observation'){<p class="feedback" role="status">设备画面暂时不可用，探索已暂停。指导已保留，请确认设备恢复响应后点击“继续探索”。</p>}
+      <details class="execution-history"><summary>执行记录（{{s.runs.length}}） · 当前展示所选轮次的操作步骤</summary>@for(run of s.runs; track run.id){<div class="run"><button class="run-link" (click)="agent.selectSession(run.id,true)">{{run.kind === 'validation' ? '用例 v' + run.case_version + ' 复跑' : '探索执行'}} · {{label(run.status)}}{{agent.currentSessionId() === run.id ? ' · 当前查看' : ''}}</button><app-test-result-badge [result]="run.test_result" [executionStatus]="run.status" [compact]="true" /></div>}</details>
+    </section>
+    <div task-result>
+    @if(showCase()) {
+    <section class="case-editor" aria-label="可复用用例"><div class="row"><strong class="case-title"><span class="material-symbols-outlined">task_alt</span>{{s.case ? '探索成果 · 可复用用例' : '本轮探索已结束'}}</strong><button (click)="generate()" [disabled]="busy() || !!s.active_run_id || generating() || dirty">{{generating() ? '正在提炼…' : s.case ? '重新提炼' : '提炼用例'}}</button></div>
+      @if(!s.case){<p class="muted">探索验收通过后自动生成；也可以先编辑并保存草稿。</p><button (click)="editorOpen.set(true)">编辑草稿</button>}
+      @if(s.generations[0]?.status === 'failed'){<p class="error">上次提炼未完成，执行结果保留。可重试提炼。</p>}
+      @if(s.case){<div class="case-compact"><h3><app-favorite-button sourceKind="exploration_case" [sourceId]="s.id" [version]="s.case.version" label="收藏此版本用例" /><span>{{s.case.name}} <small>v{{s.case.version}}</small></span></h3><p class="muted">已整理为一个可复用用例。你可以查看修改，或交给 Flash / Pro 独立复跑。</p><div class="actions"><button class="primary" (click)="copyCase()"><span class="material-symbols-outlined">content_copy</span>复制任务提示词</button><button (click)="editorOpen.set(!editorOpen())">{{editorOpen() ? '收起编辑' : '查看 / 编辑'}}</button><button (click)="rerunOpen.set(!rerunOpen())">复跑 / 批次</button></div></div>}
+      @if(editorOpen()) {
+      @if(s.case){<label>查看版本<select [ngModel]="editVersion" (ngModelChange)="selectVersion($event)" [disabled]="dirty">@for(version of versions(); track version){<option [value]="version">v{{version}}</option>}</select></label>}
+      @if(dirty && editVersion !== s.case_version){<p class="error">已有新版本；当前编辑仍保留，请先复制修改内容再加载最新版本。</p>}
+      <label>用例名称<input [(ngModel)]="editName" (ngModelChange)="dirty = true" maxlength="200" /></label><label>前置条件<textarea [(ngModel)]="editPreconditions" (ngModelChange)="dirty = true" rows="2" maxlength="10000"></textarea></label><label>任务提示词<textarea [(ngModel)]="editGoal" (ngModelChange)="dirty = true" rows="7" maxlength="10000"></textarea></label>
+      <div class="actions"><button (click)="saveCase()" [disabled]="busy() || !dirty || !editName.trim() || !editGoal.trim()">保存新版本</button><button (click)="copyCase()" [disabled]="!editGoal.trim()">复制提示词</button></div>
+      }
+      @if(rerunOpen()) {<div class="rerun-options"><datalist id="exploration-models">@for(model of models;track model){<option [value]="model"></option>}</datalist><strong>复跑与批次</strong><p class="muted">固定使用所选已保存版本。复跑通过与探索通过分别记录。</p><label>执行模式<select [(ngModel)]="rerunProfile"><option value="flash">Flash</option><option value="pro">Pro</option></select></label><label>复跑模型<input [(ngModel)]="rerunModel" list="exploration-models" placeholder="留空使用 Flash / Pro 默认模型" /></label><label>思考强度<select [(ngModel)]="rerunReasoning">@for(effort of efforts; track effort){<option [value]="effort">{{effort}}</option>}</select></label><div class="actions"><button (click)="rerun()" [disabled]="busy() || dirty || !editVersion || !!s.active_run_id">复跑此版本</button><button (click)="rerun('batch')" [disabled]="busy() || dirty || !editVersion">加入新批次草稿</button></div></div>}
+    </section>
+    }
+</div>
+  </app-agent-stream>
+  <section class="composer" [class.needs-answer]="!!s.pending_question" aria-label="指导当前探索">
+    @if(error()){<p role="alert" class="feedback error">{{error()}} <button (click)="refresh()">重试读取</button></p>}
+    @if(notice()){<p role="status" class="feedback">{{notice()}}</p>}
+    @if(s.pending_question) {
+      <div class="question-banner" role="status" aria-live="polite"><span class="material-symbols-outlined">chat_bubble</span><div><strong>探索已暂停，等待你回答</strong><p>{{pendingQuestion()?.text || '探索助手有一个问题需要你确认，请查看上方对话。'}}</p>@if(pendingQuestion()?.recommendation; as suggestion){<button (click)="fillRecommendation(suggestion)" [disabled]="busy()">填入建议回复</button>}</div></div>
+    } @else if(finalizing()) {
+      <div class="feedback" role="status">设备操作已结束，正在根据已验证的步骤提炼用例。完成后会在上方展示。</div>
+    }
+    @if(activeDemo(); as demo) {
+      <section class="demo" [class.warning]="demo.status === 'recording' && secondsLeft(demo) <= demo.warning_seconds" aria-live="polite"><strong><span class="record-dot" [class.recording]="demo.status === 'recording'"></span>{{label(demo.status)}}</strong>
+        @if(demo.status === 'recording'){<p>现在可以操作设备。剩余 {{secondsLeft(demo)}} 秒；AI 已暂停操作。</p><div class="actions"><button class="primary" (click)="demoAction('finish')" [disabled]="busy()">操作完成，开始分析</button><button (click)="demoAction('extend')" [disabled]="busy()">延长示范</button></div>}
+        @else {<p>{{demo.status === 'requested' ? '正在等待当前动作安全结束，请等“录制中”出现后再操作。' : '正在保存并分析本次示范录像。'}}</p>}
+        @if(['requested','recording'].includes(demo.status)){<button (click)="demoAction('cancel')" [disabled]="busy()">取消示范</button>}
+      </section>
+    } @else if(demoOptionsOpen() && s.active_run_id && !activeValidation() && ['running','waiting_user'].includes(s.status)) {
+      <div class="manual-control"><label>示范来源<select [(ngModel)]="source"><option value="android">安卓</option><option value="computer">电脑</option><option value="both">安卓 + 电脑</option></select></label><button (click)="action('demonstrations', {source})" [disabled]="busy()"><span class="material-symbols-outlined">touch_app</span>开始示范</button></div>
+    }
+
+    @if(activeValidation()){<p class="feedback">用例正在独立复跑，请进入复跑任务查看或指导。<button (click)="agent.selectSession(s.active_run_id!,true)">查看当前复跑</button></p>}
+    <div class="composer-heading"><label for="exploration-message">{{s.pending_question ? '回答探索助手' : '与探索助手交流'}}</label><span class="composer-status" [class.waiting]="!!s.pending_question">{{label(s.status)}}</span></div>
+    <textarea #messageInput id="exploration-message" [(ngModel)]="text" (keydown)="onKeyDown($event)" rows="2" maxlength="10000" [placeholder]="s.pending_question ? '在这里回答，发送后继续探索…' : '告诉我下一步怎么做，或补充你的想法…'" [disabled]="busy() || activeValidation() || finalizing()"></textarea>
+    <div class="actions composer-actions">
+      @if(!activeDemo() && s.active_run_id && !activeValidation() && ['running','waiting_user'].includes(s.status)){<button class="tool-button" (click)="demoOptionsOpen.set(!demoOptionsOpen())" [attr.aria-expanded]="demoOptionsOpen()"><span class="material-symbols-outlined">touch_app</span>我来操作</button>}
+      @if(!activeValidation() && !s.pending_question && (!s.active_run_id || s.status === 'paused')){<button (click)="action(s.status === 'draft' ? 'start' : 'continue')" [disabled]="busy()">{{s.status === 'draft' ? '开始探索' : '继续探索'}}</button>}
+      @if(s.active_run_id && !activeValidation() && !finalizing()){<button class="tool-button" (click)="action('pause')" [disabled]="busy() || ['paused','waiting_user','demonstrating'].includes(s.status)">暂停</button><button class="tool-button" (click)="action('stop')" [disabled]="busy()">停止</button>}
+      <small class="send-hint">Enter 发送 · Shift+Enter 换行</small>
+      <button class="primary send-button" (click)="send()" [disabled]="busy() || activeValidation() || finalizing() || !text.trim()" [attr.aria-label]="s.pending_question ? '回答并继续' : '发送'"><span>{{s.pending_question ? '回答并继续' : '发送'}}</span><span class="material-symbols-outlined">arrow_upward</span></button>
+    </div>
+    @if(!s.active_run_id){<p class="composer-footnote">可在此讨论和修订用例；继续操作设备请点击“继续探索”。</p>}
+  </section>
+} @else {<div class="empty" role="status">{{loading() ? '正在读取探索记录…' : '探索记录暂时无法读取'}} @if(error()){<p class="error">{{error()}}</p><button (click)="refresh()">重试</button>}</div>}

apps/showcase_ui/src/app/components/exploration-panel/exploration-panel.component.scss

@@ -0,0 +1,78 @@
+:host { display:flex; flex-direction:column; min-height:0; height:100%; overflow:hidden; color:#334155; font-size:13px; }
+.exploration-panel { padding:18px; display:flex; flex-direction:column; gap:16px; }
+header,.row,.actions { display:flex; align-items:center; gap:8px; flex-wrap:wrap; }
+header,.row { justify-content:space-between; }
+header strong { font-size:15px; color:#0f172a; }
+p { margin:8px 0; line-height:1.65; white-space:pre-wrap; overflow-wrap:anywhere; }
+header p,.muted,small { color:#64748b; font-size:12px; }
+button,input,select,textarea { font:inherit; box-sizing:border-box; }
+button { display:inline-flex; align-items:center; justify-content:center; gap:4px; border:1px solid #dbe3ec; border-radius:8px; background:white; color:#334155; padding:6px 10px; min-height:32px; cursor:pointer; }
+button:hover:enabled { background:#f1f5f9; border-color:#94a3b8; }
+button:disabled { opacity:.5; cursor:default; }
+button.primary { background:#0f172a; color:white; border-color:#0f172a; }
+.material-symbols-outlined { font-size:18px; }
+label { display:flex; flex-direction:column; gap:6px; color:#475569; margin:8px 0; }
+input,select,textarea { width:100%; min-width:0; border:1px solid #cbd5e1; border-radius:8px; padding:8px 10px; background:white; color:#0f172a; }
+textarea { resize:vertical; line-height:1.6; }
+input:focus,select:focus,textarea:focus { outline:2px solid #94a3b8; outline-offset:1px; }
+details { border:1px solid #e2e8f0; border-radius:10px; padding:10px 12px; }
+summary { cursor:pointer; font-size:12px; color:#475569; }
+.session-summary,.case-editor { padding:12px; border:1px solid #e2e8f0; border-radius:12px; background:#f8fafc; }
+.session-summary .actions { margin-top:10px; }
+.conversation { display:flex; flex-direction:column; gap:12px; }
+.message { padding:10px 12px; border-radius:12px; background:#f1f5f9; border:1px solid transparent; }
+.message.user { margin-left:20px; background:#ecfdf5; border-color:#d1fae5; }
+.message.system { padding:4px 8px; background:transparent; color:#64748b; font-size:12px; }
+.recommendation { border-left:2px solid #94a3b8; padding-left:10px; line-height:1.6; }
+.recommendation button { margin:6px 0 0 8px; }
+.composer { border-top:1px solid #e2e8f0; padding-top:4px; }
+.composer > button { margin-top:8px; }
+.demo { border:1px solid #bbf7d0; border-radius:12px; padding:12px; background:#f0fdf4; }
+.demo.warning { background:#fffbeb; border-color:#fcd34d; }
+.record-dot { display:inline-block; width:7px; height:7px; border-radius:50%; background:#94a3b8; margin-right:6px; }
+.record-dot.recording { background:#dc2626; animation:pulse 1.8s infinite; }
+@keyframes pulse { 50% { opacity:.4; } }
+@media(prefers-reduced-motion:reduce) { .record-dot.recording { animation:none; } }
+.feedback { padding:9px 12px; background:#f0fdf4; border-radius:8px; font-size:12px; }
+.error { color:#b91c1c; }
+.feedback.error { background:#fef2f2; }
+.run,.demo-history { padding:10px 0; border-bottom:1px solid #e2e8f0; }
+.run-link { border:none; padding-left:0; background:transparent; text-align:left; }
+video { width:100%; max-height:260px; margin-top:8px; background:#0f172a; }
+a { color:#0f766e; font-size:12px; }
+.empty { color:#64748b; text-align:center; padding:25px 10px; }
+.empty .material-symbols-outlined { font-size:32px; }
+.settings-fields,.case-editor { display:flex; flex-direction:column; gap:3px; }
+
+app-agent-stream {flex:1;min-height:0;display:block}
+.composer {flex:none;margin:0 20px 16px;padding:12px 16px;background:#fff;border:1px solid #dbe3ec;border-radius:16px;box-shadow:0 4px 18px #0f172a06;max-height:45%;overflow:auto}
+.composer label {margin:0 0 6px}.composer textarea {resize:none}.composer-actions {margin-top:8px}.composer-actions small {flex:1}
+.session-summary {margin-bottom:18px;background:#fff}.session-summary p {margin:6px 0}.session-summary .row strong {display:flex;align-items:center;gap:6px}
+.execution-history {margin-top:8px;border:0;padding:4px 0}.case-editor {margin:16px 0;background:#fff}.case-compact {display:flex;flex-wrap:wrap;align-items:center;gap:8px}.case-compact .actions {margin-left:auto}
+.demo {padding:8px 10px;margin-bottom:8px}.demo p {margin:4px 0}
+
+.composer {margin:0 24px 18px;padding:14px 18px;border-radius:22px;box-shadow:0 6px 24px #0f172a08;max-height:50%;box-sizing:border-box;min-width:0}
+.composer-heading {display:flex;align-items:center;justify-content:space-between;gap:8px;margin-bottom:4px}
+.composer-heading label {margin:0;font-size:12px;color:#64748b}
+.composer .manual-control {display:flex;align-items:center;justify-content:flex-end;gap:8px;flex-wrap:wrap;margin:0 0 10px}
+.composer .manual-control label {display:flex;flex-direction:row;align-items:center;gap:8px;margin:0;min-width:0}
+.composer .manual-control select {width:auto;min-width:110px;height:36px;padding:6px 10px}
+.composer .manual-control button {align-self:center;height:36px;white-space:nowrap}
+.composer-status {font-size:11px;border-radius:20px;padding:3px 8px;background:#eef2ff;color:#4f46e5}
+.composer-status.waiting {color:#92400e;background:#fef3c7}
+.composer textarea {border:0;border-radius:0;padding:10px 0;background:transparent;min-height:66px;max-height:140px;font-size:14px;box-shadow:none}
+.composer textarea:focus {outline:none}
+.composer:focus-within {border-color:#a5b4fc;box-shadow:0 0 0 3px #eef2ff}
+.composer-actions {gap:6px;flex-wrap:wrap;border-top:1px solid #f1f5f9;padding-top:10px}
+.composer-actions .tool-button {border:0;background:#f8fafc;color:#64748b;border-radius:10px;padding:6px 9px}
+.composer-actions .send-hint {flex:1;text-align:right;font-size:10px;color:#94a3b8;min-width:100px}
+.send-button {min-height:36px;border-radius:12px;padding:7px 12px;background:#4f46e5!important;border-color:#4f46e5!important;gap:6px}
+.composer-footnote {font-size:11px;color:#94a3b8;margin:8px 0 0}
+.composer.needs-answer {border-color:#f4c97a}
+.question-banner {display:flex;gap:10px;background:#fffbeb;border:1px solid #fde68a;border-radius:12px;padding:12px;margin-bottom:12px;color:#92400e}
+.question-banner > .material-symbols-outlined {font-size:22px;margin-top:2px}.question-banner p {margin:4px 0;font-size:13px}.question-banner strong {font-size:13px}.question-banner button {margin-top:5px;background:#fff7d6;border-color:#f4c97a;color:#92400e}
+.case-editor {border-color:#c7d2fe;background:linear-gradient(135deg,#f5f7ff,#fff);padding:20px;border-radius:18px}
+.case-title {display:flex;align-items:center;gap:8px;color:#4f46e5;font-size:14px}.case-title .material-symbols-outlined {font-size:22px}
+.case-compact {display:block}.case-compact h3 {font-size:18px;line-height:1.5;color:#1e293b;margin:14px 0 6px}.case-compact small {font-size:11px;font-weight:400;background:#e0e7ff;color:#4338ca;padding:3px 6px;border-radius:5px;white-space:nowrap}.case-compact .actions {margin:16px 0 0}.case-compact .primary {background:#4f46e5;border-color:#4f46e5}
+.session-summary {padding:12px 16px;border-radius:14px}.exploration-goal{display:flex;align-items:flex-start;gap:6px}.session-summary .exploration-goal p{flex:1;min-width:0;margin:4px 0}.case-compact h3{display:flex;align-items:flex-start;gap:6px}.case-compact h3>span{min-width:0;overflow-wrap:anywhere}.session-summary p {color:#64748b;font-size:12px}
+@media(max-width:600px) {.composer {margin:0 12px 12px;padding:12px}.send-hint {display:none}.send-button {margin-left:auto}.question-banner {max-height:160px;overflow:auto}.case-editor {padding:14px}}

apps/showcase_ui/src/app/components/exploration-panel/exploration-panel.component.spec.ts

@@ -0,0 +1,119 @@
+import { Component, input, output, signal } from '@angular/core';
+import { TestBed, fakeAsync, flushMicrotasks } from '@angular/core/testing';
+import { By } from '@angular/platform-browser';
+import { ExplorationPanelComponent } from './exploration-panel.component';
+import { ExplorationService, ExplorationSession } from '../../services/exploration.service';
+import { AgentService } from '../../services/agent.service';
+
+import { AgentStreamComponent } from '../agent-stream/agent-stream.component';
+@Component({selector:'app-agent-stream',standalone:true,template:'<ng-content></ng-content>'}) class StreamStub {conversationMessages=input<any[]>([]);demonstrations=input<any[]>([]);pendingQuestion=input('');useRecommendation=output<string>();}
+
+describe('Exploration conversation', () => {
+  const config = {model: 'gpt-6.1-sol', reasoning_effort: 'high', timeout_seconds: 1800, demonstration_timeout_seconds: 900, demonstration_warning_seconds: 60};
+  let current: ExplorationSession;
+  let api: any;
+  beforeEach(() => {
+    current = {id:'s', goal:'检查设置', status:'running', config, active_run_id:'run', active_demonstration_id:null, case_version:0, message_count:0, elapsed_seconds:10, continuation_count:0, messages:[], runs:[], case:null, demonstrations:[], generations:[]};
+    api = {drafts: new Map(),caseDrafts:new Map(),requests:new Map(),settings: jasmine.createSpy().and.resolveTo(config), list:jasmine.createSpy().and.resolveTo([]), get:jasmine.createSpy().and.callFake(async () => structuredClone(current)), act:jasmine.createSpy().and.resolveTo({}), version:jasmine.createSpy(), saveSettings:jasmine.createSpy().and.resolveTo(config)};
+    TestBed.configureTestingModule({imports:[ExplorationPanelComponent], providers:[
+      {provide:ExplorationService,useValue:api},
+      {provide:AgentService,useValue:{selectSession:jasmine.createSpy(),fetchSessions:jasmine.createSpy(),fetchStatus:jasmine.createSpy(),currentSessionId:signal('run')}}
+    ]}).overrideComponent(ExplorationPanelComponent,{remove:{imports:[AgentStreamComponent]},add:{imports:[StreamStub]}});
+  });
+  function render() {
+    const fixture = TestBed.createComponent(ExplorationPanelComponent);
+    fixture.componentRef.setInput('sessionId','s');
+    fixture.detectChanges(); flushMicrotasks(); fixture.detectChanges();
+    return fixture;
+  }
+  it('highlights unanswered questions beside the composer and hides unfinished cases', fakeAsync(() => {
+    current.status='waiting_user'; current.pending_question='q';
+    current.messages=[{id:'q',sequence:1,role:'assistant',kind:'question',text:'下一步是否回到主页？',created_at:1}];
+    current.case={version:1,name:'旧用例',goal:'旧目标',preconditions:'',evidence:[]};current.case_version=1;
+    const fixture=render();
+    expect(fixture.nativeElement.querySelector('.question-banner').textContent).toContain('下一步是否回到主页');
+    expect(fixture.nativeElement.querySelector('[aria-label="可复用用例"]')).toBeNull();
+    expect(fixture.nativeElement.textContent).not.toContain('继续探索');
+    current.active_run_id=null;current.status='completed';current.pending_question=undefined;
+    void fixture.componentInstance.refresh();flushMicrotasks();fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('[aria-label="可复用用例"]').textContent).toContain('旧用例');
+    expect(fixture.nativeElement.querySelector('.question-banner')).toBeNull();fixture.destroy();
+  }));
+  it('keeps a draft and disables guidance while the verified run is being synthesized', fakeAsync(() => {
+    current.status='finalizing';
+    const fixture=render();fixture.componentInstance.text='返回主页';
+    void fixture.componentInstance.send();flushMicrotasks();
+    expect(api.act).not.toHaveBeenCalled();expect(fixture.componentInstance.text).toBe('返回主页');
+    expect(fixture.nativeElement.querySelector('textarea').disabled).toBeTrue();fixture.destroy();
+  }));
+  it('fills a suggested reply for review from either entry point and only sends after submission', fakeAsync(() => {
+    current.status='waiting_user'; current.pending_question='q';
+    current.messages=[{id:'q',sequence:1,role:'assistant',kind:'question',text:'是否按示范流程继续？',recommendation:'确认，按示范流程继续探索。',created_at:1}];
+    const fixture=render();
+    fixture.nativeElement.querySelector('.question-banner button').click();
+    fixture.detectChanges(); flushMicrotasks();
+    const input: HTMLTextAreaElement = fixture.nativeElement.querySelector('#exploration-message');
+    expect(input.value).toBe('确认，按示范流程继续探索。');
+    expect(document.activeElement).toBe(input);
+    expect(api.act).not.toHaveBeenCalled();
+    const stream = fixture.debugElement.query(By.directive(StreamStub)).componentInstance as StreamStub;
+    stream.useRecommendation.emit(current.messages[0].recommendation!);
+    expect(api.act).not.toHaveBeenCalled();
+    fixture.componentInstance.text='确认，按示范流程继续，最后返回顶部。';
+    fixture.detectChanges();
+    fixture.nativeElement.querySelector('.send-button').click(); flushMicrotasks();
+    expect(api.act).toHaveBeenCalledTimes(1);
+    expect(api.act.calls.mostRecent().args.slice(0,3)).toEqual(['s','messages',{text:'确认，按示范流程继续，最后返回顶部。'}]);
+    expect(fixture.componentInstance.text).toBe('');
+    fixture.destroy();
+  }));
+  it('only invites manual input when the recorder is ready, and warns before its deadline', fakeAsync(() => {
+    current.active_demonstration_id='demo';
+    current.demonstrations=[{id:'demo',run_id:'run',source:'android',status:'requested',warning_seconds:60,media:[]}];
+    const fixture=render();
+    expect(fixture.nativeElement.textContent).toContain('请等“录制中”出现后再操作');
+    expect(fixture.nativeElement.textContent).not.toContain('现在可以操作设备');
+    current.demonstrations[0]={...current.demonstrations[0],status:'recording',deadline:Date.now()/1000+30};
+    void fixture.componentInstance.refresh(); flushMicrotasks(); fixture.detectChanges();
+    expect(fixture.nativeElement.textContent).toContain('现在可以操作设备');
+    expect(fixture.nativeElement.querySelector('.demo.warning')).not.toBeNull();
+    fixture.destroy();
+  }));
+  it('shows budget exhaustion as a resumable pause, never a failed test', fakeAsync(() => {
+    current.status='paused'; current.pause_reason='execution_budget'; current.active_run_id=null;
+    current.case={version:1,name:'旧用例',goal:'旧目标',preconditions:'',evidence:[]};
+    const fixture=render();
+    expect(fixture.nativeElement.textContent).toContain('不计为测试失败');
+    expect(fixture.nativeElement.querySelector('[aria-label="可复用用例"]')).toBeNull();
+    const button=[...fixture.nativeElement.querySelectorAll('button')].find((b:any)=>b.textContent.includes('继续探索')) as HTMLButtonElement;
+    button.click(); flushMicrotasks();
+    expect(api.act.calls.mostRecent().args[1]).toBe('continue');
+    fixture.destroy();
+  }));
+  it('retains the request ID and unsent text when a send response is lost', fakeAsync(() => {
+    const fixture=render();
+    api.act.and.rejectWith(new Error('network'));
+    fixture.componentInstance.text='改用右上角入口';
+    void fixture.componentInstance.send(); flushMicrotasks();
+    const first=api.act.calls.mostRecent().args[3];
+    expect(fixture.componentInstance.text).toBe('改用右上角入口');
+    api.act.and.resolveTo({});
+    void fixture.componentInstance.send(); flushMicrotasks();
+    expect(api.act.calls.mostRecent().args[3]).toBe(first);
+    expect(fixture.componentInstance.text).toBe('');
+    fixture.destroy();
+  }));
+  it('preserves unsaved edits across polling, but clears them when choosing another conversation', fakeAsync(() => {
+    current.case_version=1; current.case={version:1,name:'检查',goal:'旧目标',preconditions:'',evidence:[]};
+    const fixture=render();
+    fixture.componentInstance.editGoal='用户草稿'; fixture.componentInstance.dirty=true;
+    current.case={...current.case,version:2,goal:'其他窗口保存'}; current.case_version=2;
+    void fixture.componentInstance.refresh(); flushMicrotasks();
+    expect(fixture.componentInstance.editGoal).toBe('用户草稿');
+    current={...current,id:'other',case:null,case_version:0}; fixture.componentRef.setInput('sessionId','other');
+    fixture.detectChanges(); flushMicrotasks(); fixture.detectChanges();
+    expect(fixture.componentInstance.editGoal).toBe('');
+    expect(api.caseDrafts.get('s')?.goal).toBe('用户草稿');
+    fixture.destroy();
+  }));
+});

apps/showcase_ui/src/app/components/exploration-panel/exploration-panel.component.ts

@@ -0,0 +1,172 @@
+import { Component, ChangeDetectionStrategy, DestroyRef, ElementRef, effect, inject, signal, untracked, input, viewChild } from '@angular/core';
+import { FavoriteButtonComponent } from '../favorite-button/favorite-button.component';
+import { FormsModule } from '@angular/forms';
+import { AgentStreamComponent } from '../agent-stream/agent-stream.component';
+import { ExplorationService, ExplorationSession, ExplorationMessage, Demonstration } from '../../services/exploration.service';
+import { AgentService } from '../../services/agent.service';
+import { TestResultBadgeComponent } from '../test-result-badge/test-result-badge.component';
+
+@Component({
+  selector: 'app-exploration-panel', standalone: true,
+  imports: [FormsModule, AgentStreamComponent, TestResultBadgeComponent, FavoriteButtonComponent],
+  templateUrl: './exploration-panel.component.html', styleUrl: './exploration-panel.component.scss',
+  changeDetection: ChangeDetectionStrategy.OnPush
+})
+export class ExplorationPanelComponent {
+  api = inject(ExplorationService);
+  agent = inject(AgentService);
+  private destroyRef = inject(DestroyRef);
+  session = signal<ExplorationSession | null>(null);
+  messages = signal<ExplorationMessage[]>([]);
+  loading = signal(false);
+  busy = signal(false);
+  error = signal('');
+  notice = signal('');
+  now = signal(Date.now() / 1000);
+  sessionId = input.required<string>();
+  editorOpen = signal(false);
+  rerunOpen = signal(false);
+  demoOptionsOpen = signal(false);
+  private messageInput = viewChild<ElementRef<HTMLTextAreaElement>>('messageInput');
+  get text() { return this.api.drafts.get(this.sessionId()) || ''; }
+  set text(value: string) { this.api.drafts.set(this.sessionId(), value); }
+  source = 'android';
+  editName = '';
+  editGoal = '';
+  editPreconditions = '';
+  editVersion = 0;
+  dirty = false;
+  rerunProfile = 'flash';
+  rerunModel = '';
+  rerunReasoning = 'high';
+  private polling = false;
+  private refreshSerial = 0;
+  private followsLatest = true;
+  private requests = this.api.requests;
+  private displayedId: string | null = null;
+  readonly models = ['gpt-6.1-sol', 'gpt-6-astra', 'gpt-6-luna', 'mimo-v2.6-pro', 'mimo-v2.6-flash'];
+  readonly efforts = ['default', 'none', 'minimal', 'low', 'medium', 'high', 'xhigh', 'max'];
+
+  constructor() {
+    effect(() => {
+      const id = this.sessionId();
+      untracked(() => {
+        this.rememberDraft(); this.displayedId = id;
+        this.session.set(null); this.messages.set([]); this.dirty = false; this.editVersion = 0; this.followsLatest = true;
+        this.editName = ''; this.editGoal = ''; this.editPreconditions = '';
+        this.editorOpen.set(false); this.rerunOpen.set(false); this.error.set(''); this.notice.set('');
+        const draft = this.api.caseDrafts.get(id);
+        if (draft) { this.loadCase(draft); this.dirty = true; this.editorOpen.set(true); }
+        if (id) void this.refresh(id);
+      });
+    });
+    const interval = setInterval(() => {
+      this.now.set(Date.now() / 1000);
+      const id = this.sessionId();
+      if (id && !this.polling) void this.refresh(id, false);
+    }, 2000);
+    this.destroyRef.onDestroy(() => { this.rememberDraft(); this.refreshSerial++; clearInterval(interval); });
+  }
+
+  private rememberDraft() {
+    if (!this.displayedId) return;
+    if (this.dirty) this.api.caseDrafts.set(this.displayedId, {name:this.editName,goal:this.editGoal,preconditions:this.editPreconditions,version:this.editVersion});
+    else this.api.caseDrafts.delete(this.displayedId);
+  }
+  async refresh(id = this.sessionId(), initial = true) {
+    if (!id || this.destroyRef.destroyed) return;
+    const serial = ++this.refreshSerial;
+    if (initial) this.loading.set(true);
+    this.polling = true;
+    try {
+      const current = this.session();
+      const after = current?.id === id ? (this.messages().at(-1)?.sequence || 0) : 0;
+      const value = await this.api.get(id, after);
+      if (id !== this.sessionId() || serial !== this.refreshSerial) return;
+      const combined = new Map((current?.id === id ? this.messages() : []).map(m => [m.id, m]));
+      value.messages.forEach(m => combined.set(m.id, m));
+      this.messages.set([...combined.values()].sort((a, b) => a.sequence - b.sequence));
+      this.session.set(value);
+      if (!current) this.source = value.task_config?.['execution_target'] === 'computer' ? 'computer' : 'android';
+      if (value.case && this.followsLatest && !this.dirty && this.editVersion !== value.case.version) this.loadCase(value.case);
+    } catch (e) { if (id === this.sessionId()) this.showError(e); }
+    finally { if (serial === this.refreshSerial) { this.polling = false; this.loading.set(false); } }
+  }
+  private loadCase(value: {name: string; goal: string; preconditions: string; version: number}) {
+    this.editName = value.name; this.editGoal = value.goal; this.editPreconditions = value.preconditions;
+    this.editVersion = value.version; this.dirty = false;
+  }
+  async selectVersion(version: string) {
+    const id = this.sessionId(); if (!id || this.dirty) return;
+    this.followsLatest = Number(version) === this.session()?.case_version;
+    try { const value = await this.api.version(id, Number(version)); if (id === this.sessionId()) this.loadCase(value); }
+    catch(e) { this.showError(e); }
+  }
+  versions() { return Array.from({length: this.session()?.case_version || 0}, (_, i) => i + 1); }
+  label(status: string) {
+    return ({draft:'未开始', queued:'等待执行', reserved:'等待执行', running:'探索中', waiting_user:'等待你回答', demonstrating:'人工示范', analyzing_demo:'分析示范', finalizing:'正在提炼用例', paused:'已暂停，可继续', completed:'探索已结束', cancelled:'已取消', failed:'执行失败', interrupted:'已中断，可继续', requested:'等待安全接管', recording:'录制中', analyzing:'分析中', timed_out:'示范到时，已保存'} as Record<string,string>)[status] || status;
+  }
+  pendingQuestion() { return this.messages().find(m => m.id === this.session()?.pending_question); }
+  finalizing() { return this.session()?.status === 'finalizing'; }
+  showCase() { const s = this.session(); return !!s && !s.active_run_id && ['completed', 'cancelled'].includes(s.status); }
+  activeDemo() { const s = this.session(); return s?.demonstrations.find(d => d.id === s.active_demonstration_id); }
+  activeValidation() { const s = this.session(); return s?.runs.some(r => r.id === s.active_run_id && r.kind === 'validation') || false; }
+  secondsLeft(demo: Demonstration) { return Math.max(0, Math.ceil((demo.deadline || this.now()) - this.now())); }
+  async action(action: string, payload: Record<string, unknown> = {}): Promise<Record<string, unknown> | null> {
+    const id = this.sessionId(); if (!id || this.busy()) return null;
+    const selectedRun = this.agent.currentSessionId();
+    const key = JSON.stringify([id, action, payload]);
+    const requestId = this.requests.get(key) || crypto.randomUUID(); this.requests.set(key, requestId);
+    this.busy.set(true); this.error.set(''); this.notice.set('');
+    try {
+      const value = await this.api.act(id, action, payload, requestId);
+      this.requests.delete(key);
+      await this.refresh(id, false);
+      const run = value['run'] as {id?: string} | undefined;
+      if (id === this.sessionId() && selectedRun === this.agent.currentSessionId() && run?.id) { this.agent.selectSession(run.id, true); }
+      this.agent.fetchSessions(); this.agent.fetchStatus();
+      return id === this.sessionId() ? value : null;
+    } catch (e) { if (id === this.sessionId()) this.showError(e); return null; }
+    finally { this.busy.set(false); }
+  }
+  async send() {
+    const text = this.text.trim(), id = this.sessionId(); if (!text || this.activeValidation() || this.finalizing()) return;
+    if (await this.action('messages', {text}) && this.api.drafts.get(id)?.trim() === text) this.api.drafts.delete(id);
+  }
+  fillRecommendation(reply: string) {
+    if (this.busy() || this.activeValidation() || this.finalizing() || this.pendingQuestion()?.recommendation !== reply) return;
+    this.text = reply;
+    this.messageInput()?.nativeElement.focus();
+  }
+  demoAction(action: string) {
+    const demo = this.activeDemo();
+    if (demo) void this.action(`demonstrations/${demo.id}/${action}`, action === 'extend' ? {seconds: this.session()?.config.demonstration_timeout_seconds || 900} : {});
+  }
+  onKeyDown(event: KeyboardEvent) {
+    if (event.key === 'Enter' && !event.shiftKey && !event.isComposing) { event.preventDefault(); void this.send(); }
+  }
+  async saveCase() {
+    if (!this.editName.trim() || !this.editGoal.trim()) return;
+    const result = await this.action('case', {expected_version: this.editVersion, case: {name: this.editName, goal: this.editGoal, preconditions: this.editPreconditions, evidence: this.session()?.case?.evidence || []}});
+    if (result) { this.dirty = false; this.followsLatest = true; await this.refresh(); this.notice.set('用例已保存为新版本'); }
+  }
+  generate() {
+    const source = [...(this.session()?.runs || [])].reverse().find(r => r.kind === 'exploration' && r.test_result?.status === 'passed' && r.status === 'completed');
+    if (source) void this.action('generate', {source_run_id: source.id, expected_version: this.session()?.case_version || 0});
+    else this.error.set('尚无验证通过的探索执行，请继续探索；也可以手动保存草稿。');
+  }
+  generating() { return this.session()?.generations.some(g => g.status === 'running') || false; }
+  rerun(action = 'validate') {
+    const task_config: Record<string, unknown> = {};
+    if (this.rerunModel.trim()) { task_config['model'] = this.rerunModel.trim(); task_config['reasoning_effort'] = this.rerunReasoning; }
+    const payload: Record<string, unknown> = {version: this.editVersion, profile: this.rerunProfile, task_config};
+    if (action === 'batch') payload['name'] = this.editName;
+    void this.action(action, payload).then(value => { if (value && action === 'batch') this.notice.set('已创建批次草稿，可在“报告与批量”中查看并启动。'); });
+  }
+  async copyCase() {
+    try { await navigator.clipboard.writeText(`${this.editPreconditions ? '前置条件：' + this.editPreconditions + '\n\n' : ''}${this.editGoal}`); this.notice.set('用例提示词已复制'); }
+    catch { this.error.set('剪贴板不可用，请选中下方提示词复制'); }
+  }
+  mediaUrl(runId: string, path: string) { return '/videos/' + [runId, ...path.split('/')].map(encodeURIComponent).join('/'); }
+  private showError(error: unknown) { const e = error as {error?: {detail?: unknown}; message?: string}; this.error.set(typeof e.error?.detail === 'string' ? e.error.detail : e.message || '操作失败，请刷新状态后重试'); }
+}

apps/showcase_ui/src/app/components/favorite-button/favorite-button.component.ts

@@ -0,0 +1,32 @@
+import { Component, ChangeDetectionStrategy, computed, inject, input, signal } from '@angular/core';
+import { FavoriteKind, FavoritesService } from '../../services/favorites.service';
+
+@Component({
+  selector: 'app-favorite-button', standalone: true, changeDetection: ChangeDetectionStrategy.OnPush,
+  template: `<button type="button" [class.saved]="!!favorite()" [disabled]="busy() || !api.loaded() || !sourceId()"
+    [attr.aria-pressed]="!!favorite()" [attr.aria-label]="favorite() ? '取消收藏' : label()" [title]="favorite() ? '取消收藏' : label()"
+    (click)="toggle($event)"><span class="material-symbols-outlined" aria-hidden="true">{{favorite() ? 'star' : 'star_outline'}}</span></button>
+    @if(error()){<span class="feedback" role="alert">{{error()}}</span>}
+    @if(api.error() && !api.loaded()){<button type="button" class="retry" (click)="$event.stopPropagation(); api.load(true)" title="重试加载收藏">重试</button>}`,
+  styles: [`:host{display:inline-flex;align-items:center;position:relative;flex-shrink:0}button{display:inline-flex;align-items:center;justify-content:center;width:28px;height:28px;border:0;border-radius:7px;background:transparent;color:#64748b;cursor:pointer}button:hover{background:#fef3c7;color:#b45309}button.saved{color:#c47b0b}button.saved span{font-variation-settings:'FILL' 1}button:focus-visible{outline:2px solid #818cf8;outline-offset:2px}button:disabled{opacity:.45;cursor:default}.material-symbols-outlined{font-size:19px}.feedback{position:absolute;left:0;top:30px;min-width:170px;max-width:260px;padding:8px 10px;z-index:20;border:1px solid #fecaca;border-radius:8px;background:#fff7f7;color:#b91c1c;font-size:12px;line-height:1.5;box-shadow:0 4px 16px #0f172a12}.retry{font-size:10px;width:auto}`]
+})
+export class FavoriteButtonComponent {
+  api = inject(FavoritesService);
+  sourceId = input.required<string>();
+  sourceKind = input<FavoriteKind>('task');
+  version = input<number>();
+  label = input('收藏到精选任务');
+  busy = signal(false); error = signal('');
+  favorite = computed(() => this.api.items().find(i => i.source_kind === this.sourceKind() && i.source_id === this.sourceId() && (!this.version() || i.case_version === this.version())));
+  constructor() { void this.api.load(); }
+  async toggle(event: Event) {
+    event.stopPropagation(); if (this.busy()) return;
+    this.busy.set(true); this.error.set('');
+    try {
+      const saved = this.favorite();
+      if (saved) await this.api.remove(saved.id);
+      else await this.api.create(this.sourceKind(), this.sourceId(), this.version());
+    } catch(e) { this.error.set((e as any)?.error?.detail || '收藏操作失败，请重试'); }
+    finally { this.busy.set(false); }
+  }
+}

apps/showcase_ui/src/app/components/featured-tasks/featured-tasks.component.html

@@ -0,0 +1,38 @@
+<section class="featured" aria-label="精选任务">
+  <div class="heading"><div><h2><span class="material-symbols-outlined" aria-hidden="true">star</span>精选任务</h2><p>收藏常用流程，沿用原配置一键复跑。</p></div><input type="search" aria-label="搜索精选任务" placeholder="搜索收藏…" [ngModel]="query()" (ngModelChange)="query.set($event)"></div>
+  <div class="filters" aria-label="精选任务类型">@for(tab of filters; track tab.id){<button type="button" [class.selected]="filter() === tab.id" [attr.aria-pressed]="filter() === tab.id" (click)="filter.set(tab.id)">{{tab.label}}</button>}</div>
+  @if(error() || api.error()){<p class="error" role="alert">{{error() || api.error()}} @if(api.error()){<button (click)="api.load(true)">重试</button>}</p>}
+  @if(notice()){<p class="notice" role="status">{{notice()}}</p>}
+  @if(api.loading() && !api.loaded()){<p class="empty" role="status">正在读取精选任务…</p>}
+  @else if(!api.items().length && !api.error()){<div class="empty"><span class="material-symbols-outlined" aria-hidden="true">star_outline</span><strong>把常用测试留在这里</strong><p>在任务队列、历史记录或批次旁点击星标，即可收藏。删除原记录不会影响收藏。</p></div>}
+  @else if(!filtered().length){<p class="empty">没有匹配的精选任务。</p>}
+  <div class="cards">@for(item of filtered(); track item.id){
+    <article class="card" [class.expanded]="expanded() === item.id">
+      <div class="card-overview">
+      <div class="card-heading"><span class="mode" [class.pro]="options(item)['profile'] === 'pro'" [class.explore]="item.kind === 'exploration'">{{mode(item)}}</span>@if(item.case_version){<span class="case-version">用例 v{{item.case_version}}</span>}</div>
+      <div class="card-title"><button class="star" type="button" (click)="remove(item)" [disabled]="!!busy()" aria-label="取消收藏" title="取消收藏"><span class="material-symbols-outlined" aria-hidden="true">star</span></button><h3 [title]="item.name">{{item.name}}</h3></div><p class="model-summary">{{summary(item)}}</p>
+      <div class="card-actions"><button type="button" class="configure" (click)="toggle(item)" [attr.aria-expanded]="expanded() === item.id"><span class="material-symbols-outlined" aria-hidden="true">tune</span>配置<span class="material-symbols-outlined" aria-hidden="true">{{expanded() === item.id ? 'expand_less' : 'expand_more'}}</span></button><button type="button" class="primary" (click)="run(item)" [disabled]="!!busy()"><span class="material-symbols-outlined" aria-hidden="true">play_arrow</span>{{busy() === item.id ? '处理中…' : item.kind === 'exploration' ? '重新探索' : '运行'}}</button></div>
+      </div>
+      @if(expanded() === item.id){@let d = draft(item);
+        <div class="configuration"><label>收藏名称<input [(ngModel)]="d.name" maxlength="200" aria-label="收藏名称"></label>
+          @if(item.kind === 'batch'){
+            <label class="checkbox"><input type="checkbox" [(ngModel)]="d.uniform">本次统一模型和思考强度</label>
+            @if(d.uniform){<app-model-selector label="整批模型" [(value)]="d.model" /><app-setting-slider label="整批思考强度" [(value)]="d.reasoning_effort" [values]="efforts" [labels]="effortLabels" />}
+            @for(c of item.snapshot.cases; track $index; let i = $index){<details><summary>{{i + 1}}. {{c.name}}</summary>
+              <label>执行模式<select [(ngModel)]="d.cases[i].profile"><option value="flash">Flash</option><option value="pro">Pro</option></select></label>
+              <app-model-selector [label]="'用例 ' + (i + 1) + ' 模型'" [(value)]="d.cases[i].model" [disabled]="d.uniform" />
+              <app-setting-slider label="思考强度" [(value)]="d.cases[i].reasoning_effort" [values]="efforts" [labels]="effortLabels" [disabled]="d.uniform" />
+              @if(!computer(c.task_config)){<label>执行设备<select [(ngModel)]="d.cases[i].device_serial"><option value="">请选择在线设备</option>@if(d.cases[i].device_serial){<option [value]="d.cases[i].device_serial">原选择：{{d.cases[i].device_serial}}</option>}@for(device of system.connectedDevices(); track device.serial){<option [value]="device.serial">{{device.serial}}</option>}</select></label>}
+            </details>}
+          }@else{
+            @if(item.kind !== 'exploration'){<label>执行模式<select [(ngModel)]="d.profile"><option value="flash">Flash</option><option value="pro">Pro</option></select></label>}
+            <app-model-selector label="复跑模型" [(value)]="d.model" />
+            <app-setting-slider label="思考强度" [(value)]="d.reasoning_effort" [values]="efforts" [labels]="effortLabels" />
+            @if(!computer(options(item))){<label>执行设备<select [(ngModel)]="d.device_serial" aria-label="复跑设备"><option value="">请选择在线设备</option>@if(d.device_serial){<option [value]="d.device_serial">原选择：{{d.device_serial}}</option>}@for(device of system.connectedDevices(); track device.serial){<option [value]="device.serial">{{device.serial}}</option>}</select></label>}
+          }
+          <p class="hint">调整仅用于本次运行；保存后才会改变收藏的默认配置。</p><button class="save" (click)="save(item)" [disabled]="!!busy() || !d.name.trim() || (d.uniform && !d.model.trim())">保存为默认配置</button>
+        </div>
+      }
+    </article>
+  }</div>
+</section>

apps/showcase_ui/src/app/components/featured-tasks/featured-tasks.component.scss

@@ -0,0 +1,60 @@
+:host { display: block; width: 100%; color: #334155; }
+.featured { margin-top: 26px; }
+.heading { display: flex; align-items: center; justify-content: space-between; gap: 16px; }
+h2 { display: flex; align-items: center; gap: 7px; margin: 0; font-size: 16px; font-weight: 650; color: #1e293b; }
+h2 span { font-size: 20px; color: #d29a32; }
+.heading p { font-size: 12px; color: #64748b; margin: 6px 0 0; }
+.heading > input { max-width: 210px; width: 40%; border-radius: 20px; }
+.filters { display: flex; gap: 5px; flex-wrap: wrap; margin: 17px 0; }
+.filters button { border: 0; padding: 6px 13px; border-radius: 18px; font-size: 12px; background: transparent; color: #64748b; }
+.filters button.selected { background: #e0e7ff; color: #4338ca; font-weight: 600; }
+.cards { display: grid; grid-template-columns: repeat(auto-fill, minmax(min(100%, 260px), 1fr)); gap: 14px; align-items: start; }
+.card { padding: 16px; border: 1px solid #e2e8f0; border-radius: 16px; background: #fff; min-width: 0; transition: border-color .15s, box-shadow .15s; }
+.card:hover { border-color: #c7d2fe; box-shadow: 0 4px 16px #33415508; }
+.card.expanded { border-color: #a5b4fc; }
+.card-overview { display: flex; flex-direction: column; min-height: 178px; }
+.card-heading { display: flex; align-items: center; justify-content: space-between; gap: 6px; min-height: 22px; }
+.mode { padding: 3px 8px; background: #f1f5f9; color: #475569; border-radius: 6px; font-size: 11px; line-height: 16px; }
+.mode.pro { background: #ede9fe; color: #6d28d9; }
+.mode.explore { background: #e0f2fe; color: #0369a1; }
+.case-version { font-size: 11px; color: #94a3b8; white-space: nowrap; }
+.card-title { display: flex; align-items: flex-start; gap: 6px; margin: 13px 0 7px; min-height: 42px; }
+.card-title h3 { flex: 1; min-width: 0; margin: 0; font-size: 14px; line-height: 21px; overflow-wrap: anywhere; color: #1e293b;
+  display: -webkit-box; -webkit-box-orient: vertical; -webkit-line-clamp: 2; overflow: hidden; }
+.star { display: flex; flex-shrink: 0; width: 24px; height: 24px; align-items: center; justify-content: center; padding: 0; border: 0; background: transparent; color: #c58b23; }
+.star span { font-size: 19px; font-variation-settings: 'FILL' 1; }
+.model-summary { font-size: 12px; line-height: 18px; color: #64748b; margin: 0 0 12px; min-height: 36px; overflow-wrap: anywhere;
+  display: -webkit-box; -webkit-box-orient: vertical; -webkit-line-clamp: 2; overflow: hidden; }
+.card-actions { display: flex; align-items: center; justify-content: space-between; margin-top: auto; gap: 10px; }
+.card-actions button { display: flex; align-items: center; gap: 4px; min-height: 34px; padding: 7px 10px; border-radius: 8px; font-size: 12px; white-space: nowrap; }
+.card-actions span { font-size: 16px; }
+.configure { color: #64748b; border: 0; background: transparent; }
+.primary { border: 1px solid #6366f1; background: #6366f1; color: white; }
+.primary:hover { background: #4f46e5; }
+.configuration { border-top: 1px solid #e2e8f0; margin-top: 14px; padding-top: 14px; display: grid; gap: 12px; }
+.configuration label { display: grid; gap: 6px; font-size: 12px; color: #475569; }
+.configuration .checkbox { display: flex; align-items: center; }
+.checkbox input { width: auto; min-height: auto; }
+input, select { box-sizing: border-box; min-width: 0; width: 100%; min-height: 40px; padding: 10px 12px; border: 1px solid #dbe3ec; border-radius: 10px; color: #334155; background: white; font: inherit; font-size: 13px; }
+option { color: #334155; background: #fff; }
+.hint { font-size: 11px; color: #64748b; line-height: 1.5; margin: 7px 0 0; }
+.save { justify-self: start; padding: 7px 10px; border: 1px solid #cbd5e1; background: #fff; color: #475569; border-radius: 8px; font-size: 12px; }
+details { border: 1px solid #e2e8f0; padding: 10px; border-radius: 9px; }
+details summary { font-size: 12px; cursor: pointer; overflow-wrap: anywhere; }
+details[open] > label, details[open] > app-setting-slider { margin-top: 12px; display: block; }
+details[open] > app-model-selector { margin-top: 12px; }
+button { font-family: inherit; cursor: pointer; }
+button:disabled { opacity: .5; cursor: default; }
+button:focus-visible, input:focus-visible, select:focus-visible, summary:focus-visible { outline: 2px solid #818cf8; outline-offset: 2px; }
+.empty { display: flex; flex-direction: column; align-items: center; text-align: center; padding: 28px 18px; border: 1px dashed #cbd5e1; border-radius: 16px; color: #64748b; font-size: 12px; line-height: 1.7; }
+.empty > span { font-size: 28px; color: #c5a365; margin-bottom: 9px; }
+.empty strong { font-size: 14px; color: #475569; }
+.empty p { max-width: 440px; margin: 8px 0 0; }
+.error, .notice { padding: 10px 12px; border-radius: 8px; font-size: 12px; line-height: 1.6; }
+.error { background: #fff1f2; color: #b91c1c; }
+.notice { background: #f0fdf4; color: #166534; }
+@media (max-width: 560px) {
+  .heading { align-items: flex-start; flex-direction: column; }
+  .heading > input { width: 100%; max-width: none; }
+  .cards { grid-template-columns: 1fr; }
+}

apps/showcase_ui/src/app/components/featured-tasks/featured-tasks.component.spec.ts

@@ -0,0 +1,54 @@
+import { TestBed, fakeAsync, flushMicrotasks } from '@angular/core/testing';
+import { signal } from '@angular/core';
+import { Router } from '@angular/router';
+import { FeaturedTasksComponent } from './featured-tasks.component';
+import { FavoritesService, Favorite } from '../../services/favorites.service';
+import { AgentService } from '../../services/agent.service';
+import { SystemService } from '../../services/system.service';
+
+describe('Featured tasks',()=>{
+  let api:any, agent:any, system:any, router:any;
+  const favorite:Favorite={id:'favorite',name:'检查页面',kind:'task',source_kind:'task',source_id:'source',inherited_fields:[],snapshot:{goal:'检查主页',runtime_overrides:{profile:'pro',model:'original-model',reasoning_effort:'high',device_serial:'device'}}};
+  beforeEach(()=>{
+    api={items:signal([structuredClone(favorite)]),loaded:signal(true),loading:signal(false),error:signal(''),load:jasmine.createSpy().and.resolveTo(),run:jasmine.createSpy().and.resolveTo({kind:'task',session_id:'new'}),update:jasmine.createSpy().and.resolveTo(favorite),remove:jasmine.createSpy().and.resolveTo()};
+    agent={fetchSessions:jasmine.createSpy(),selectSession:jasmine.createSpy(),batchFilter:signal(''),activeTab:signal('tasks')};system={connectedDevices:signal([{serial:'device',state:'device'}])};router={navigate:jasmine.createSpy().and.resolveTo(true)};
+    TestBed.configureTestingModule({providers:[{provide:FavoritesService,useValue:api},{provide:AgentService,useValue:agent},{provide:SystemService,useValue:system},{provide:Router,useValue:router}]});
+  });
+  it('runs a card with original settings and keeps configuration folded initially',fakeAsync(()=>{
+    const fixture=TestBed.createComponent(FeaturedTasksComponent);fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.configuration')).toBeNull();
+    fixture.nativeElement.querySelector('.primary').click();flushMicrotasks();
+    expect(api.run).toHaveBeenCalledWith('favorite',{});expect(api.update).not.toHaveBeenCalled();expect(agent.selectSession).toHaveBeenCalledWith('new',true);expect(router.navigate).toHaveBeenCalledWith(['/workspace']);
+  }));
+  it('blocks unavailable original devices and opens a device chooser without submitting',fakeAsync(()=>{
+    system.connectedDevices.set([{serial:'other',state:'device'}]);const fixture=TestBed.createComponent(FeaturedTasksComponent);fixture.detectChanges();fixture.nativeElement.querySelector('.primary').click();flushMicrotasks();fixture.detectChanges();
+    expect(api.run).not.toHaveBeenCalled();expect(fixture.nativeElement.querySelector('[aria-label="复跑设备"]')).not.toBeNull();expect(fixture.nativeElement.querySelector('[role="alert"]').textContent).toContain('原设备不在线');
+  }));
+  it('applies model changes only to this run unless defaults are explicitly saved',fakeAsync(()=>{
+    const fixture=TestBed.createComponent(FeaturedTasksComponent);const c=fixture.componentInstance;c.draft(favorite).model='temporary';void c.run(favorite);flushMicrotasks();
+    expect(api.run).toHaveBeenCalledWith('favorite',{task_config:{model:'temporary'}});expect(api.update).not.toHaveBeenCalled();expect(favorite.snapshot.runtime_overrides!['model']).toBe('original-model');
+    void c.save(favorite);flushMicrotasks();expect(api.update).toHaveBeenCalledWith('favorite',{task_config:{model:'temporary'},name:'检查页面'});
+  }));
+  it('uses the shared model picker and automatically keeps completed legacy settings without a warning',fakeAsync(()=>{
+    api.items.set([{...structuredClone(favorite),inherited_fields:['record_video','recording']}]);
+    const fixture=TestBed.createComponent(FeaturedTasksComponent);fixture.detectChanges();
+    expect(fixture.nativeElement.textContent).not.toContain('旧任务部分配置缺失');
+    fixture.nativeElement.querySelector('.configure').click();fixture.detectChanges();flushMicrotasks();
+    const select:HTMLSelectElement=fixture.nativeElement.querySelector('app-model-selector select');
+    expect(Array.from(select.options,option=>option.value)).toEqual(['gpt-6-luna','gpt-6.1-sol','gpt-6-astra','mimo-v2.6-pro','__custom__']);
+    expect(select.value).toBe('__custom__');
+    expect(fixture.nativeElement.querySelector('input[aria-label="复跑模型（自定义）"]').value).toBe('original-model');
+    select.value='gpt-6-luna';select.dispatchEvent(new Event('change'));fixture.detectChanges();
+    fixture.nativeElement.querySelector('.primary').click();flushMicrotasks();
+    expect(api.run).toHaveBeenCalledWith('favorite',{task_config:{model:'gpt-6-luna'}});
+    expect(api.update).not.toHaveBeenCalled();
+  }));
+  it('computer-only favorites need no Android device and can be filtered by name',fakeAsync(()=>{
+    const pc:Favorite={...structuredClone(favorite),kind:'exploration_case',case_version:1,snapshot:{goal:'检查电脑',runtime_overrides:{profile:'flash',model:'original-model',reasoning_effort:'high',execution_target:'computer'}}};api.items.set([pc]);system.connectedDevices.set([]);
+    const fixture=TestBed.createComponent(FeaturedTasksComponent),c=fixture.componentInstance;c.query.set('不存在');expect(c.filtered()).toEqual([]);c.query.set('页面');expect(c.filtered().length).toBe(1);void c.run(pc);flushMicrotasks();expect(api.run).toHaveBeenCalled();
+  }));
+  it('batch defaults keep each case and repetition count, with explicit uniform override',()=>{
+    const batch:Favorite={...structuredClone(favorite),kind:'batch',snapshot:{repeat_count:3,cases:[{name:'甲',goal:'检查甲',task_config:{profile:'flash',model:'a',reasoning_effort:'high',device_serial:'device'}},{name:'乙',goal:'检查乙',task_config:{profile:'pro',model:'b',reasoning_effort:'max',device_serial:'device'}}]}};
+    const fixture=TestBed.createComponent(FeaturedTasksComponent),c=fixture.componentInstance;expect(c.payload(batch)).toEqual({});const d=c.draft(batch);expect(c.payload(batch)).toEqual({case_configs:{}});d.uniform=true;d.model='uniform';d.reasoning_effort='high';expect(c.payload(batch)['task_config']).toEqual({model:'uniform',reasoning_effort:'high'});expect(batch.snapshot.repeat_count).toBe(3);
+  });
+});

apps/showcase_ui/src/app/components/featured-tasks/featured-tasks.component.ts

@@ -0,0 +1,91 @@
+import { Component, ChangeDetectionStrategy, computed, inject, signal } from '@angular/core';
+import { FormsModule } from '@angular/forms';
+import { Router } from '@angular/router';
+import { FavoritesService, Favorite } from '../../services/favorites.service';
+import { AgentService } from '../../services/agent.service';
+import { SystemService } from '../../services/system.service';
+import { SettingSliderComponent, REASONING_VALUES, REASONING_LABELS } from '../setting-slider/setting-slider.component';
+import { ModelSelectorComponent } from '../model-selector/model-selector.component';
+
+interface ModelDraft {profile: string; model: string; reasoning_effort: string; device_serial: string;}
+interface Draft extends ModelDraft {name: string; uniform: boolean; cases: ModelDraft[];}
+
+@Component({
+  selector: 'app-featured-tasks', standalone: true, imports: [FormsModule, SettingSliderComponent, ModelSelectorComponent],
+  templateUrl: './featured-tasks.component.html', styleUrl: './featured-tasks.component.scss',
+  changeDetection: ChangeDetectionStrategy.OnPush
+})
+export class FeaturedTasksComponent {
+  api = inject(FavoritesService); system = inject(SystemService); private agent = inject(AgentService); private router = inject(Router);
+  query = signal(''); filter = signal('all'); expanded = signal<string | null>(null);
+  busy = signal<string | null>(null); error = signal(''); notice = signal('');
+  drafts = new Map<string, Draft>();
+  readonly efforts = REASONING_VALUES; readonly effortLabels = REASONING_LABELS;
+  readonly filters = [{id:'all',label:'全部'},{id:'flash',label:'Flash'},{id:'pro',label:'Pro'},{id:'exploration',label:'探索'},{id:'batch',label:'批量'}];
+  filtered = computed(() => this.api.items().filter(item => {
+    const filter = this.filter(), query = this.query().trim().toLowerCase();
+    return (filter === 'all' || (filter === 'batch' ? item.kind === 'batch' : filter === 'exploration' ? ['exploration','exploration_case'].includes(item.kind) : item.kind !== 'batch' && this.options(item)['profile'] === filter)) && (!query || `${item.name} ${item.snapshot.goal || ''}`.toLowerCase().includes(query));
+  }));
+  constructor() { void this.api.load(); }
+  options(item: Favorite): Record<string, any> { return item.kind === 'exploration' ? {...item.snapshot.task_config,...item.snapshot.config,profile:'exploration'} : item.snapshot.runtime_overrides || {}; }
+  mode(item: Favorite) { return item.kind === 'batch' ? '批量' : item.kind === 'exploration' ? '重新探索' : (this.options(item)['profile'] === 'pro' ? 'Pro' : 'Flash') + (item.kind === 'exploration_case' ? ' · 用例复跑' : ''); }
+  summary(item: Favorite) { return item.kind === 'batch' ? `${item.snapshot.cases?.length || 0} 个用例 · ${item.snapshot.repeat_count || 1} 轮 · 保留各项配置` : `${this.options(item)['model'] || '当前默认'} / ${this.options(item)['reasoning_effort'] || '当前默认'}`; }
+  computer(options: Record<string, any>) { return options['execution_target'] === 'computer'; }
+  modelDraft(options: Record<string, any>): ModelDraft { return {profile:options['profile'] || 'flash', model:options['model'] || '', reasoning_effort:options['reasoning_effort'] || 'default', device_serial:this.computer(options) ? '' : options['device_serial'] || ''}; }
+  draft(item: Favorite): Draft {
+    let draft = this.drafts.get(item.id);
+    if (!draft) { draft = {...this.modelDraft(item.kind === 'batch' ? item.snapshot.cases?.[0]?.task_config || {} : this.options(item)), name:item.name, uniform:false, cases:(item.snapshot.cases || []).map(c => this.modelDraft(c.task_config))}; this.drafts.set(item.id,draft); }
+    return draft;
+  }
+  toggle(item: Favorite) { this.expanded.set(this.expanded() === item.id ? null : item.id); }
+  private changes(current: ModelDraft, original: Record<string, any>) {
+    const baseline = this.modelDraft(original), result: Record<string, unknown> = {};
+    for (const key of ['profile','model','reasoning_effort','device_serial'] as const) {
+      if (current[key] !== baseline[key] && current[key]) result[key] = current[key];
+    }
+    return result;
+  }
+  payload(item: Favorite): Record<string, unknown> {
+    const draft = this.drafts.get(item.id); if (!draft) return {};
+    if (item.kind === 'batch') {
+      const case_configs: Record<number, unknown> = {};
+      draft.cases.forEach((c,i) => { const changes = this.changes(c,item.snapshot.cases![i].task_config); if(Object.keys(changes).length)case_configs[i]=changes; });
+      return {case_configs, ...(draft.uniform ? {task_config:{model:draft.model,reasoning_effort:draft.reasoning_effort}} : {})};
+    }
+    return {task_config:this.changes(draft,this.options(item))};
+  }
+  private deviceMissing(item: Favorite) {
+    const devices = this.system.connectedDevices();
+    const cases = item.kind === 'batch' ? item.snapshot.cases!.map((c,i) => ({...c.task_config,...(this.payload(item)['case_configs'] as any)?.[i]})) : [{...this.options(item),...(this.payload(item)['task_config'] as any)}];
+    return cases.some(options => !this.computer(options) && (!options['device_serial'] || !devices.some(d => d.serial === options['device_serial'] && d.state === 'device')));
+  }
+  async save(item: Favorite) {
+    if(!this.validModels(item)){this.error.set('模型名称不能为空');return;}
+    if(this.busy())return;this.busy.set(item.id);this.error.set('');this.notice.set('');
+    try { const saved = await this.api.update(item.id,{...this.payload(item),name:this.draft(item).name});this.drafts.delete(item.id);this.draft(saved);this.notice.set('已保存收藏的默认配置'); }
+    catch(e){this.error.set((e as any)?.error?.detail || '保存失败，请重试');}finally{this.busy.set(null);}
+  }
+  async remove(item: Favorite) {
+    if(this.busy())return;this.busy.set(item.id);this.error.set('');
+    try{await this.api.remove(item.id);this.drafts.delete(item.id);}catch(e){this.error.set((e as any)?.error?.detail || '取消收藏失败');}finally{this.busy.set(null);}
+  }
+  async run(item: Favorite) {
+    if(this.busy())return;this.error.set('');this.notice.set('');
+    if(!this.validModels(item)){this.error.set('模型名称不能为空');return;}
+    if(this.deviceMissing(item)){this.expanded.set(item.id);this.error.set('原设备不在线，请在配置中选择可用设备后运行。');return;}
+    const draft=this.drafts.get(item.id);if(draft?.uniform && !draft.model.trim()){this.error.set('请填写统一使用的模型');return;}
+    this.busy.set(item.id);
+    try{
+      const result=await this.api.run(item.id,this.payload(item));
+      this.agent.fetchSessions();
+      if(result.batch_id){this.agent.batchFilter.set(result.batch_id);this.agent.activeTab.set('tasks');}
+      else{const id=result.session_id || result.admission?.run?.id;if(id)this.agent.selectSession(id,true);}
+      await this.router.navigate(['/workspace']);
+    }catch(e){this.error.set((e as any)?.error?.detail || '运行请求失败；重试将沿用相同请求编号，避免重复执行。');}
+    finally{this.busy.set(null);}
+  }
+  private validModels(item: Favorite) {
+    const draft=this.drafts.get(item.id);if(!draft)return true;
+    return item.kind === 'batch' ? draft.uniform ? !!draft.model.trim() : draft.cases.every(c => !!c.model.trim()) : !!draft.model.trim();
+  }
+}

apps/showcase_ui/src/app/components/floating-video-player/floating-video-player.component.html

@@ -29,6 +29,7 @@ limitations under the License.
     [class.theater]="isTheaterMode()"
     [style.left.px]="!isTheaterMode() ? posX() : null"
     [style.top.px]="!isTheaterMode() ? posY() : null"
+    [style.--player-top]="posY() + 'px'"
     >
     <!-- Window Header (Draggable Handle) -->
     <header class="window-header" (mousedown)="startDrag($event)">
@@ -95,12 +96,31 @@ limitations under the License.
         </button>
       </div>
     </header>
+    @if (sources().length && !agentService.isVideoMinimized()) {
+      <div class="recording-sources">
+        <label for="recording-source">录像来源</label>
+        <select id="recording-source" [ngModel]="agentService.selectedRecordingSource()" (ngModelChange)="selectSource($event)">
+          @for (source of sources(); track source.id) {
+            <option [value]="source.id">{{ source.label }}{{ source.status === 'failed' ? '（录制异常）' : '' }}</option>
+          }
+        </select>
+      </div>
+      @if (agentService.recordingPlaybackMessage()) {
+        <div class="recording-notice" role="status">{{ agentService.recordingPlaybackMessage() }}</div>
+      }
+      @if (!agentService.activeVideoUrl()) {
+        <div class="recording-notice">此来源暂无可播放录像，请查看步骤截图或日志。</div>
+      }
+    }
+    @if (!agentService.isVideoMinimized() && agentService.playerMode() === 'video' && missingRecordingLabels().length) {
+      <div class="recording-notice" role="status">{{ missingRecordingLabels().join('、') }}步骤没有对应录像，可切换「步骤」查看已有截图。</div>
+    }
     <!-- Video Main Content Area -->
     @if (!agentService.isVideoMinimized()) {
       <div class="video-body">
         <!-- 1. LIVE SCREEN STREAM (AUTOMATICALLY ACTIVE DURING RUNNING TEST) -->
         @if (agentService.recordingPlaybackStatus() === 'live') {
-          @if (!liveStreamError()) {
+          @if (!liveStreamError() && pageVisible()) {
             <div class="live-stream-container">
               <img
                 class="live-screen-img"
@@ -332,12 +352,16 @@ limitations under the License.
                 @if (agentService.playerMode() === 'video' && agentService.recordingPlaybackStatus() === 'ready' && !agentService.isVideoLoading() && agentService.activeVideoUrl(); as videoUrl) {
                   @if (!videoLoadError()) {
                     <div class="video-container">
+                        @if (gapMessage()) { <div class="recording-gap" role="status">{{ gapMessage() }}</div> }
                       <video
                         #videoRef
+                          [style.visibility]="gapMessage() ? 'hidden' : 'visible'"
                         [src]="currentVideoUrl()"
                         playsinline
                         (loadedmetadata)="onLoadedMetadata()"
                         (timeupdate)="onTimeUpdate()"
+                        (seeked)="onSeeked()"
+                        (emptied)="onPause()"
                         (play)="onPlay()"
                         (pause)="onPause()"
                         (ended)="onEnded()"
@@ -349,11 +373,18 @@ limitations under the License.
                   <!-- Dedicated Bottom Controls Bar for Video -->
                   @if (!videoLoadError()) {
                     <div class="video-controls-footer">
+                      @if (hasRecordingGaps()) {
+                        <div class="recording-coverage-summary">
+                          <span>已录制 {{ formatTime(recordedDuration()) }}</span>
+                          <span>空白区间自动跳过</span>
+                        </div>
+                      }
                       <!-- Timeline Scrubber Bar -->
-                      <div class="scrubber-bar-container">
+                      <div class="scrubber-bar-container" [class.has-recording-gaps]="hasRecordingGaps()">
                         <input
                           type="range"
                           class="scrubber-range"
+                          aria-label="录像进度"
                           [min]="0"
                           [max]="duration() || 100"
                           [step]="0.1"
@@ -361,6 +392,30 @@ limitations under the License.
                           (input)="onScrub($event)"
                           />
                           <div class="scrubber-progress-fill" [style.width.%]="progressPercent()"></div>
+                          @if (hasRecordingGaps()) {
+                            <div class="recording-coverage" aria-hidden="true">
+                              @for (range of recordedCoverage(); track $index) {
+                                <span [style.left.%]="range.left" [style.width.%]="range.width"></span>
+                              }
+                              <i [style.left.%]="progressPercent()"></i>
+                            </div>
+                          }
+                          @if (videoStepMarkers().length > 0) {
+                            <div class="video-markers" aria-label="录像关键步骤">
+                              @for (marker of videoStepMarkers(); track marker.time) {
+                                <button
+                                  type="button"
+                                  class="video-marker"
+                                  [class.passed]="currentTime() >= marker.time"
+                                  [class.failed]="marker.failed"
+                                  [style.left.%]="marker.percent"
+                                  [title]="videoMarkerLabel(marker)"
+                                  [attr.aria-label]="'跳转到' + videoMarkerLabel(marker)"
+                                  (click)="seekToVideoMarker(marker)"
+                                ></button>
+                              }
+                            </div>
+                          }
                         </div>
                         <!-- Bottom Controls Row -->
                         <div class="controls-row">
@@ -375,7 +430,7 @@ limitations under the License.
                             </button>
                             <!-- Time Display -->
                             <span class="time-display">
-                              {{ formatTime(currentTime()) }} / {{ formatTime(duration()) }}
+                              <span [title]="agentService.usesSessionTimeline() ? '任务时间；蓝色区间有录像，空白区间未录制' : '录像时间'">{{ formatTime(currentTime()) }} / {{ formatTime(duration()) }}</span>
                             </span>
                           </div>
                           <div class="controls-right">
@@ -406,6 +461,9 @@ limitations under the License.
                             </button>
                           </div>
                         </div>
+                        @if (skippedInterval()) {
+                          <div class="recording-skip-notice" role="status">{{ skippedInterval() }}</div>
+                        }
                       </div>
                     }
                   }

apps/showcase_ui/src/app/components/floating-video-player/floating-video-player.component.scss

@@ -31,6 +31,7 @@
   position: fixed;
   width: 400px;
   max-width: calc(100vw - 32px);
+  max-height: calc(100dvh - var(--player-top, 60px) - 12px);
   background: rgba(255, 255, 255, 0.84);
   backdrop-filter: blur(24px) saturate(180%);
   -webkit-backdrop-filter: blur(24px) saturate(180%);
@@ -298,6 +299,7 @@
 
   /* --- Video Body Area --- */
   .video-body {
+    min-height: 0;
     position: relative;
     background: #ffffff;
     display: flex;
@@ -655,6 +657,7 @@
     }
 
     .video-container {
+      min-height: 0;
       position: relative;
       width: 100%;
       display: flex;
@@ -664,6 +667,7 @@
       background: #ffffff;
 
       video {
+        min-height: 0;
         width: 100%;
         height: auto;
         max-height: 520px;
@@ -724,6 +728,38 @@
           box-shadow: 0 0 8px rgba(2, 132, 199, 0.4);
         }
 
+        .video-markers {
+          position: absolute;
+          inset: 0;
+          z-index: 6;
+          pointer-events: none;
+
+          .video-marker {
+            position: absolute;
+            top: 50%;
+            width: 12px;
+            height: 12px;
+            padding: 0;
+            border: 2px solid #fff;
+            border-radius: 50%;
+            background: #64748b;
+            box-shadow: 0 1px 4px rgba(15, 23, 42, 0.25);
+            transform: translate(-50%, -50%);
+            cursor: pointer;
+            pointer-events: auto;
+            transition: transform 0.15s ease, background 0.15s ease;
+
+            &.passed { background: #0284c7; }
+            &.failed { background: #ef4444; }
+            &:hover, &:focus-visible {
+              z-index: 1;
+              transform: translate(-50%, -50%) scale(1.35);
+              outline: 2px solid #38bdf8;
+              outline-offset: 2px;
+            }
+          }
+        }
+
         .step-ticks-wrapper {
           position: absolute;
           left: 0;
@@ -985,3 +1021,86 @@
     }
   }
 }
+.recording-sources {
+  display: flex;
+  gap: 8px;
+  align-items: center;
+  padding: 8px 12px;
+  background: #f8fafc;
+  border-bottom: 1px solid #e2e8f0;
+  font-size: 12px;
+  color: #475569;
+
+  select {
+    min-width: 0;
+    flex: 1;
+    border: 1px solid #cbd5e1;
+    border-radius: 6px;
+    padding: 4px;
+    background: white;
+  }
+}
+
+.recording-notice {
+  padding: 8px 12px;
+  background: #fffbeb;
+  color: #92400e;
+  font-size: 12px;
+}
+
+.recording-gap {
+  position: absolute;
+  inset: 0;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  padding: 24px;
+  color: #475569;
+  background: #f1f5f9;
+  text-align: center;
+  z-index: 2;
+}
+
+.recording-coverage-summary {
+  display: flex;
+  justify-content: space-between;
+  flex-wrap: wrap;
+  gap: 4px 12px;
+  color: #64748b;
+  font-size: 11px;
+  line-height: 16px;
+}
+
+.recording-skip-notice {
+  color: #64748b;
+  font-size: 11px;
+  line-height: 16px;
+}
+
+.has-recording-gaps {
+  .scrubber-progress-fill { opacity: 0; }
+
+  .recording-coverage {
+    position: absolute;
+    inset: 0;
+    pointer-events: none;
+    overflow: hidden;
+
+    span {
+      position: absolute;
+      top: 2px;
+      height: 5px;
+      border-radius: 3px;
+      background: #0284c7;
+    }
+
+    i {
+      position: absolute;
+      top: 0;
+      width: 3px;
+      height: 8px;
+      background: #0f172a;
+      transform: translateX(-50%);
+    }
+  }
+}

apps/showcase_ui/src/app/components/floating-video-player/floating-video-player.component.spec.ts

@@ -0,0 +1,507 @@
+import { ComponentFixture, TestBed } from '@angular/core/testing';
+import { signal } from '@angular/core';
+import { AgentService, RecordingSource, VideoSegment } from '../../services/agent.service';
+import { StepReplayFrame } from '../../core/models/stream.model';
+import { FloatingVideoPlayerComponent } from './floating-video-player.component';
+
+describe('FloatingVideoPlayerComponent video markers', () => {
+  let fixture: ComponentFixture<FloatingVideoPlayerComponent>;
+  let component: FloatingVideoPlayerComponent;
+  let video: HTMLVideoElement;
+  let paused: boolean;
+  let seeking: boolean;
+  let time: number;
+  let play: jasmine.Spy;
+  let agentService: ReturnType<typeof createAgentService>;
+
+  function createAgentService() {
+    const start = 1_800_000_000;
+    const frame: StepReplayFrame = {
+      index: 0, stepNumber: 1, title: '步骤 1: 点击设置', imageUrl: 'step.png',
+      actionText: '点击设置', timestamp: start + 5
+    };
+    const service = {
+      isVideoWindowOpen: signal(true),
+      isVideoMinimized: signal(false),
+      recordingPlaybackStatus: signal('ready'),
+      playerMode: signal('video'),
+      hasCurrentSessionStepFrames: signal(true),
+      activeVideoTitle: signal('测试录像'),
+      isVideoLoading: signal(false),
+      activeVideoUrl: signal('/test.mp4'),
+      activeVideoSegments: signal<VideoSegment[]>([{ url: '/test.mp4', start: 0, duration: 20, width: 400, height: 800 }]),
+      recordingSources: signal<RecordingSource[]>([]),
+      selectedRecordingSource: signal(''),
+      usesSessionTimeline: signal(false),
+      recordingSessionDuration: signal(0),
+      recordingPlaybackMessage: signal(''),
+      selectRecordingSource: jasmine.createSpy('selectRecordingSource'),
+      currentSessionStepFrames: signal([frame]),
+      currentSession: signal({ session_id: 'test', start_time: start }),
+      currentSessionId: signal('test'),
+      videoSeekRequest: signal<{ seconds: number; requestId: number; sourceId?: string } | null>(null),
+      stepSeekRequest: signal(null),
+      consumeVideoAutoplay: jasmine.createSpy('consumeVideoAutoplay').and.returnValue(false)
+    };
+    service.selectRecordingSource.and.callFake((id: string) => {
+      const source = service.recordingSources().find(item => item.id === id);
+      if (!source) return false;
+      if (service.selectedRecordingSource() === id) return true;
+      service.selectedRecordingSource.set(id);
+      service.activeVideoSegments.set(source.segments);
+      service.activeVideoUrl.set(source.url || source.segments[0]?.url || '');
+      return true;
+    });
+    return service;
+  }
+
+  beforeEach(() => {
+    agentService = createAgentService();
+    TestBed.configureTestingModule({
+      imports: [FloatingVideoPlayerComponent],
+      providers: [{ provide: AgentService, useValue: agentService }]
+    });
+
+    fixture = TestBed.createComponent(FloatingVideoPlayerComponent);
+    fixture.detectChanges();
+    component = fixture.componentInstance;
+    video = component.videoRef!.nativeElement;
+    paused = true;
+    seeking = false;
+    time = 0;
+    Object.defineProperties(video, {
+      paused: { configurable: true, get: () => paused },
+      seeking: { configurable: true, get: () => seeking },
+      ended: { configurable: true, get: () => false },
+      duration: { configurable: true, get: () => 20 },
+      readyState: { configurable: true, get: () => 1 },
+      currentTime: {
+        configurable: true, get: () => time,
+        set: (value: number) => { time = value; seeking = true; }
+      }
+    });
+    spyOn(video, 'pause').and.callFake(() => {
+      paused = true;
+      video.dispatchEvent(new Event('pause'));
+    });
+    play = spyOn(video, 'play').and.callFake(() => {
+      paused = false;
+      video.dispatchEvent(new Event('play'));
+      return Promise.resolve();
+    });
+    video.dispatchEvent(new Event('loadedmetadata'));
+  });
+
+  function finishSeek(): void {
+    seeking = false;
+    video.dispatchEvent(new Event('seeked'));
+    fixture.detectChanges();
+  }
+
+  it('removes the live image while hidden and reconnects it on return', () => {
+    const hidden = spyOnProperty(document, 'hidden', 'get').and.returnValue(false);
+    agentService.recordingPlaybackStatus.set('live');
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('img.live-screen-img')).not.toBeNull();
+
+    hidden.and.returnValue(true);
+    document.dispatchEvent(new Event('visibilitychange'));
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('img.live-screen-img')).toBeNull();
+
+    hidden.and.returnValue(false);
+    document.dispatchEvent(new Event('visibilitychange'));
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('img.live-screen-img')).not.toBeNull();
+    expect(component.liveStreamError()).toBeFalse();
+  });
+
+  function playButton(): HTMLButtonElement {
+    return fixture.nativeElement.querySelector('.play-btn');
+  }
+
+  it('shows a recorded step on the video timeline and seeks when clicked', () => {
+    const marker = fixture.nativeElement.querySelector('.video-marker') as HTMLButtonElement;
+    expect(marker).not.toBeNull();
+    expect(marker.getAttribute('aria-label')).toContain('步骤 1');
+    expect(marker.style.left).toBe('25%');
+
+    const seek = spyOn(fixture.componentInstance, 'seek');
+    marker.click();
+    expect(seek).toHaveBeenCalledOnceWith(5);
+  });
+
+  it('resumes a playing video after a marker seek finishes and synchronizes the button', async () => {
+    component.togglePlay();
+    await Promise.resolve();
+    play.calls.reset();
+    fixture.nativeElement.querySelector('.video-marker').click();
+    expect(time).toBe(5);
+    expect(play).not.toHaveBeenCalled();
+    finishSeek();
+    await Promise.resolve();
+    fixture.detectChanges();
+    expect(paused).toBeFalse();
+    expect(play).toHaveBeenCalledTimes(1);
+    expect(playButton().title).toBe('暂停');
+  });
+
+  it('keeps a paused video paused after a marker seek and resumes with one click', async () => {
+    fixture.nativeElement.querySelector('.video-marker').click();
+    finishSeek();
+    expect(paused).toBeTrue();
+    expect(play).not.toHaveBeenCalled();
+    expect(playButton().title).toBe('播放');
+    playButton().click();
+    await Promise.resolve();
+    fixture.detectChanges();
+    expect(paused).toBeFalse();
+    expect(playButton().title).toBe('暂停');
+  });
+
+  it('preserves playing intent across rapid consecutive marker seeks', () => {
+    component.togglePlay();
+    component.seek(5);
+    component.seek(8);
+    play.calls.reset();
+    finishSeek();
+    expect(time).toBe(8);
+    expect(play).toHaveBeenCalledTimes(1);
+    expect(paused).toBeFalse();
+  });
+
+  it('honors play and pause clicks while the destination is still seeking', () => {
+    component.togglePlay();
+    component.seek(5);
+    play.calls.reset();
+    fixture.detectChanges();
+    expect(playButton().title).toBe('播放');
+    playButton().click();
+    fixture.detectChanges();
+    expect(playButton().title).toBe('暂停');
+    expect(play).not.toHaveBeenCalled();
+    playButton().click();
+    finishSeek();
+    expect(paused).toBeTrue();
+    expect(play).not.toHaveBeenCalled();
+    expect(playButton().title).toBe('播放');
+  });
+
+  it('waits for the destination segment to seek before resuming playback', () => {
+    agentService.activeVideoSegments.set([
+      { url: '/test.mp4', start: 0, duration: 20, width: 400, height: 800 },
+      { url: '/second.mp4', start: 20, duration: 20, width: 400, height: 800 }
+    ]);
+    fixture.detectChanges();
+    component.togglePlay();
+    component.seek(25);
+    play.calls.reset();
+    fixture.detectChanges();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    expect(component.activeSegmentIndex()).toBe(1);
+    expect(time).toBe(5);
+    expect(play).not.toHaveBeenCalled();
+    finishSeek();
+    expect(play).toHaveBeenCalledTimes(1);
+    expect(paused).toBeFalse();
+    expect(component.currentTime()).toBe(25);
+  });
+
+  it('keeps a paused video paused when jumping to another segment', () => {
+    agentService.activeVideoSegments.set([
+      { url: '/test.mp4', start: 0, duration: 20, width: 400, height: 800 },
+      { url: '/second.mp4', start: 20, duration: 20, width: 400, height: 800 }
+    ]);
+    fixture.detectChanges();
+    component.seek(25);
+    fixture.detectChanges();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    finishSeek();
+    expect(time).toBe(5);
+    expect(paused).toBeTrue();
+    expect(play).not.toHaveBeenCalled();
+    expect(playButton().title).toBe('播放');
+  });
+
+  it('resumes after seeking a recording without a segment manifest', () => {
+    agentService.activeVideoSegments.set([]);
+    agentService.usesSessionTimeline.set(true);
+    fixture.detectChanges();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    component.togglePlay();
+    component.seek(5);
+    play.calls.reset();
+    finishSeek();
+    expect(time).toBe(5);
+    expect(component.currentTime()).toBe(5);
+    expect(paused).toBeFalse();
+    expect(play).toHaveBeenCalledTimes(1);
+  });
+
+  function setSources(computerStart = 0): void {
+    agentService.recordingSources.set([
+      { id: 'android', label: '安卓', kind: 'android', target_ids: ['android'], status: 'ready',
+        segments: [{ url: '/test.mp4', start: 0, duration: 20, width: 400, height: 800 }] },
+      { id: 'screen', label: '电脑', kind: 'computer', target_ids: ['window-1'], status: 'ready',
+        segments: [{ url: '/computer.mp4', start: computerStart, duration: 20, width: 1280, height: 720 }] }
+    ]);
+    agentService.usesSessionTimeline.set(true);
+    agentService.recordingSessionDuration.set(40);
+    agentService.selectRecordingSource('android');
+    fixture.detectChanges();
+  }
+
+  function setGappedRecording(): void {
+    agentService.usesSessionTimeline.set(true);
+    agentService.recordingSessionDuration.set(1040);
+    agentService.activeVideoSegments.set([
+      { url: '/test.mp4', start: 2, offset_ms: 2000, duration: 20, width: 400, height: 800 },
+      { url: '/resume.mp4', start: 1000, offset_ms: 1000000, duration: 20, width: 400, height: 800 }
+    ]);
+    fixture.detectChanges();
+    video.dispatchEvent(new Event('loadedmetadata'));
+  }
+
+  it('automatically skips a long pause and plays the next clip after it loads', () => {
+    setGappedRecording();
+    component.togglePlay();
+    play.calls.reset();
+    time = 20;
+    paused = true;
+    component.onEnded();
+    expect(component.activeSegmentIndex()).toBe(1);
+    expect(component.currentTime()).toBe(1000);
+    expect(component.gapMessage()).toBe('');
+    expect(component.skippedInterval()).toContain('00:22–16:40');
+    expect(play).not.toHaveBeenCalled();
+    component.onTimeUpdate(); // Late timeupdate from the ended clip must not move the new timeline.
+    expect(component.currentTime()).toBe(1000);
+    fixture.detectChanges();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    finishSeek();
+    expect(time).toBe(0);
+    expect(paused).toBeFalse();
+    expect(play).toHaveBeenCalledTimes(1);
+  });
+
+  it('restarts and loops at the first recorded frame, skipping leading and trailing gaps', () => {
+    setGappedRecording();
+    component.restart();
+    finishSeek();
+    expect(component.currentTime()).toBe(2);
+    expect(paused).toBeFalse();
+    component.seek(1010, true);
+    fixture.detectChanges();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    finishSeek();
+    component.isLooping.set(true);
+    paused = true;
+    component.onEnded();
+    fixture.detectChanges();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    finishSeek();
+    expect(component.activeSegmentIndex()).toBe(0);
+    expect(component.currentTime()).toBe(2);
+    expect(component.gapMessage()).toBe('');
+    expect(paused).toBeFalse();
+  });
+
+  for (const playing of [false, true]) {
+    it(`snaps a scrub in a gap forward while preserving ${playing ? 'playing' : 'paused'} intent`, () => {
+      setGappedRecording();
+      if (playing) component.togglePlay();
+      const input = fixture.nativeElement.querySelector('.scrubber-range') as HTMLInputElement;
+      input.value = '500';
+      input.dispatchEvent(new Event('input'));
+      fixture.detectChanges();
+      video.dispatchEvent(new Event('loadedmetadata'));
+      finishSeek();
+      expect(component.currentTime()).toBe(1000);
+      expect(component.skippedInterval()).toContain('08:20–16:40');
+      expect(paused).toBe(!playing);
+    });
+  }
+
+  it('does not substitute another time for an exact step seek in a gap', () => {
+    setGappedRecording();
+    component.seekToSessionTime(500, true);
+    expect(component.currentTime()).toBe(500);
+    expect(component.gapMessage()).toContain('未录制');
+    expect(paused).toBeTrue();
+    expect(play).not.toHaveBeenCalled();
+  });
+
+  it('ends at the last recorded frame without playing the trailing gap', () => {
+    setGappedRecording();
+    component.seek(1010, true);
+    fixture.detectChanges();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    finishSeek();
+    play.calls.reset();
+    paused = true;
+    component.onEnded();
+    expect(component.currentTime()).toBe(1020);
+    expect(component.isPlaying()).toBeFalse();
+    expect(component.skippedInterval()).toContain('录像已结束');
+    expect(play).not.toHaveBeenCalled();
+  });
+
+  it('shows actual recorded duration and coverage separately from task duration', () => {
+    setGappedRecording();
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.recording-coverage-summary').textContent).toContain('00:40');
+    expect(fixture.nativeElement.querySelector('.time-display').textContent).toContain('17:20');
+    expect(fixture.nativeElement.querySelectorAll('.recording-coverage span').length).toBe(2);
+    expect(component.recordedCoverage()[1].left).toBeCloseTo(1000 / 1040 * 100);
+  });
+
+  it('explains missing computer footage without treating Android video as its evidence', () => {
+    agentService.currentSessionStepFrames.update(frames => [...frames,
+      { ...frames[0], sourceId: 'plugin:open-computer-use', sourceLabel: '电脑', index: 1 }
+    ]);
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.recording-notice').textContent).toContain('电脑步骤没有对应录像');
+    agentService.recordingSources.set([
+      { id: 'pc', label: '电脑', kind: 'computer', target_ids: ['plugin:open-computer-use'],
+        status: 'ready', segments: agentService.activeVideoSegments() }
+    ]);
+    fixture.detectChanges();
+    expect(component.missingRecordingLabels()).toEqual([]);
+  });
+
+  it('handles a seek request once without seeking back on play or pause', () => {
+    const seek = spyOn(component, 'seekToSessionTime').and.callThrough();
+    agentService.videoSeekRequest.set({ seconds: 5, requestId: 1 });
+    fixture.detectChanges();
+    finishSeek();
+    time = 9;
+    component.togglePlay();
+    fixture.detectChanges();
+    expect(time).toBe(9);
+    component.togglePlay();
+    fixture.detectChanges();
+    expect(time).toBe(9);
+    expect(seek).toHaveBeenCalledOnceWith(5, false);
+  });
+
+  it('defers a seek command until the recording is ready', () => {
+    const seek = spyOn(component, 'seekToSessionTime');
+    agentService.recordingPlaybackStatus.set('processing');
+    agentService.videoSeekRequest.set({ seconds: 5, requestId: 2 });
+    fixture.detectChanges();
+    expect(seek).not.toHaveBeenCalled();
+    agentService.recordingPlaybackStatus.set('ready');
+    fixture.detectChanges();
+    component.onLoadedMetadata();
+    expect(seek).toHaveBeenCalledOnceWith(5, false);
+  });
+
+  it('switches to the requested source, waits for seeked and allows later manual source changes', () => {
+    setSources(2);
+    component.togglePlay();
+    agentService.videoSeekRequest.set({ seconds: 7, requestId: 3, sourceId: 'window-1' });
+    fixture.detectChanges();
+    play.calls.reset();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    expect(time).toBe(5);
+    expect(play).not.toHaveBeenCalled();
+    finishSeek();
+    expect(component.currentTime()).toBe(7);
+    expect(paused).toBeFalse();
+    expect(play).toHaveBeenCalledTimes(1);
+    component.selectSource('android');
+    fixture.detectChanges();
+    expect(agentService.selectedRecordingSource()).toBe('android');
+    video.dispatchEvent(new Event('loadedmetadata'));
+    finishSeek();
+    expect(time).toBe(7);
+    expect(agentService.selectedRecordingSource()).toBe('android');
+  });
+
+  it('preserves task time when switching into a gap and advances only on explicit play', () => {
+    setSources(20);
+    time = 12;
+    component.onTimeUpdate();
+    component.togglePlay();
+    component.selectSource('screen');
+    fixture.detectChanges();
+    play.calls.reset();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    expect(component.currentTime()).toBe(12);
+    expect(component.gapMessage()).toContain('未录制');
+    expect(paused).toBeTrue();
+    expect(play).not.toHaveBeenCalled();
+    component.togglePlay();
+    finishSeek();
+    expect(component.currentTime()).toBe(20);
+    expect(component.gapMessage()).toBe('');
+    expect(paused).toBeFalse();
+  });
+
+  it('retains a paused source switch and resumes with one play click', () => {
+    setSources();
+    time = 6;
+    component.onTimeUpdate();
+    component.selectSource('screen');
+    fixture.detectChanges();
+    video.dispatchEvent(new Event('loadedmetadata'));
+    finishSeek();
+    expect(component.currentTime()).toBe(6);
+    expect(paused).toBeTrue();
+    expect(play).not.toHaveBeenCalled();
+    component.togglePlay();
+    expect(play).toHaveBeenCalledTimes(1);
+    expect(paused).toBeFalse();
+  });
+
+  it('does not play another source when the requested source has no recording', () => {
+    setSources();
+    agentService.videoSeekRequest.set({ seconds: 5, requestId: 4, sourceId: 'missing-window' });
+    fixture.detectChanges();
+    component.togglePlay();
+    expect(component.gapMessage()).toContain('来源没有录像');
+    expect(component.currentTime()).toBe(5);
+    expect(paused).toBeTrue();
+    expect(play).not.toHaveBeenCalled();
+  });
+
+  it('resets a rejected play request so a single subsequent click can resume', async () => {
+    play.and.callFake(() => {
+      paused = false;
+      video.dispatchEvent(new Event('play'));
+      return Promise.reject(new DOMException('Interrupted', 'AbortError'));
+    });
+    component.seek(5, true);
+    finishSeek();
+    await Promise.resolve();
+    await Promise.resolve();
+    fixture.detectChanges();
+    expect(paused).toBeTrue();
+    expect(playButton().title).toBe('播放');
+    play.and.callFake(() => {
+      paused = false;
+      video.dispatchEvent(new Event('play'));
+      return Promise.resolve();
+    });
+    playButton().click();
+    await Promise.resolve();
+    fixture.detectChanges();
+    expect(paused).toBeFalse();
+    expect(playButton().title).toBe('暂停');
+  });
+
+  it('ignores an older aborted play request after a newer seek resumes', async () => {
+    let rejectOldPlay!: (reason: unknown) => void;
+    play.and.returnValue(new Promise<void>((_, reject) => { rejectOldPlay = reject; }));
+    component.togglePlay();
+    paused = false;
+    component.seek(5);
+    play.and.callFake(() => { paused = false; return Promise.resolve(); });
+    finishSeek();
+    rejectOldPlay(new DOMException('Interrupted', 'AbortError'));
+    await Promise.resolve();
+    await Promise.resolve();
+    expect(paused).toBeFalse();
+    expect(component.isPlaying()).toBeTrue();
+  });
+});

apps/showcase_ui/src/app/components/floating-video-player/floating-video-player.component.ts

@@ -25,6 +25,7 @@ import {
   ViewChild,
   HostListener,
   effect,
+  untracked,
   OnDestroy
 } from '@angular/core';
 
@@ -33,7 +34,8 @@ import { AgentService } from '../../services/agent.service';
 import { StepReplayFrame } from '../../core/models/stream.model';
 import { drawActionCoordinatesOnOverlay } from '../../utils/image-overlay.util';
 import { getActionIcon } from '../../utils/action-formatter.util';
-import { locateTimelineTime, sessionTimeToTimelineTime } from '../../utils/recording-timeline.util';
+import { locateCoveredSessionTime, locateTimelineTime, sessionTimeToTimelineTime } from '../../utils/recording-timeline.util';
+import { buildVideoStepMarkers, VideoStepMarker } from '../../utils/video-step-markers.util';
 
 @Component({
   selector: 'app-floating-video-player',
@@ -60,8 +62,29 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
   public isMuted = signal<boolean>(false);
   public isTheaterMode = signal<boolean>(false);
   public videoLoadError = signal<boolean>(false);
+  public gapMessage = signal('');
+  public skippedInterval = signal('');
+  private sourceSwitch: { time: number; play: boolean } | null = null;
+  private lastHandledVideoSeekRequestId: number | null = null;
+  private gapCanAdvance = true;
+  public sources = computed(() => this.agentService.recordingSources?.() || []);
   public liveStreamUrl = signal<string>('/api/stream/device-live');
   public liveStreamError = signal<boolean>(false);
+  public pageVisible = signal(!document.hidden);
+
+  @HostListener('document:visibilitychange')
+  @HostListener('window:pageshow')
+  onPageVisibilityChange(): void {
+    this.pageVisible.set(!document.hidden);
+    if (!document.hidden && this.agentService.recordingPlaybackStatus() === 'live') {
+      this.retryLiveStream();
+    }
+  }
+
+  @HostListener('window:pagehide')
+  onPageHide(): void {
+    this.pageVisible.set(false);
+  }
   public activeSegmentIndex = signal<number>(0);
   public currentVideoUrl = computed(() => {
     const segments = this.agentService.activeVideoSegments();
@@ -70,11 +93,14 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
   private pendingLocalTime: number | null = null;
   private pendingAutoplay = false;
   private pendingAbsoluteSeek: number | null = null;
+  private waitingForMetadata = true;
+  private seekPlaybackIntent: boolean | null = null;
+  private playRequestId = 0;
 
   // Step Replay state signals
   public activeStepIndex = signal<number>(0);
   public isStepPlaying = signal<boolean>(false);
-  public viewMode = signal<'pre' | 'post'>('pre'); // 'pre' = action preview with coords, 'post' = action result
+  public viewMode = signal<'pre' | 'post'>('post'); // default to the observed result
   public isCardHovered = signal<boolean>(false);
   public isCardPinned = signal<boolean>(false);
   public isCardVisible = computed(() => this.isCardHovered() || this.isCardPinned());
@@ -83,6 +109,33 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
   private stepTimer: any = null;
 
   public stepFrames = computed(() => this.agentService.currentSessionStepFrames());
+  public missingRecordingLabels = computed(() => {
+    const sources = this.sources();
+    return [...new Set(this.stepFrames()
+      .filter(frame => frame.sourceId && frame.sourceId !== 'android' && !sources.some(source =>
+        (source.id === frame.sourceId || source.target_ids.includes(frame.sourceId!)) &&
+        (source.segments.length > 0 || source.url)))
+      .map(frame => frame.sourceLabel || '插件'))];
+  });
+  public recordedDuration = computed(() => this.agentService.activeVideoSegments()
+    .reduce((sum, segment) => sum + segment.duration, 0));
+  public hasRecordingGaps = computed(() => !!this.agentService.usesSessionTimeline?.() &&
+    this.agentService.activeVideoSegments().length > 0 && this.duration() - this.recordedDuration() > 0.1);
+  public recordedCoverage = computed(() => this.agentService.activeVideoSegments().map(segment => ({
+    start: segment.start,
+    left: this.duration() > 0 ? segment.start / this.duration() * 100 : 0,
+    width: this.duration() > 0 ? segment.duration / this.duration() * 100 : 0
+  })));
+  public videoStepMarkers = computed(() => buildVideoStepMarkers(
+    this.stepFrames().filter(frame => {
+      const source = this.sources().find(s => s.id === this.agentService.selectedRecordingSource?.());
+      return !source || source.target_ids.includes(frame.sourceId || 'android');
+    }),
+    this.agentService.currentSession()?.start_time,
+    this.agentService.activeVideoSegments(),
+    this.duration(),
+    this.agentService.usesSessionTimeline?.() || false
+  ));
   public totalStepFrames = computed(() => this.stepFrames().length);
   public currentStepFrame = computed<StepReplayFrame | null>(() => {
     const frames = this.stepFrames();
@@ -127,10 +180,26 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
         const url = this.agentService.activeVideoUrl();
         const segments = this.agentService.activeVideoSegments();
         this.videoLoadError.set(false);
+        this.waitingForMetadata = true;
         this.activeSegmentIndex.set(0);
-        this.currentTime.set(0);
+        const switching = this.sourceSwitch;
-        this.duration.set(segments.reduce((sum, segment) => sum + segment.duration, 0));
+        this.sourceSwitch = null;
+        this.currentTime.set(switching?.time || 0);
+        this.gapMessage.set('');
+        this.skippedInterval.set('');
+        this.gapCanAdvance = true;
+        this.duration.set(this.agentService.usesSessionTimeline?.()
+          ? Math.max(this.agentService.recordingSessionDuration?.() || 0, ...segments.map(s => s.start + s.duration))
+          : segments.reduce((sum, segment) => sum + segment.duration, 0));
         this.isPlaying.set(false);
+        this.pendingLocalTime = null;
+        this.pendingAutoplay = false;
+        if (switching) {
+          this.pendingAbsoluteSeek = switching.time;
+          this.pendingAutoplay = switching.play;
+        }
+        this.seekPlaybackIntent = null;
+        this.playRequestId++;
       },
       { allowSignalWrites: true }
     );
@@ -144,7 +213,7 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
         if (curSessionId !== lastTrackedSessionId) {
           lastTrackedSessionId = curSessionId;
           this.activeStepIndex.set(0);
-          this.viewMode.set('pre');
+          this.viewMode.set('post');
           this.pauseStepPlay();
         } else if (total > 0 && this.activeStepIndex() >= total) {
           this.activeStepIndex.set(total - 1);
@@ -173,15 +242,42 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
 
     effect(() => {
       const request = this.agentService.videoSeekRequest();
-      if (!request) return;
+      const status = this.agentService.recordingPlaybackStatus();
+      const sources = this.sources();
+      if (!request || status !== 'ready' || request.requestId === this.lastHandledVideoSeekRequestId) return;
+      // A seek is a command, not a binding to playback/source state. Reading
+      // those signals in this effect would replay the command after play/pause
+      // or a manual source selection.
+      untracked(() => {
+        this.lastHandledVideoSeekRequestId = request.requestId;
         this.pendingAbsoluteSeek = request.seconds;
-      if (this.agentService.recordingPlaybackStatus() !== 'ready') return;
+        if (request.sourceId && sources.length) {
-      const video = this.videoRef?.nativeElement;
+          const source = sources.find(s => s.id === request.sourceId || s.target_ids.includes(request.sourceId!));
-      if (video && video.readyState >= 1) {
+          if (!source) {
-        const target = this.pendingAbsoluteSeek;
             this.pendingAbsoluteSeek = null;
-        this.seekToSessionTime(target, true);
+            this.pendingLocalTime = null;
+            this.pendingAutoplay = false;
+            this.seekPlaybackIntent = null;
+            this.playRequestId++;
+            this.gapCanAdvance = false;
+            this.currentTime.set(request.seconds);
+            this.gapMessage.set('此步骤的来源没有录像，请查看步骤截图或日志。');
+            this.videoRef?.nativeElement.pause();
+            this.isPlaying.set(false);
+            return;
           }
+          if (source.id !== this.agentService.selectedRecordingSource?.()) {
+            this.sourceSwitch = { time: request.seconds, play: this.isPlaying() };
+            this.agentService.selectRecordingSource(source.id);
+            return;
+          }
+        }
+        const video = this.videoRef?.nativeElement;
+        if (video && video.readyState >= 1 && !this.waitingForMetadata) {
+          this.pendingAbsoluteSeek = null;
+          this.seekToSessionTime(request.seconds, this.isPlaying());
+        }
+      });
     });
 
     effect(() => {
@@ -292,6 +388,7 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
   public onLoadedMetadata(): void {
     const v = this.videoRef?.nativeElement;
     if (v) {
+      this.waitingForMetadata = false;
       const segments = this.agentService.activeVideoSegments();
       if (!segments.length) {
         this.duration.set(v.duration || 0);
@@ -300,33 +397,34 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
       v.playbackRate = this.playbackRate();
       v.loop = this.isLooping() && !segments.length;
       v.muted = this.isMuted();
+      const autoplay = this.agentService.consumeVideoAutoplay();
+      if (autoplay) {
+        this.isMuted.set(true);
+        v.muted = true;
+      }
+      const shouldPlay = this.pendingAutoplay || autoplay;
+      this.pendingAutoplay = false;
       if (this.pendingAbsoluteSeek !== null) {
         const target = this.pendingAbsoluteSeek;
         this.pendingAbsoluteSeek = null;
-        this.seekToSessionTime(target, true);
+        this.seekToSessionTime(target, shouldPlay);
-        // A seek into another finalized segment changes the video source. Let
+        // Same-source seeks resume on seeked; cross-segment seeks resume after
-        // that segment's metadata event apply pendingLocalTime.
+        // the destination metadata and seeked. Never play the old frame here.
-        if (this.pendingLocalTime !== null) return;
+        return;
       }
       if (this.pendingLocalTime !== null) {
-        v.currentTime = Math.max(0, Math.min(v.duration || this.pendingLocalTime, this.pendingLocalTime));
+        const localTime = Math.max(0, Math.min(v.duration || this.pendingLocalTime, this.pendingLocalTime));
         this.pendingLocalTime = null;
+        this.seekWithinVideo(localTime, shouldPlay);
+      } else if (shouldPlay) {
+        this.playVideo();
       }
-      if (this.pendingAutoplay) {
+      if (this.pendingLocalTime === null) this.onTimeUpdate();
-        this.pendingAutoplay = false;
-        v.play().catch(() => {});
-      }
-      if (this.agentService.consumeVideoAutoplay()) {
-        this.isMuted.set(true);
-        v.muted = true;
-        v.play().catch(() => {
-          this.isPlaying.set(false);
-        });
-      }
     }
   }
 
   public onTimeUpdate(): void {
+    if (this.gapMessage() || this.pendingLocalTime !== null) return;
     const v = this.videoRef?.nativeElement;
     if (v) {
       const segments = this.agentService.activeVideoSegments();
@@ -339,29 +437,57 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
   }
 
   public onPlay(): void {
-    this.isPlaying.set(true);
+    this.isPlaying.set(!this.videoRef?.nativeElement.paused);
   }
 
   public onPause(): void {
     this.isPlaying.set(false);
   }
 
+  public onSeeked(): void {
+    const v = this.videoRef?.nativeElement;
+    if (!v || v.seeking || this.pendingLocalTime !== null) return;
+    const shouldPlay = this.seekPlaybackIntent;
+    this.seekPlaybackIntent = null;
+    this.onTimeUpdate();
+    if (shouldPlay) {
+      this.playVideo();
+    } else {
+      this.isPlaying.set(!v.paused && !v.ended);
+    }
+  }
+
   public onEnded(): void {
+    this.isPlaying.set(false);
     const segments = this.agentService.activeVideoSegments();
     if (segments.length && this.activeSegmentIndex() < segments.length - 1) {
-      this.pendingLocalTime = 0;
+      const current = segments[this.activeSegmentIndex()];
-      this.pendingAutoplay = true;
+      const next = segments[this.activeSegmentIndex() + 1];
-      this.activeSegmentIndex.update(index => index + 1);
+      this.seek(next.start, true);
+      if (this.agentService.usesSessionTimeline?.() && next.start - current.start - current.duration > 0.05) {
+        this.showSkippedInterval(current.start + current.duration, next.start);
+      }
       return;
     }
     if (segments.length && this.isLooping()) {
-      this.seek(0, true);
+      this.restart();
       return;
     }
+    const last = segments[segments.length - 1];
+    if (last && this.agentService.usesSessionTimeline?.()) {
+      this.currentTime.set(last.start + last.duration);
+      if (this.duration() - this.currentTime() > 0.05) {
+        this.skippedInterval.set('录像已结束，之后的任务时间未录制。');
+      }
+    }
     this.isPlaying.set(false);
   }
 
   public onError(): void {
+    this.playRequestId++;
+    this.pendingLocalTime = null;
+    this.pendingAutoplay = false;
+    this.seekPlaybackIntent = null;
     this.videoLoadError.set(true);
     this.isPlaying.set(false);
   }
@@ -370,17 +496,73 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
    * Playback actions
    */
   public togglePlay(): void {
+    if (this.gapMessage()) {
+      if (!this.gapCanAdvance) return;
+      const next = this.agentService.activeVideoSegments().find(s => s.start > this.currentTime());
+      if (next) {
+        const from = this.currentTime();
+        this.seek(next.start, true);
+        this.showSkippedInterval(from, next.start);
+      }
+      return;
+    }
     const v = this.videoRef?.nativeElement;
     if (!v) return;
+    if (v.ended && this.agentService.activeVideoSegments().length) {
+      this.restart();
+      return;
+    }
+    if (this.pendingLocalTime !== null) {
+      this.pendingAutoplay = !this.isPlaying();
+      this.isPlaying.set(this.pendingAutoplay);
+      return;
+    }
+    if (this.seekPlaybackIntent !== null) {
+      this.seekPlaybackIntent = !this.isPlaying();
+      this.isPlaying.set(this.seekPlaybackIntent);
+      return;
+    }
     if (v.paused) {
-      v.play().catch(() => {});
+      this.playVideo();
     } else {
+      this.playRequestId++;
       v.pause();
+      this.isPlaying.set(false);
     }
   }
 
+  private playVideo(): void {
+    const v = this.videoRef?.nativeElement;
+    if (!v) return;
+    const requestId = ++this.playRequestId;
+    v.play().then(() => {
+      if (requestId === this.playRequestId) this.isPlaying.set(!v.paused && !v.ended);
+    }).catch(() => {
+      // A rejected play request can leave paused=false without advancing frames.
+      // Reset the native state too, so the next click really starts playback.
+      if (requestId === this.playRequestId) {
+        v.pause();
+        this.isPlaying.set(false);
+      }
+    });
+  }
+
+  private seekWithinVideo(seconds: number, shouldPlay: boolean): void {
+    const v = this.videoRef?.nativeElement;
+    if (!v) return;
+    this.playRequestId++;
+    this.seekPlaybackIntent = shouldPlay;
+    // Cancel any unfinished play request before seeking; resume only once the
+    // browser has decoded the destination, including after a source switch.
+    v.pause();
+    this.isPlaying.set(false);
+    v.currentTime = seconds;
+    if (!v.seeking) this.onSeeked();
+  }
+
   public restart(): void {
-    this.seek(0, true);
+    const first = this.agentService.activeVideoSegments()[0];
+    this.seek(this.agentService.usesSessionTimeline?.() ? first?.start || 0 : 0, true);
   }
 
   /**
@@ -391,29 +573,66 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
    */
   public seekToSessionTime(sessionSeconds: number, autoplay = false): void {
     const segments = this.agentService.activeVideoSegments();
-    this.seek(sessionTimeToTimelineTime(segments, sessionSeconds), autoplay);
+    this.seek(this.agentService.usesSessionTimeline?.() ? sessionSeconds : sessionTimeToTimelineTime(segments, sessionSeconds), autoplay);
   }
 
-  /** Seek on the back-to-back playlist axis used by the scrubber. */
+  public selectSource(id: string): void {
+    if (id === this.agentService.selectedRecordingSource?.()) return;
+    this.sourceSwitch = { time: this.currentTime(), play: this.isPlaying() };
+    if (!this.agentService.selectRecordingSource(id)) this.sourceSwitch = null;
+  }
+
+  public seekToVideoMarker(marker: VideoStepMarker): void {
+    this.seek(marker.time);
+  }
+
+  public videoMarkerLabel(marker: VideoStepMarker): string {
+    const first = marker.steps[0];
+    const count = marker.steps.length;
+    const action = first.actionText || first.title;
+    return count > 1
+      ? `步骤 ${first.stepNumber} 等 ${count} 步 · ${this.formatTime(marker.time)} · ${action}`
+      : `步骤 ${first.stepNumber} · ${this.formatTime(marker.time)} · ${action}`;
+  }
+
+  /** Seek exactly: a step/source request must never display unrelated evidence. */
   public seek(seconds: number, autoplay = false): void {
     const v = this.videoRef?.nativeElement;
     if (!v) return;
+    this.skippedInterval.set('');
     const target = Math.max(0, Math.min(this.duration(), seconds));
     const segments = this.agentService.activeVideoSegments();
-    const location = locateTimelineTime(segments, target);
+    // URL-only legacy recordings have no coverage manifest. Preserve their
-    if (!location) {
+    // native video seeking instead of treating every timestamp as a gap.
-      v.currentTime = target;
+    const sessionTimeline = this.agentService.usesSessionTimeline?.() && segments.length > 0;
-      if (autoplay) v.play().catch(() => {});
+    const location = sessionTimeline ? locateCoveredSessionTime(segments, target) : locateTimelineTime(segments, target);
+    if (sessionTimeline && !location) {
+      this.playRequestId++;
+      this.pendingLocalTime = null;
+      this.pendingAutoplay = false;
+      this.seekPlaybackIntent = null;
+      v.pause();
+      this.isPlaying.set(false);
+      this.currentTime.set(target);
+      this.gapCanAdvance = true;
+      this.gapMessage.set('此时间段未录制（尚未开始、暂停或录制中断），可查看步骤截图与日志。');
       return;
     }
-    const { index, localTime } = location;
+    this.gapMessage.set('');
-    if (index === this.activeSegmentIndex()) {
+    const shouldPlay = autoplay || this.pendingAutoplay || (this.seekPlaybackIntent ?? !v.paused);
-      v.currentTime = localTime;
+    if (!location) {
-      if (autoplay) v.play().catch(() => {});
+      this.seekWithinVideo(target, shouldPlay);
+    } else if (location.index === this.activeSegmentIndex() && this.pendingLocalTime === null) {
+      this.seekWithinVideo(location.localTime, shouldPlay);
     } else {
-      this.pendingLocalTime = localTime;
+      this.playRequestId++;
-      this.pendingAutoplay = autoplay || !v.paused;
+      this.seekPlaybackIntent = null;
-      this.activeSegmentIndex.set(index);
+      v.pause();
+      this.isPlaying.set(false);
+      this.pendingLocalTime = location.localTime;
+      this.pendingAutoplay = shouldPlay;
+      this.waitingForMetadata = true;
+      this.activeSegmentIndex.set(location.index);
     }
     this.currentTime.set(target);
   }
@@ -421,9 +640,22 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
   public onScrub(event: Event): void {
     const input = event.target as HTMLInputElement;
     const targetTime = parseFloat(input.value);
+    const segments = this.agentService.activeVideoSegments();
+    if (this.agentService.usesSessionTimeline?.() && segments.length && !locateCoveredSessionTime(segments, targetTime)) {
+      const next = segments.find(segment => segment.start > targetTime);
+      if (next) {
+        this.seek(next.start);
+        this.showSkippedInterval(targetTime, next.start);
+        return;
+      }
+    }
     this.seek(targetTime);
   }
 
+  private showSkippedInterval(from: number, to: number): void {
+    this.skippedInterval.set(`已跳过未录制区间 ${this.formatTime(from)}–${this.formatTime(to)}`);
+  }
+
   public setSpeed(rate: number): void {
     this.playbackRate.set(rate);
     const v = this.videoRef?.nativeElement;
@@ -621,6 +853,7 @@ export class FloatingVideoPlayerComponent implements OnDestroy {
   }
 
   public ngOnDestroy(): void {
+    this.playRequestId++;
     if (this.stepTimer) {
       clearTimeout(this.stepTimer);
       this.stepTimer = null;

apps/showcase_ui/src/app/components/llm-request-status/llm-request-status.component.html

@@ -0,0 +1,80 @@
+@if (sessionId()) {
+  <section class="model-requests" aria-label="模型请求状态">
+    @for (review of activeReviews(); track review.review_id) {
+      <div class="request-live review-live" [class.stale]="unavailable()" aria-live="polite">
+        <div class="request-heading">
+          <span class="status-dot" [class.waiting]="!unavailable()"></span>
+          <strong>{{ unavailable() ? '验收状态暂时无法更新' : reviewStage(review) }}</strong>
+          <span class="elapsed">验收已用 {{ reviewElapsed(review) }} 秒</span>
+        </div>
+        @if (!unavailable()) {
+          <p>完整结论确认前，测试仍为待判定。剩余验收预算 {{ reviewRemaining(review) }} 秒 / 总计 {{ review.total_seconds }} 秒。</p>
+          <p class="notice">已发起 {{ review.request_count }} 次模型请求，已重试 {{ review.retry_count }} 次；重试复用已有证据。</p>
+        }
+      </div>
+    }
+    @for (request of active(); track request.attempt_id) {
+      <div class="request-live" [class.retrying]="request.phase === 'retry_wait'" [class.stale]="unavailable()">
+        <div class="request-heading" aria-live="polite">
+          <span class="status-dot" [class.waiting]="!unavailable()"></span>
+          <strong>{{ unavailable() ? '请求状态暂时无法更新' : phase(request) }}</strong>
+          <span class="model">{{ request.model || '模型未记录' }} · {{ role(request) }}{{ request.is_fallback ? ' · 兜底模型' : '' }}</span>
+          <span class="elapsed">已用 {{ elapsed(request) }} 秒</span>
+        </div>
+        @if (!unavailable()) {
+          <p>{{ hint(request, now()) }} @if (request.phase !== 'retry_wait') { 本次请求最长等待 {{ request.timeout_seconds }} 秒。 }</p>
+        }
+        @if (request.error) { <p class="error">{{ request.error }}</p> }
+      </div>
+    }
+    @if (unavailable()) { <p class="notice" role="status">诊断信息暂时不可用；显示的是上次记录，不能据此判断任务是否卡住。</p> }
+    <details>
+      <summary>模型请求诊断 @if (requests().length) { <span>· 最近 {{ requests().length }} 次尝试</span> }</summary>
+      <div class="diagnostics-header"><span>时间按本机时区显示；下载文件同时包含 UTC 时间。</span><a [href]="downloadUrl()" download>下载完整诊断</a></div>
+      @if (loading()) { <p class="notice">正在加载请求记录…</p> }
+      @if (!loading() && !unavailable() && !requests().length) { <p class="notice">暂无请求诊断。旧任务没有记录的时间和用量无法补推。</p> }
+      <div class="request-history">
+        @for (review of reviews(); track review.review_id) {
+          <details class="request-detail review-detail">
+            <summary>{{ review.started_at * 1000 | date:'HH:mm:ss' }} · 最终验收 · {{ reviewStage(review) }}</summary>
+            <dl>
+              <dt>验收编号</dt><dd>{{ review.review_id }} · {{ review.attempt_id }}</dd>
+              <dt>验收预算</dt><dd>总计 {{ review.total_seconds }} 秒 · 已用 {{ reviewElapsed(review) }} 秒</dd>
+              <dt>模型请求</dt><dd>{{ review.request_count }} 次，实际重试 {{ review.retry_count }} 次；每个请求最多重试 {{ review.max_retries }} 次，总预算内执行</dd>
+              <dt>开始时间</dt><dd>{{ review.started_at * 1000 | date:'yyyy-MM-dd HH:mm:ss.SSS Z' }}</dd>
+              <dt>预算截止</dt><dd>{{ review.deadline_at * 1000 | date:'yyyy-MM-dd HH:mm:ss.SSS Z' }}</dd>
+              <dt>结束时间</dt><dd>{{ review.ended_at ? (review.ended_at * 1000 | date:'HH:mm:ss.SSS') : '未记录' }}</dd>
+              @if (review.last_failure_category) { <dt>最近错误类别</dt><dd>{{ review.last_failure_category }}</dd> }
+              @if (review.status === 'budget_exhausted') { <dt>结束原因</dt><dd class="error">取证、模型请求与重试已耗尽验收总预算；未确认通过。</dd> }
+              @if (review.status === 'request_retries_exhausted') { <dt>结束原因</dt><dd class="error">验收模型请求重试已用尽；未确认通过，没有重跑设备操作。</dd> }
+              @if (review.status === 'cancelled') { <dt>结束原因</dt><dd>验收被停止或指导中断，不作为请求超时重试。</dd> }
+            </dl>
+          </details>
+        }
+        @for (request of requests(); track request.attempt_id) {
+          <details class="request-detail">
+            <summary><span>{{ request.started_at * 1000 | date:'HH:mm:ss' }}</span> {{ request.model || '模型' }} · {{ role(request) }} · {{ phase(request) }} @if (!request.active && !request.ended_at) { <span>（未记录结束，当前未在执行）</span> } @if (request.is_fallback) { <span class="fallback">兜底</span> }</summary>
+            <dl>
+              <dt>本地请求编号</dt><dd>{{ request.request_id }}</dd>
+              <dt>本次尝试编号</dt><dd>{{ request.attempt_id }}</dd>
+              <dt>服务端请求编号</dt><dd>{{ request.server_request_id || '服务未提供' }}</dd>
+              <dt>模型配置</dt><dd>{{ request.model || '未记录' }} / {{ request.reasoning_effort || '默认' }} · 第 {{ request.attempt }} 次请求，最多重试 {{ request.max_retries }} 次</dd>
+              @if (request.final_review; as review) { <dt>所属最终验收</dt><dd>{{ review.review_id }} · {{ reviewStage(review) }} · 请求记录时剩余 {{ review.remaining_seconds }} 秒</dd> }
+              <dt>输出 Token 上限</dt><dd>{{ request.max_output_tokens || '未设置（由模型服务决定）' }}</dd>
+              <dt>有效输出片段</dt><dd>{{ request.output_chunks ?? '未记录' }}（客户端计数，与服务端 SSE 事件数口径不同）</dd>
+              <dt>工具参数</dt><dd>{{ request.tool_argument_chars ?? '未记录' }} 字符 @for (tool of request.tools || []; track $index) { <span class="image-size">{{ tool.name || '工具名未提供' }} · {{ tool.argument_chars }} 字符</span> }</dd>
+              <dt>请求开始</dt><dd>{{ request.started_at * 1000 | date:'yyyy-MM-dd HH:mm:ss.SSS Z' }}</dd>
+              <dt>收到响应头</dt><dd>{{ request.headers_at ? (request.headers_at * 1000 | date:'HH:mm:ss.SSS') : '未记录' }} · HTTP {{ request.http_status || '未知' }}</dd>
+              <dt>首个有效输出</dt><dd>{{ request.first_output_at ? (request.first_output_at * 1000 | date:'HH:mm:ss.SSS') : '未收到／未记录' }}</dd>
+              <dt>最后有效输出</dt><dd>{{ request.last_output_at ? (request.last_output_at * 1000 | date:'HH:mm:ss.SSS') : '未收到／未记录' }}</dd>
+              <dt>结束时间</dt><dd>{{ request.ended_at ? (request.ended_at * 1000 | date:'HH:mm:ss.SSS') : '尚未记录' }}</dd>
+              <dt>输入图片</dt><dd>{{ request.images.length }} 张 @for (image of request.images; track $index) { <span class="image-size">{{ image.width && image.height ? image.width + ' × ' + image.height : '尺寸未知' }}</span> }</dd>
+              <dt>Token 用量</dt><dd>{{ request.usage ? '输入 ' + request.usage.prompt_tokens + ' / 输出 ' + request.usage.completion_tokens + ' / 合计 ' + request.usage.total_tokens : '未知（服务未报告）' }}</dd>
+              @if (request.error) { <dt>失败原因</dt><dd class="error">{{ request.error_type }} · {{ request.error }} · 阶段 {{ phase(request.failure_phase || '') }} @if (request.category) { · {{ request.category }} }</dd> }
+            </dl>
+          </details>
+        }
+      </div>
+    </details>
+  </section>
+}

apps/showcase_ui/src/app/components/llm-request-status/llm-request-status.component.scss

@@ -0,0 +1,26 @@
+:host { display: block; }
+.model-requests { margin: 10px 0 16px; color: #475569; font-size: 12px; }
+.request-live { padding: 12px 14px; border: 1px solid #dbeafe; border-radius: 10px; background: #f8fbff; margin-bottom: 8px; }
+.request-heading { display: flex; align-items: center; flex-wrap: wrap; gap: 8px; }
+strong { color: #334155; font-size: 13px; }
+.model { color: #64748b; overflow-wrap: anywhere; }
+.elapsed { margin-left: auto; color: #1d4ed8; font-variant-numeric: tabular-nums; }
+.status-dot { height: 7px; width: 7px; border-radius: 50%; background: #64748b; flex: 0 0 auto; }
+.waiting { background: #3b82f6; animation: pulse 1.5s ease-in-out infinite; }
+.retrying { border-color: #fde68a; background: #fffbeb; .elapsed { color: #92400e; } .waiting { background: #d97706; } }
+.stale { border-color: #cbd5e1; background: #f8fafc; }
+p { margin: 6px 0 0; line-height: 1.7; }
+.error { color: #b45309; overflow-wrap: anywhere; }
+summary { cursor: pointer; padding: 5px 0; line-height: 1.7; } summary:hover { color: #2563eb; }
+summary:focus-visible, a:focus-visible { outline: 2px solid #3b82f6; outline-offset: 2px; }
+.diagnostics-header { display: flex; align-items: center; justify-content: space-between; flex-wrap: wrap; gap: 8px; margin: 8px 0; color: #64748b; }
+a { color: #2563eb; text-decoration: none; border: 1px solid #dbeafe; border-radius: 6px; padding: 4px 9px; }
+.request-history { max-height: 360px; overflow-y: auto; }
+.request-detail { border-top: 1px solid #e2e8f0; padding: 4px 0; }
+dl { display: grid; grid-template-columns: 110px minmax(0, 1fr); gap: 7px 10px; margin: 8px 0 12px; line-height: 1.6; }
+dt { color: #64748b; } dd { margin: 0; overflow-wrap: anywhere; font-variant-numeric: tabular-nums; }
+.image-size, .fallback { display: inline-block; padding: 0 5px; margin: 2px 3px; background: #f1f5f9; border-radius: 4px; }
+.notice { color: #64748b; }
+@keyframes pulse { 50% { opacity: .35; } }
+@media (prefers-reduced-motion: reduce) { .waiting { animation: none; } }
+@media (max-width: 480px) { dl { grid-template-columns: 90px minmax(0, 1fr); } .elapsed { margin-left: 15px; } }

apps/showcase_ui/src/app/components/llm-request-status/llm-request-status.component.spec.ts

@@ -0,0 +1,121 @@
+import { fakeAsync, TestBed, tick } from '@angular/core/testing';
+import { provideHttpClient } from '@angular/common/http';
+import { HttpTestingController, provideHttpClientTesting } from '@angular/common/http/testing';
+import { FinalReview, LlmRequestStatusComponent, ModelRequest, requestHint, reviewStage } from './llm-request-status.component';
+
+const waiting: ModelRequest = {
+  request_id: 'call-1', attempt_id: 'attempt-1', model: 'gpt-6-luna', role: 'FlashRunner',
+  is_fallback: false, phase: 'waiting_output', active: true, attempt: 1, max_retries: 5,
+  timeout_seconds: 300, elapsed_seconds: 42, started_at: Date.now() / 1000 - 42,
+  http_status: 200, usage: null, images: []
+};
+
+describe('模型请求诊断', () => {
+  beforeEach(() => TestBed.configureTestingModule({ providers: [provideHttpClient(), provideHttpClientTesting()] }));
+  afterEach(() => TestBed.inject(HttpTestingController).verify());
+
+  it('收到200仍显示等待，未知用量不当作0；完成后取消轮询', fakeAsync(() => {
+    const fixture = TestBed.createComponent(LlmRequestStatusComponent);
+    fixture.componentRef.setInput('sessionId', 'task-one');
+    fixture.componentRef.setInput('running', true);
+    fixture.detectChanges();
+    const http = TestBed.inject(HttpTestingController);
+    http.expectOne('/api/sessions/task-one/request-diagnostics').flush({ requests: [waiting] });
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.request-live').textContent).toContain('等待模型输出');
+    expect(fixture.nativeElement.textContent).toContain('未知（服务未报告）');
+    expect(fixture.nativeElement.querySelector('details').open).toBeFalse();
+    tick(3000);
+    http.expectOne('/api/sessions/task-one/request-diagnostics').flush({ requests: [{ ...waiting, active: false, phase: 'completed', ended_at: Date.now() / 1000 }] });
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.request-live')).toBeNull();
+    fixture.destroy(); tick(5000);
+    http.expectNone('/api/sessions/task-one/request-diagnostics');
+  }));
+
+  it('切换会话取消旧请求，历史失败不会显示正在重试', () => {
+    const fixture = TestBed.createComponent(LlmRequestStatusComponent);
+    const http = TestBed.inject(HttpTestingController);
+    fixture.componentRef.setInput('sessionId', 'old'); fixture.detectChanges();
+    const old = http.expectOne('/api/sessions/old/request-diagnostics');
+    fixture.componentRef.setInput('sessionId', 'new'); fixture.detectChanges();
+    expect(old.cancelled).toBeTrue();
+    http.expectOne('/api/sessions/new/request-diagnostics').flush({ requests: [{ ...waiting, phase: 'retry_wait' }] });
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.request-live')).toBeNull();
+    expect(fixture.nativeElement.querySelector('a').href).toContain('/new/request-diagnostics?download=true');
+    fixture.destroy();
+  });
+
+  it('诊断接口失败时明确提示状态未知', () => {
+    const fixture = TestBed.createComponent(LlmRequestStatusComponent);
+    fixture.componentRef.setInput('sessionId', 'task'); fixture.detectChanges();
+    TestBed.inject(HttpTestingController).expectOne('/api/sessions/task/request-diagnostics').flush({}, { status: 503, statusText: 'Unavailable' });
+    fixture.detectChanges();
+    expect(fixture.nativeElement.textContent).toContain('不能据此判断任务是否卡住');
+    fixture.destroy();
+  });
+
+  it('重试倒计时到0后不谎称已发送请求', () => {
+    expect(requestHint({ ...waiting, phase: 'retry_wait', retry_at: 110 }, 105)).toContain('5 秒后进行第 1/5 次重试');
+    expect(requestHint({ ...waiting, phase: 'retry_wait', retry_at: 110 }, 111)).toContain('等待下一次请求状态');
+  });
+
+  it('持续生成工具参数时展示工具名和字符数，不误报工具已执行', () => {
+    const fixture = TestBed.createComponent(LlmRequestStatusComponent);
+    fixture.componentRef.setInput('sessionId', 'tool-task');
+    fixture.componentRef.setInput('running', true);
+    fixture.detectChanges();
+    TestBed.inject(HttpTestingController).expectOne('/api/sessions/tool-task/request-diagnostics').flush({ requests: [{
+      ...waiting, phase: 'generating_tool_arguments', tool_argument_chars: 15000,
+      tools: [{ name: 'click', argument_chars: 15000 }], output_chunks: 200
+    }] });
+    fixture.detectChanges();
+    const text = fixture.nativeElement.querySelector('.request-live').textContent;
+    expect(text).toContain('正在生成工具参数');
+    expect(text).toContain('click');
+    expect(text).toContain((15000).toLocaleString());
+    expect(text).toContain('本次工具尚未执行');
+    fixture.destroy();
+  });
+
+  it('无模型请求时仍展示验收阶段与总预算，耗尽后保留明确原因', fakeAsync(() => {
+    const fixture = TestBed.createComponent(LlmRequestStatusComponent);
+    const http = TestBed.inject(HttpTestingController);
+    const now = Date.now() / 1000;
+    const review: FinalReview = {
+      review_id: 'review-1', attempt_id: 'final#1', status: 'running', stage: 'reading_evidence', active: true,
+      started_at: now - 20, deadline_at: now + 80, total_seconds: 100, elapsed_seconds: 20, remaining_seconds: 80,
+      request_count: 2, retry_count: 1, max_retries: 5
+    };
+    fixture.componentRef.setInput('sessionId', 'review-task');
+    fixture.componentRef.setInput('running', true);
+    fixture.detectChanges();
+    http.expectOne('/api/sessions/review-task/request-diagnostics').flush({ requests: [], final_reviews: [review] });
+    fixture.detectChanges();
+    const text = fixture.nativeElement.querySelector('.review-live').textContent;
+    expect(text).toContain('正在读取验收证据');
+    expect(text).toContain('测试仍为待判定');
+    expect(text).toContain('剩余验收预算 80 秒');
+    expect(text).toContain('已重试 1 次');
+    tick(1000); fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.review-live').textContent).toContain('剩余验收预算 79 秒');
+    tick(2000);
+    http.expectOne('/api/sessions/review-task/request-diagnostics').flush({ requests: [], final_reviews: [{
+      ...review, active: false, status: 'budget_exhausted', ended_at: now + 3, elapsed_seconds: 100
+    }] });
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.review-live')).toBeNull();
+    expect(fixture.nativeElement.textContent).toContain('验收总预算已用尽');
+    expect(fixture.nativeElement.textContent).toContain('未确认通过');
+    fixture.destroy(); tick(3000);
+    http.expectNone('/api/sessions/review-task/request-diagnostics');
+  }));
+
+  it('区分重试耗尽、验收中断与完成，不把完成当成通过', () => {
+    const review = { status: 'request_retries_exhausted' } as FinalReview;
+    expect(reviewStage(review)).toBe('验收请求重试已用尽');
+    expect(reviewStage({ ...review, status: 'cancelled' })).toBe('验收已中断');
+    expect(reviewStage({ ...review, status: 'completed' })).toBe('验收完成');
+  });
+});

apps/showcase_ui/src/app/components/llm-request-status/llm-request-status.component.ts

@@ -0,0 +1,124 @@
+import { ChangeDetectionStrategy, Component, computed, effect, inject, input, output, signal } from '@angular/core';
+import { DatePipe } from '@angular/common';
+import { HttpClient } from '@angular/common/http';
+import { Subscription } from 'rxjs';
+
+export interface ModelRequest {
+  request_id: string; attempt_id: string; model?: string; reasoning_effort?: string;
+  role: string; is_fallback: boolean; phase: string; active: boolean;
+  attempt: number; max_retries: number; timeout_seconds: number; elapsed_seconds: number;
+  started_at: number; headers_at?: number; first_output_at?: number; last_output_at?: number;
+  ended_at?: number; retry_at?: number; http_status?: number; server_request_id?: string;
+  error?: string; error_type?: string; failure_phase?: string;
+  output_chunks?: number; tool_argument_chars?: number; max_output_tokens?: number | null;
+  tools?: { name: string | null; argument_chars: number }[];
+  usage: { prompt_tokens: number; completion_tokens: number; total_tokens: number } | null;
+  images: { width: number | null; height: number | null }[];
+  final_review?: FinalReview;
+  category?: string;
+}
+
+export interface FinalReview {
+  review_id: string; attempt_id: string; status: string; stage: string; active: boolean;
+  started_at: number; deadline_at: number; ended_at?: number;
+  total_seconds: number; elapsed_seconds: number; remaining_seconds: number;
+  request_count: number; retry_count: number; request_attempt?: number; max_retries: number;
+  request_phase?: string; last_failure_category?: string;
+}
+
+export function reviewStage(review: FinalReview): string {
+  if (review.status !== 'running') {
+    return ({ completed: '验收完成', budget_exhausted: '验收总预算已用尽', request_retries_exhausted: '验收请求重试已用尽', cancelled: '验收已中断', failed: '验收未完成' } as Record<string, string>)[review.status] || '验收状态待核实';
+  }
+  return ({ preparing_evidence: '正在准备验收证据', reviewing_evidence: '正在核对证据与结论', reading_evidence: '正在读取验收证据', structuring_verdict: '正在整理验收结论', validating_verdict: '正在校验完整结论' } as Record<string, string>)[review.stage] || '正在进行最终验收';
+}
+
+export function requestPhase(request: ModelRequest | string): string {
+  const phase = typeof request === 'string' ? request : request.phase;
+  return ({ waiting_response: '等待模型响应', waiting_output: '等待模型输出', receiving_output: '正在接收模型输出', generating_tool_arguments: '正在生成工具参数', retry_wait: '等待重试', completed: '请求完成', failed: '请求失败', cancelled: '请求已取消' } as Record<string, string>)[phase] || '请求状态待核实';
+}
+
+export function requestHint(request: ModelRequest, now: number): string {
+  if (request.phase === 'retry_wait') {
+    const remaining = Math.max(0, Math.ceil((request.retry_at || now) - now));
+    return remaining ? `${remaining} 秒后进行第 ${request.attempt}/${request.max_retries} 次重试。` : '重试等待已结束，正在等待下一次请求状态。';
+  }
+  if (request.phase === 'waiting_output') return '已收到服务响应，尚未收到有效模型输出。';
+  if (request.phase === 'generating_tool_arguments') {
+    const names = (request.tools || []).map(tool => tool.name).filter(Boolean).join('、');
+    return `${names ? names + ' · ' : ''}已接收 ${(request.tool_argument_chars || 0).toLocaleString()} 个参数字符，最近输出在 ${Math.max(0, Math.floor(now - (request.last_output_at || now)))} 秒前；本次工具尚未执行。`;
+  }
+  if (request.phase === 'receiving_output') return `最近一次输出在 ${Math.max(0, Math.floor(now - (request.last_output_at || now)))} 秒前；完整响应仍在接收中。`;
+  return '正在等待服务响应；尚不能确认模型是否开始生成。';
+}
+
+@Component({
+  selector: 'app-llm-request-status', standalone: true, imports: [DatePipe],
+  changeDetection: ChangeDetectionStrategy.OnPush,
+  templateUrl: './llm-request-status.component.html', styleUrl: './llm-request-status.component.scss'
+})
+export class LlmRequestStatusComponent {
+  private http = inject(HttpClient);
+  sessionId = input<string | null>(null);
+  running = input(false);
+  activeChange = output<boolean>();
+  requests = signal<ModelRequest[]>([]);
+  reviews = signal<FinalReview[]>([]);
+  now = signal(Date.now() / 1000);
+  unavailable = signal(false);
+  loading = signal(false);
+  active = computed(() => this.running() ? this.requests().filter(r => r.active) : []);
+  activeReviews = computed(() => this.running() ? this.reviews().filter(r => r.active) : []);
+  downloadUrl = computed(() => `/api/sessions/${encodeURIComponent(this.sessionId() || '')}/request-diagnostics?download=true`);
+  phase = requestPhase;
+  hint = requestHint;
+  reviewStage = reviewStage;
+
+  constructor() {
+    effect(() => this.activeChange.emit(this.active().length > 0 || this.activeReviews().length > 0));
+    effect(onCleanup => {
+      const id = this.sessionId();
+      const running = this.running();
+      this.requests.set([]);
+      this.reviews.set([]);
+      this.unavailable.set(false);
+      this.loading.set(!!id);
+      if (!id) return;
+      let stopped = false;
+      let timer: ReturnType<typeof setTimeout> | undefined;
+      let subscription: Subscription | undefined;
+      const schedule = () => { if (running && !stopped) timer = setTimeout(load, 3000); };
+      const load = () => {
+        subscription = this.http.get<{ requests: ModelRequest[]; final_reviews?: FinalReview[] }>(`/api/sessions/${encodeURIComponent(id)}/request-diagnostics`).subscribe({
+          next: data => {
+            this.requests.set(data.requests || []);
+            this.reviews.set(data.final_reviews || []);
+            this.unavailable.set(false);
+            this.loading.set(false);
+            schedule();
+          },
+          error: () => { this.unavailable.set(true); this.loading.set(false); schedule(); }
+        });
+      };
+      load();
+      const clock = running ? setInterval(() => this.now.set(Date.now() / 1000), 1000) : undefined;
+      onCleanup(() => { stopped = true; clearTimeout(timer); clearInterval(clock); subscription?.unsubscribe(); });
+    });
+  }
+
+  elapsed(request: ModelRequest): number {
+    return Math.max(0, Math.floor(request.ended_at ? request.elapsed_seconds : this.now() - request.started_at));
+  }
+
+  reviewElapsed(review: FinalReview): number {
+    return Math.max(0, Math.floor(review.ended_at ? review.elapsed_seconds : this.now() - review.started_at));
+  }
+
+  reviewRemaining(review: FinalReview): number {
+    return Math.max(0, Math.ceil(review.deadline_at - this.now()));
+  }
+
+  role(request: ModelRequest): string {
+    return ({ FlashRunner: '任务执行', operator: '任务执行', planner: '规划', checker: '结果检查', outputter: '报告整理', step_summarizer: '步骤整理' } as Record<string, string>)[request.role] || request.role;
+  }
+}

apps/showcase_ui/src/app/components/model-selector/model-selector.component.spec.ts

@@ -0,0 +1,28 @@
+import { TestBed } from '@angular/core/testing';
+import { ModelSelectorComponent } from './model-selector.component';
+
+describe('Shared model selector', () => {
+  it('preserves a saved custom or retired model without silently replacing it', () => {
+    const fixture = TestBed.createComponent(ModelSelectorComponent);
+    fixture.componentRef.setInput('value', 'mimo-v2.6-flash');
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('select').value).toBe('__custom__');
+    expect(fixture.nativeElement.querySelector('input').value).toBe('mimo-v2.6-flash');
+    expect(Array.from(fixture.nativeElement.querySelectorAll('option'), (o: any) => o.value))
+      .not.toContain('mimo-v2.6-flash');
+  });
+
+  it('allows editing custom names even when the typed name matches a preset', () => {
+    const fixture = TestBed.createComponent(ModelSelectorComponent);
+    fixture.componentRef.setInput('value', 'gpt-6-luna');
+    fixture.detectChanges();
+    const select: HTMLSelectElement = fixture.nativeElement.querySelector('select');
+    select.value = '__custom__'; select.dispatchEvent(new Event('change')); fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('input').value).toBe('gpt-6-luna');
+    fixture.componentRef.setInput('value', 'gpt-6.1-sol'); fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('input').value).toBe('gpt-6.1-sol');
+    fixture.componentRef.setInput('disabled', true); fixture.detectChanges();
+    expect(select.disabled).toBeTrue();
+    expect(fixture.nativeElement.querySelector('input').disabled).toBeTrue();
+  });
+});

apps/showcase_ui/src/app/components/model-selector/model-selector.component.ts

@@ -0,0 +1,56 @@
+import { ChangeDetectionStrategy, Component, computed, input, output, signal } from '@angular/core';
+import { MODEL_PRESETS } from '../../core/model-presets';
+
+let nextId = 0;
+
+@Component({
+  selector: 'app-model-selector',
+  standalone: true,
+  changeDetection: ChangeDetectionStrategy.OnPush,
+  template: `
+    <label [for]="id">{{label()}}</label>
+    <div class="selector">
+      <select [id]="id" [value]="selection()" [disabled]="disabled()"
+        (change)="choose($any($event.target).value)">
+        @for (preset of presets(); track preset) { <option [value]="preset">{{preset}}</option> }
+        <option value="__custom__">自定义模型…</option>
+      </select>
+      <span class="material-symbols-outlined" aria-hidden="true">expand_more</span>
+    </div>
+    @if (selection() === '__custom__') {
+      <input type="text" [value]="value()" (input)="valueChange.emit($any($event.target).value)"
+        [attr.aria-label]="label() + '（自定义）'" [disabled]="disabled()" required maxlength="200"
+        autocomplete="off" autocapitalize="off" spellcheck="false" placeholder="输入接口支持的模型名称" />
+    }
+  `,
+  styles: [`
+    :host { display: grid; gap: 7px; min-width: 0; color: #334155; font-size: 13px; }
+    label { font-weight: 600; line-height: 1.5; }
+    .selector { position: relative; min-width: 0; }
+    select, input { box-sizing: border-box; width: 100%; min-width: 0; min-height: 40px;
+      padding: 10px 12px; border: 1px solid #dbe3ec; border-radius: 10px; background: #fff;
+      color: inherit; font: inherit; }
+    select { appearance: none; padding-right: 36px; cursor: pointer; text-overflow: ellipsis; }
+    .selector > span { position: absolute; right: 11px; top: 50%; transform: translateY(-50%);
+      pointer-events: none; font-size: 19px; color: #64748b; }
+    select:hover:not(:disabled), input:hover:not(:disabled) { border-color: #a5b4fc; }
+    select:focus-visible, input:focus-visible { outline: 2px solid #a5b4fc; outline-offset: 2px; }
+    select:disabled, input:disabled { opacity: .5; cursor: default; }
+    option { background: #fff; color: #334155; }
+  `]
+})
+export class ModelSelectorComponent {
+  readonly id = `model-selector-${nextId++}`;
+  label = input('模型');
+  value = input.required<string>();
+  presets = input<string[]>(MODEL_PRESETS);
+  disabled = input(false);
+  valueChange = output<string>();
+  private custom = signal(false);
+  selection = computed(() => this.custom() || !this.presets().includes(this.value()) ? '__custom__' : this.value());
+
+  choose(value: string): void {
+    this.custom.set(value === '__custom__');
+    if (value !== '__custom__') this.valueChange.emit(value);
+  }
+}

apps/showcase_ui/src/app/components/model-settings/capability-settings.component.ts

@@ -1,12 +1,26 @@
 import { Component, inject, OnInit, signal, output } from '@angular/core';
 import { FormsModule } from '@angular/forms';
 import { HttpClient } from '@angular/common/http';
-import { finalize } from 'rxjs';
+import { finalize, timeout } from 'rxjs';
+import { UI_READ_TIMEOUT_MS } from '../../core/http-timeouts';
 
 interface Capabilities {
+  recording?: { codec: string; max_long_edge: number; fps: number; h265_kbps: number; h264_kbps: number; computer_scope: string; minimum_disk_mb: number; hevc_playback: boolean };
+  adb_auto_connect: boolean;
+  adb_discovery_interval_seconds: number;
+  adb_discovery_timeout_seconds: number;
   video_enabled: boolean;
+  recording_enabled: boolean;
   outputter_enabled: boolean;
   step_summarizer_enabled: boolean;
+  screenshot_limit_enabled: boolean;
+  screenshot_max_long_edge: number;
+  task_timeout_seconds: number;
+  observation_timeout_seconds: number;
+  observation_max_attempts: number;
+  device_command_timeout_seconds: number;
+  action_failure_attempts: number;
+  worker_cleanup_timeout_seconds: number;
 }
 
 @Component({
@@ -15,13 +29,47 @@ interface Capabilities {
     <section aria-label="功能开关">
       <h3>功能开关</h3>
       <p>按需启用，保存后对新任务生效。</p>
+      <label><input type="checkbox" role="switch" [(ngModel)]="config.adb_auto_connect" [disabled]="busy() || loading() || !loaded()" />
+        <span>自动连接设备<small>刷新设备时发现本机常见模拟器，并重连手动连接过的 TCP 设备；重启 ADB 后也会重连。不扫描局域网，不自动选择任务设备。保存后立即生效。</small></span></label>
       <label><input type="checkbox" role="switch" [(ngModel)]="config.video_enabled" [disabled]="busy() || loading() || !loaded()" />
-        <span>视频分析与录屏<small>录制设备画面，并允许智能体分析录像；关闭可减少资源占用。</small></span></label>
+        <span>视频分析<small>允许智能体分析任务画面。运行期间会临时录制，完成后不保留录像（除非开启录屏）。</small></span></label>
+      <label><input type="checkbox" role="switch" [(ngModel)]="config.recording_enabled" [disabled]="busy() || loading() || !loaded()" />
+        <span>保存录屏<small>保留任务录像以便复盘，可单独开启，不要求启用视频分析。</small></span></label>
+      @if (config.recording; as recording) {
+        <details class="execution-settings">
+          <summary>录像留档设置</summary>
+          <p>默认低码率留档，不影响模型观察截图。仅录实际使用的来源；屏幕模式包含该屏幕其他可见内容。</p>
+          <div class="number-fields">
+            <label>编码<select [(ngModel)]="recording.codec" [disabled]="busy() || !loaded()"><option value="auto">自动选择兼容编码</option><option value="h265">优先 H.265</option><option value="h264">H.264</option></select></label>
+            <label>电脑录制范围<select [(ngModel)]="recording.computer_scope" [disabled]="busy() || !loaded()"><option value="screen">当前控制的屏幕</option><option value="window">仅目标窗口</option></select></label>
+            <label>最长边（像素）<input type="number" min="320" max="4096" [(ngModel)]="recording.max_long_edge" [disabled]="busy() || !loaded()" /></label>
+            <label>最高帧率<input type="number" min="5" max="60" [(ngModel)]="recording.fps" [disabled]="busy() || !loaded()" /></label>
+            <label>H.265 目标码率（kbps）<input type="number" min="128" max="10000" [(ngModel)]="recording.h265_kbps" [disabled]="busy() || !loaded()" /></label>
+            <label>H.264 目标码率（kbps）<input type="number" min="128" max="20000" [(ngModel)]="recording.h264_kbps" [disabled]="busy() || !loaded()" /></label>
+          </div>
+        </details>
+      }
       <label><input type="checkbox" role="switch" [(ngModel)]="config.outputter_enabled" [disabled]="busy() || loading() || !loaded()" />
         <span>最终报告生成<small>根据执行记录生成任务结果报告。</small></span></label>
       <label><input type="checkbox" role="switch" [(ngModel)]="config.step_summarizer_enabled" [disabled]="busy() || loading() || !loaded()" />
         <span>步骤摘要压缩<small>Flash 模式下在后台生成摘要，减少长任务的上下文占用。</small></span></label>
-      <button type="button" (click)="save()" [disabled]="busy() || loading() || !loaded()">{{ busy() ? '正在保存…' : '保存功能设置' }}</button>
+      <label><input type="checkbox" role="switch" [(ngModel)]="config.screenshot_limit_enabled" [disabled]="busy() || loading() || !loaded()" />
+        <span>限制截图分辨率<small>开启时长边最多 {{ config.screenshot_max_long_edge }} 像素；关闭后使用设备原始分辨率。JSON 配置可覆盖开关和上限。</small></span></label>
+      <details class="execution-settings">
+        <summary>执行保护</summary>
+        <p>限制故障等待与反复失败；保存后对新任务生效，任务 JSON 可单独覆盖。</p>
+        <div class="number-fields">
+          <label>任务总超时（秒）<input type="number" min="30" max="86400" step="1" [(ngModel)]="config.task_timeout_seconds" [disabled]="busy() || !loaded()" /><small>不包含排队和用户暂停时间。</small></label>
+          <label>屏幕观测超时（秒）<input type="number" min="1" max="300" [(ngModel)]="config.observation_timeout_seconds" [disabled]="busy() || !loaded()" /></label>
+          <label>观测恢复最多尝试次数<input type="number" min="1" max="5" step="1" [(ngModel)]="config.observation_max_attempts" [disabled]="busy() || !loaded()" /></label>
+          <label>ADB 命令超时（秒）<input type="number" min="1" max="60" [(ngModel)]="config.device_command_timeout_seconds" [disabled]="busy() || !loaded()" /></label>
+          <label>连续动作失败次数<input type="number" min="1" max="10" step="1" [(ngModel)]="config.action_failure_attempts" [disabled]="busy() || !loaded()" /><small>先重新观察，再让兜底模型处理当前步骤。</small></label>
+          <label>任务收尾超时（秒）<input type="number" min="1" max="300" [(ngModel)]="config.worker_cleanup_timeout_seconds" [disabled]="busy() || !loaded()" /><small>已有最终结果但进程未退出时，回收进程以释放队列。</small></label>
+          <label>设备自动发现间隔（秒）<input type="number" min="1" max="300" [(ngModel)]="config.adb_discovery_interval_seconds" [disabled]="busy() || !loaded()" /></label>
+          <label>每轮自动连接预算（秒）<input type="number" min="1" max="30" [(ngModel)]="config.adb_discovery_timeout_seconds" [disabled]="busy() || !loaded()" /></label>
+        </div>
+      </details>
+      <button type="button" (click)="save()" [disabled]="busy() || loading() || !loaded() || !valid()">{{ busy() ? '正在保存…' : '保存功能设置' }}</button>
       @if (message()) { <p role="status">{{ message() }}</p> }
       @if (error()) { <p role="alert" class="error">{{ error() }}</p> }
     </section>
@@ -31,24 +79,56 @@ interface Capabilities {
     label { display: flex; gap: 12px; align-items: flex-start; padding: 14px 0; cursor: pointer; }
     input { width: 20px; height: 20px; accent-color: #0284c7; flex-shrink: 0; }
     small { display: block; margin-top: 4px; } button { padding: 10px 18px; border-radius: 10px; border: 1px solid #38bdf877; background: #0284c7; color: white; font: inherit; cursor: pointer; }
-    button:disabled { opacity: .5; cursor: default; } .error { color: #be123c; }`]
+    button:disabled { opacity: .5; cursor: default; } .error { color: #be123c; }
+    .execution-settings { margin: 12px 0 20px; padding: 12px 14px; border: 1px solid #cbd5e1; border-radius: 10px; background: #f8fafc; }
+    summary { cursor: pointer; font-weight: 600; color: #334155; }
+    .number-fields { display: grid; grid-template-columns: 1fr 1fr; gap: 12px 16px; }
+    .number-fields label { display: block; padding: 0; color: #334155; }
+    .number-fields input, .number-fields select { box-sizing: border-box; display: block; width: 100%; height: auto; margin-top: 8px; padding: 9px; border: 1px solid #cbd5e1; border-radius: 8px; font: inherit; }
+    section {font-size:13px;border-color:#e2e8f0}h3 {font-size:15px;color:#334155}
+    section > label {align-items:center;flex-direction:row-reverse;justify-content:space-between;border-bottom:1px solid #f1f5f9;gap:20px}section > label > span {flex:1}small {font-size:12px}
+    input[role=switch] {appearance:none;position:relative;width:36px;height:22px;border:1px solid #cbd5e1;border-radius:20px;background:#e2e8f0;cursor:pointer;transition:background .15s}
+    input[role=switch]::after {content:'';position:absolute;top:2px;left:2px;width:16px;height:16px;border-radius:50%;background:#fff;box-shadow:0 1px 3px #0f172a25;transition:transform .15s}
+    input[role=switch]:checked {background:#6366f1;border-color:#6366f1}input[role=switch]:checked::after {transform:translateX(14px)}input[role=switch]:focus-visible {outline:2px solid #a5b4fc;outline-offset:3px}input:disabled {opacity:.5;cursor:default}
+    button {background:#4f46e5;border-color:#4f46e5;border-radius:12px;font-weight:600}.execution-settings {border-color:#e2e8f0;border-radius:14px;padding:14px 16px}.number-fields input,.number-fields select {background:#fff;border-color:#dbe3ec}
+    @media(max-width: 480px) { .number-fields { grid-template-columns: 1fr; } }`]
 })
 export class CapabilitySettingsComponent implements OnInit {
   changed = output<Capabilities>();
   private http = inject(HttpClient);
-  config: Capabilities = { video_enabled: false, outputter_enabled: true, step_summarizer_enabled: true };
+  config: Capabilities = { adb_auto_connect: true, adb_discovery_interval_seconds: 10, adb_discovery_timeout_seconds: 5, video_enabled: false, recording_enabled: false, outputter_enabled: true, step_summarizer_enabled: true, screenshot_limit_enabled: true, screenshot_max_long_edge: 2048, task_timeout_seconds: 1800, observation_timeout_seconds: 60, observation_max_attempts: 3, device_command_timeout_seconds: 10, action_failure_attempts: 3, worker_cleanup_timeout_seconds: 30 };
   busy = signal(false); loading = signal(true); loaded = signal(false); message = signal(''); error = signal('');
   ngOnInit(): void {
-    this.http.get<Capabilities>('/api/system/capability-settings').pipe(finalize(() => this.loading.set(false))).subscribe({
+    this.http.get<Capabilities>('/api/system/capability-settings').pipe(timeout(UI_READ_TIMEOUT_MS), finalize(() => this.loading.set(false))).subscribe({
       next: config => { this.config = config; this.loaded.set(true); this.changed.emit(config); }, error: () => this.error.set('功能配置加载失败，请刷新页面重试。')
     });
   }
   save(): void {
-    if (this.busy() || this.loading() || !this.loaded()) return;
+    if (this.busy() || this.loading() || !this.loaded() || !this.valid()) return;
+    if (this.config.recording) {
+      this.config.recording.hevc_playback = !!document.createElement('video').canPlayType('video/mp4; codecs="hvc1"');
+    }
     this.busy.set(true); this.message.set(''); this.error.set('');
     this.http.post<Capabilities>('/api/system/capability-settings', this.config).pipe(finalize(() => this.busy.set(false))).subscribe({
       next: config => { this.config = config; this.changed.emit(config); this.message.set('功能设置已保存，新任务将使用此配置。'); },
       error: () => this.error.set('保存失败，请检查服务器连接和 .env 文件权限。')
     });
   }
+  valid(): boolean {
+    const c = this.config;
+    const r = c.recording;
+    if (r && (!['auto', 'h264', 'h265'].includes(r.codec) || !['screen', 'window'].includes(r.computer_scope)
+      || !Number.isInteger(r.max_long_edge) || r.max_long_edge < 320 || r.max_long_edge > 4096
+      || !Number.isInteger(r.fps) || r.fps < 5 || r.fps > 60
+      || !Number.isInteger(r.h265_kbps) || r.h265_kbps < 128 || r.h265_kbps > 10000
+      || !Number.isInteger(r.h264_kbps) || r.h264_kbps < 128 || r.h264_kbps > 20000)) return false;
+    return Number.isInteger(c.task_timeout_seconds) && c.task_timeout_seconds >= 30 && c.task_timeout_seconds <= 86400
+      && c.observation_timeout_seconds > 0 && c.observation_timeout_seconds <= 300
+      && Number.isInteger(c.observation_max_attempts) && c.observation_max_attempts >= 1 && c.observation_max_attempts <= 5
+      && c.device_command_timeout_seconds > 0 && c.device_command_timeout_seconds <= 60
+      && Number.isInteger(c.action_failure_attempts) && c.action_failure_attempts >= 1 && c.action_failure_attempts <= 10
+      && c.worker_cleanup_timeout_seconds > 0 && c.worker_cleanup_timeout_seconds <= 300
+      && c.adb_discovery_interval_seconds >= 1 && c.adb_discovery_interval_seconds <= 300
+      && c.adb_discovery_timeout_seconds > 0 && c.adb_discovery_timeout_seconds <= 30;
+  }
 }

apps/showcase_ui/src/app/components/model-settings/model-settings.component.spec.ts

@@ -1,4 +1,5 @@
-import { TestBed } from '@angular/core/testing';
+import { TestBed, fakeAsync, tick } from '@angular/core/testing';
+import { UI_READ_TIMEOUT_MS } from '../../core/http-timeouts';
 import { provideHttpClient } from '@angular/common/http';
 import { HttpTestingController, provideHttpClientTesting } from '@angular/common/http/testing';
 import { ModelSettingsComponent } from './model-settings.component';
@@ -6,14 +7,29 @@ import { CapabilitySettingsComponent } from './capability-settings.component';
 
 describe('持久化模型与功能设置', () => {
   let http: HttpTestingController;
-  const config = { base_url: 'https://example.test/v1', model: 'mimo-v2.6-flash',
+  const config = { base_url: 'https://example.test/v1', model: 'gpt-6-luna',
-    api_key_configured: true, timeout_seconds: 300, max_retries: 5, retry_interval_seconds: 10, reasoning_effort: 'high', env_path: '/app/.env' };
+    api_key_configured: true, timeout_seconds: 300, max_retries: 5, retry_interval_seconds: 10, reasoning_effort: 'high',
+    rate_limit_retry_seconds: 60, tool_fallback_model: 'gpt-6-luna', tool_fallback_reasoning_effort: 'high',
+    tool_fallback_max_retries: 3, tool_fallback_no_progress_attempts: 3,
+    model_presets: ['gpt-6-luna', 'gpt-6.1-sol', 'gpt-6-astra', 'mimo-v2.6-pro'], env_path: '/app/.env' };
   beforeEach(() => {
     TestBed.configureTestingModule({ providers: [provideHttpClient(), provideHttpClientTesting()] });
     http = TestBed.inject(HttpTestingController);
   });
   afterEach(() => http.verify());
 
+  it('配置请求排队超时后退出加载状态，不会发送保存请求', fakeAsync(() => {
+    const fixture = TestBed.createComponent(ModelSettingsComponent);
+    fixture.detectChanges();
+    const request = http.expectOne('/api/system/model-settings');
+    tick(UI_READ_TIMEOUT_MS);
+    expect(request.cancelled).toBeTrue();
+    expect(fixture.componentInstance.loading()).toBeFalse();
+    expect(fixture.componentInstance.error()).toBeTruthy();
+    expect(fixture.componentInstance.valid()).toBeFalse();
+    http.expectNone('/api/system/model-settings');
+  }));
+
   it('加载配置后保留密钥，保存新模型且不发送空密钥', () => {
     const fixture = TestBed.createComponent(ModelSettingsComponent);
     fixture.detectChanges();
@@ -30,37 +46,40 @@ describe('持久化模型与功能设置', () => {
     expect(request.request.body.timeout_seconds).toBe(300);
     expect(request.request.body.max_retries).toBe(5);
     expect(request.request.body.retry_interval_seconds).toBe(10);
+    expect(request.request.body.rate_limit_retry_seconds).toBe(60);
+    expect(request.request.body.tool_fallback_model).toBe('gpt-6-luna');
+    expect(request.request.body.tool_fallback_no_progress_attempts).toBe(3);
     expect(request.request.body.api_key).toBeUndefined();
     request.flush({ ...config, model: 'another-model' });
     expect(component.busy()).toBeNull();
     expect(component.message()).toContain('保存');
   });
 
-  it('预设按钮填入名称，自定义模型与 max 可以保存并重新加载', async () => {
+  it('统一选择预设或自定义模型，自定义名称与 max 可以保存并重新加载', async () => {
     const fixture = TestBed.createComponent(ModelSettingsComponent);
     fixture.detectChanges();
     http.expectOne('/api/system/model-settings').flush(config);
     fixture.detectChanges();
     await fixture.whenStable();
     const root: HTMLElement = fixture.nativeElement;
-    const presets = root.querySelectorAll<HTMLButtonElement>('.model-presets button');
+    const select = root.querySelector<HTMLSelectElement>('app-model-selector select')!;
-    expect(presets.length).toBe(2);
+    expect(Array.from(select.options, option => option.value)).toEqual([...config.model_presets, '__custom__']);
-    presets[0].click();
+    expect(select.value).toBe('gpt-6-luna');
+    select.value = 'mimo-v2.6-pro';
+    select.dispatchEvent(new Event('change'));
+    expect(fixture.componentInstance.model).toBe('mimo-v2.6-pro');
+    select.value = '__custom__';
+    select.dispatchEvent(new Event('change'));
     fixture.detectChanges();
-    await fixture.whenStable();
+    const input = root.querySelector<HTMLInputElement>('input[aria-label="模型名称（自定义）"]')!;
-    const input = root.querySelector<HTMLInputElement>('#model-name')!;
-    expect(input.value).toBe('gpt-5.6-luna');
-    expect(presets[0].getAttribute('aria-pressed')).toBe('true');
-    expect(input.getAttribute('list')).toBeNull();
     expect(input.autocomplete).toBe('off');
     input.value = 'custom/model-v2';
     input.dispatchEvent(new Event('input'));
-    const reasoning = root.querySelector<HTMLSelectElement>('#model-reasoning')!;
+    const reasoning = root.querySelector<HTMLInputElement>('input[aria-label="思考强度"]')!;
-    reasoning.value = 'max';
+    reasoning.value = '7';
-    reasoning.dispatchEvent(new Event('change'));
+    reasoning.dispatchEvent(new Event('input'));
     fixture.detectChanges();
     await fixture.whenStable();
-    expect(presets[0].getAttribute('aria-pressed')).toBe('false');
     root.querySelector<HTMLFormElement>('form')!.dispatchEvent(new Event('submit', { cancelable: true }));
     const request = http.expectOne('/api/system/model-settings');
     expect(request.request.body.model).toBe('custom/model-v2');
@@ -73,8 +92,36 @@ describe('持久化模型与功能设置', () => {
     http.expectOne('/api/system/model-settings').flush(saved);
     reopened.detectChanges();
     await reopened.whenStable();
-    expect(reopened.nativeElement.querySelector('#model-name').value).toBe('custom/model-v2');
+    expect(reopened.nativeElement.querySelector('input[aria-label="模型名称（自定义）"]').value).toBe('custom/model-v2');
-    expect(reopened.nativeElement.querySelector('#model-reasoning').value).toBe('max');
+    expect(reopened.nativeElement.querySelector('input[aria-label="思考强度"]').getAttribute('aria-valuetext')).toBe('最高 · max');
+  });
+
+  it('高级稳定性设置可编辑并保存', async () => {
+    const fixture = TestBed.createComponent(ModelSettingsComponent);
+    fixture.detectChanges();
+    http.expectOne('/api/system/model-settings').flush(config);
+    fixture.detectChanges();
+    await fixture.whenStable();
+    const root: HTMLElement = fixture.nativeElement;
+    root.querySelector<HTMLDetailsElement>('.recovery-settings')!.open = true;
+    const fallbackSelect = root.querySelectorAll<HTMLSelectElement>('app-model-selector select')[1];
+    expect(Array.from(fallbackSelect.options, option => option.value)).toEqual([...config.model_presets, '__custom__']);
+    fallbackSelect.value = '__custom__';
+    fallbackSelect.dispatchEvent(new Event('change'));
+    fixture.detectChanges();
+    const fallback = root.querySelector<HTMLInputElement>('input[aria-label="兜底模型名称（自定义）"]')!;
+    fallback.value = 'custom/fallback';
+    fallback.dispatchEvent(new Event('input'));
+    const wait = root.querySelector<HTMLInputElement>('input[name="rateLimitRetry"]')!;
+    wait.value = '90';
+    wait.dispatchEvent(new Event('input'));
+    fixture.detectChanges();
+    await fixture.whenStable();
+    root.querySelector<HTMLFormElement>('form')!.dispatchEvent(new Event('submit', { cancelable: true }));
+    const request = http.expectOne('/api/system/model-settings');
+    expect(request.request.body.tool_fallback_model).toBe('custom/fallback');
+    expect(request.request.body.rate_limit_retry_seconds).toBe(90);
+    request.flush({ ...config, tool_fallback_model: 'custom/fallback', rate_limit_retry_seconds: 90 });
   });
 
   it('连接测试使用当前输入而不写入保存接口', () => {
@@ -83,7 +130,7 @@ describe('持久化模型与功能设置', () => {
     http.expectOne('/api/system/model-settings').flush(config);
     fixture.componentInstance.test();
     const request = http.expectOne('/api/system/model-settings/test');
-    expect(request.request.body.model).toBe('mimo-v2.6-flash');
+    expect(request.request.body.model).toBe('gpt-6-luna');
     request.flush({ message: 'Responses 连接测试成功' });
     http.expectNone('/api/system/model-settings');
     expect(fixture.componentInstance.message()).toContain('成功');
@@ -98,21 +145,79 @@ describe('持久化模型与功能设置', () => {
     http.expectNone('/api/system/capability-settings');
   });
 
-  it('保存关闭视频分析的配置，重新打开时恢复关闭状态', () => {
+  it('录屏与视频分析独立保存，重新打开后保留各自状态', () => {
-    const capabilities = { video_enabled: true, outputter_enabled: true, step_summarizer_enabled: false };
+    const capabilities = { adb_auto_connect: true, adb_discovery_interval_seconds: 10, adb_discovery_timeout_seconds: 5, video_enabled: true, recording_enabled: false, outputter_enabled: true, step_summarizer_enabled: false, screenshot_limit_enabled: true, screenshot_max_long_edge: 2048, task_timeout_seconds: 1800, observation_timeout_seconds: 60, observation_max_attempts: 3, device_command_timeout_seconds: 10, action_failure_attempts: 3, worker_cleanup_timeout_seconds: 30 };
     const fixture = TestBed.createComponent(CapabilitySettingsComponent);
     fixture.detectChanges();
     http.expectOne('/api/system/capability-settings').flush(capabilities);
     fixture.componentInstance.config.video_enabled = false;
+    fixture.componentInstance.config.recording_enabled = true;
+    fixture.componentInstance.config.screenshot_limit_enabled = false;
     fixture.componentInstance.save();
     const request = http.expectOne('/api/system/capability-settings');
     expect(request.request.body.video_enabled).toBeFalse();
-    const saved = { ...capabilities, video_enabled: false };
+    expect(request.request.body.recording_enabled).toBeTrue();
+    expect(request.request.body.screenshot_limit_enabled).toBeFalse();
+    const saved = { ...capabilities, video_enabled: false, recording_enabled: true, screenshot_limit_enabled: false };
     request.flush(saved);
     fixture.destroy();
     const reopened = TestBed.createComponent(CapabilitySettingsComponent);
     reopened.detectChanges();
     http.expectOne('/api/system/capability-settings').flush(saved);
     expect(reopened.componentInstance.config.video_enabled).toBeFalse();
+    expect(reopened.componentInstance.config.recording_enabled).toBeTrue();
+    expect(reopened.componentInstance.config.screenshot_limit_enabled).toBeFalse();
+  });
+
+  it('执行保护默认折叠，修改时限可保存且拒绝无效数值', async () => {
+    const capabilities = { adb_auto_connect: true, adb_discovery_interval_seconds: 10, adb_discovery_timeout_seconds: 5, video_enabled: false, recording_enabled: false, outputter_enabled: true, step_summarizer_enabled: true, screenshot_limit_enabled: true, screenshot_max_long_edge: 2048, task_timeout_seconds: 1800, observation_timeout_seconds: 60, observation_max_attempts: 3, device_command_timeout_seconds: 10, action_failure_attempts: 3, worker_cleanup_timeout_seconds: 30 };
+    const fixture = TestBed.createComponent(CapabilitySettingsComponent);
+    fixture.detectChanges();
+    http.expectOne('/api/system/capability-settings').flush(capabilities);
+    fixture.detectChanges();
+    await fixture.whenStable();
+    const root: HTMLElement = fixture.nativeElement;
+    expect(root.querySelector<HTMLDetailsElement>('.execution-settings')!.open).toBeFalse();
+    root.querySelector<HTMLElement>('summary')!.click();
+    const timeout = root.querySelector<HTMLInputElement>('input[type="number"]')!;
+    timeout.value = '3600';
+    timeout.dispatchEvent(new Event('input'));
+    fixture.detectChanges();
+    await fixture.whenStable();
+    root.querySelector<HTMLButtonElement>('button')!.click();
+    const request = http.expectOne('/api/system/capability-settings');
+    expect(request.request.body.task_timeout_seconds).toBe(3600);
+    expect(request.request.body.observation_timeout_seconds).toBe(60);
+    request.flush({ ...capabilities, task_timeout_seconds: 3600 });
+    expect(fixture.componentInstance.message()).toContain('已保存');
+    timeout.value = '0';
+    timeout.dispatchEvent(new Event('input'));
+    fixture.detectChanges();
+    await fixture.whenStable();
+    expect(root.querySelector<HTMLButtonElement>('button')!.disabled).toBeTrue();
+    fixture.componentInstance.save();
+    http.expectNone('/api/system/capability-settings');
+  });
+
+  it('自动连接默认开启，可独立关闭并保存发现预算', () => {
+    const fixture = TestBed.createComponent(CapabilitySettingsComponent);
+    const capabilities = { ...fixture.componentInstance.config };
+    fixture.detectChanges();
+    http.expectOne('/api/system/capability-settings').flush(capabilities);
+    fixture.detectChanges();
+    const root: HTMLElement = fixture.nativeElement;
+    expect(root.textContent).toContain('自动连接设备');
+    expect(fixture.componentInstance.config.adb_auto_connect).toBeTrue();
+    fixture.componentInstance.config.adb_auto_connect = false;
+    fixture.componentInstance.config.adb_discovery_timeout_seconds = 8;
+    fixture.componentInstance.save();
+    const request = http.expectOne('/api/system/capability-settings');
+    expect(request.request.body.adb_auto_connect).toBeFalse();
+    expect(request.request.body.adb_discovery_timeout_seconds).toBe(8);
+    expect(request.request.body.recording_enabled).toBeFalse();
+    request.flush(request.request.body);
+    fixture.componentInstance.config.adb_discovery_timeout_seconds = 0;
+    fixture.componentInstance.save();
+    http.expectNone('/api/system/capability-settings');
   });
 });

apps/showcase_ui/src/app/components/model-settings/model-settings.component.ts

@@ -1,7 +1,11 @@
 import { Component, OnInit, inject, signal, output } from '@angular/core';
 import { FormsModule } from '@angular/forms';
 import { HttpClient } from '@angular/common/http';
-import { finalize } from 'rxjs';
+import { finalize, timeout } from 'rxjs';
+import { UI_READ_TIMEOUT_MS } from '../../core/http-timeouts';
+import { SettingSliderComponent, REASONING_VALUES, REASONING_LABELS } from '../setting-slider/setting-slider.component';
+import { ModelSelectorComponent } from '../model-selector/model-selector.component';
+import { MODEL_PRESETS } from '../../core/model-presets';
 
 interface ModelSettings {
   base_url: string;
@@ -10,7 +14,13 @@ interface ModelSettings {
   timeout_seconds: number;
   max_retries: number;
   retry_interval_seconds: number;
+  rate_limit_retry_seconds: number;
+  tool_fallback_model: string;
+  tool_fallback_reasoning_effort: string;
+  tool_fallback_max_retries: number;
+  tool_fallback_no_progress_attempts: number;
   reasoning_effort: string;
+  model_presets: string[];
   env_path: string;
   message?: string;
 }
@@ -18,33 +28,17 @@ interface ModelSettings {
 @Component({
   selector: 'app-model-settings',
   standalone: true,
-  imports: [FormsModule],
+  imports: [FormsModule, SettingSliderComponent, ModelSelectorComponent],
   template: `
     <form class="model-settings" (ngSubmit)="save()">
       <p>统一使用 OpenAI Responses。保存后新任务立即生效，重启后自动加载。</p>
       <label for="model-base-url">接口地址</label>
       <input id="model-base-url" name="baseUrl" type="url" [(ngModel)]="baseUrl" required
-        placeholder="https://api.openai.com/v1" [disabled]="busy() || loading()" />
+        placeholder="https://api-n-cd.playones.com/v1" [disabled]="busy() || loading()" />
       <small>填写接口根地址，通常包含 /v1，无需添加 /responses。</small>
-      <label for="model-name">模型名称</label>
+      <app-model-selector label="模型名称" [(value)]="model" [presets]="modelPresets" [disabled]="!!busy() || loading()" />
-      <div class="model-presets" role="group" aria-label="预设模型">
+      <small>选择预设模型，或选择“自定义模型”填写名称。</small>
-        @for (preset of modelPresets; track preset) {
+      <app-setting-slider label="思考强度" [(value)]="reasoning" [values]="reasoningValues" [labels]="reasoningLabels" [disabled]="!!busy() || loading()" hint="更高的强度适合复杂推理；需接口支持，也可选择模型默认。" />
-          <button type="button" [attr.aria-pressed]="model === preset" (click)="model = preset"
-            [disabled]="busy() || loading()">{{ preset }}</button>
-        }
-      </div>
-      <input id="model-name" name="llmModelName" type="text" autocomplete="off" autocapitalize="off" spellcheck="false"
-        [(ngModel)]="model" required placeholder="输入接口支持的模型名" aria-describedby="model-name-help"
-        [disabled]="busy() || loading()" />
-      <small id="model-name-help">点击预设填入，或直接输入自定义模型名称；保存后可继续使用。</small>
-      <label for="model-reasoning">思考强度</label>
-      <select id="model-reasoning" name="reasoning" [(ngModel)]="reasoning" [disabled]="busy() || loading()">
-        <option value="default">由模型默认</option><option value="none">关闭（none）</option>
-        <option value="minimal">最低（minimal）</option><option value="low">低（low）</option>
-        <option value="medium">中（medium）</option><option value="high">高（high）</option>
-        <option value="xhigh">极高（xhigh）</option><option value="max">最高（max）</option>
-      </select>
-      <small>接口需支持所选强度；不支持时会明确报错，可手动选择“由模型默认”。</small>
       <label for="model-key">API Key <span>{{ keyConfigured() ? '（已配置，留空保留）' : '（尚未配置）' }}</span></label>
       <input id="model-key" name="apiKey" type="password" [(ngModel)]="apiKey" autocomplete="new-password"
         placeholder="输入新密钥" [disabled]="busy() || loading()" />
@@ -55,6 +49,22 @@ interface ModelSettings {
           [(ngModel)]="retries" [disabled]="busy() || loading()" /></label>
         <label>重试间隔（秒）<input name="retryInterval" type="number" min="0" max="300" required [(ngModel)]="retryInterval" [disabled]="busy() || loading()" /></label>
       </div>
+      <details class="recovery-settings">
+        <summary>稳定性与兜底模型</summary>
+        <div class="recovery-content">
+          <label>429 等待重试（秒）<input name="rateLimitRetry" type="number" min="0" max="3600" required
+            [(ngModel)]="rateLimitRetry" [disabled]="busy() || loading()" /></label>
+          <app-model-selector label="兜底模型名称" [(value)]="toolFallbackModel" [presets]="modelPresets" [disabled]="!!busy() || loading()" />
+          <app-setting-slider label="兜底思考强度" [(value)]="toolFallbackReasoning" [values]="reasoningValues" [labels]="reasoningLabels" [disabled]="!!busy() || loading()" />
+          <div class="number-fields">
+            <label>兜底请求重试次数<input name="toolFallbackRetries" type="number" min="0" max="10" required
+              [(ngModel)]="toolFallbackRetries" [disabled]="busy() || loading()" /></label>
+            <label>兜底无效调用上限<input name="toolFallbackNoProgress" type="number" min="1" max="10" required
+              [(ngModel)]="toolFallbackNoProgress" [disabled]="busy() || loading()" /></label>
+          </div>
+          <small>首选模型连续 3 次未形成有效工具调用后，本步骤使用兜底模型；下一步恢复首选模型。</small>
+        </div>
+      </details>
       <div class="actions">
         <button type="submit" [disabled]="busy() || loading() || !valid()">{{ busy() === 'save' ? '正在保存…' : '保存配置' }}</button>
         <button type="button" (click)="test()" [disabled]="busy() || loading() || !valid()">
@@ -76,43 +86,63 @@ interface ModelSettings {
     input:focus, select:focus { outline: 2px solid #38bdf8; }
     .number-fields { display: grid; grid-template-columns: 1fr 1fr; gap: 16px; margin-top: 8px; }
     .number-fields input { margin-top: 8px; }
+    .recovery-settings { border: 1px solid #cbd5e1; border-radius: 10px; background: #f8fafc; padding: 10px 14px; }
+    .recovery-settings summary { cursor: pointer; font-weight: 600; color: #334155; }
+    .recovery-content { display: grid; gap: 10px; padding-top: 12px; }
+    .recovery-content label input, .recovery-content label select { margin-top: 8px; }
     .actions { display: flex; gap: 12px; margin-top: 8px; }
     button { padding: 11px 20px; border-radius: 10px; border: 1px solid #38bdf877; background: #0284c7; color: white; cursor: pointer; font: inherit; }
     button:disabled { opacity: .5; cursor: default; } .success { color: #047857; } .error { color: #be123c; }
-    .model-presets { display: flex; flex-wrap: wrap; gap: 8px; }
-    .model-presets button { padding: 8px 12px; background: #f8fafc; color: #475569; border-color: #cbd5e1; font-size: 14px; }
-    .model-presets button:hover:not(:disabled) { background: #e0f2fe; border-color: #38bdf8; }
-    .model-presets button[aria-pressed="true"] { background: #e0f2fe; border-color: #0284c7; color: #075985; }
     button:focus-visible { outline: 2px solid #0284c7; outline-offset: 2px; }
     small { overflow-wrap: anywhere; } @media(max-width: 480px) { .number-fields { grid-template-columns: 1fr; } }
+    .model-settings {gap:12px;font-size:13px} label {margin-top:6px}input,select {background:#fff;border-color:#dbe3ec;padding:11px 12px}
+    input:focus,select:focus {outline:2px solid #a5b4fc;outline-offset:2px}
+    .actions {border-top:1px solid #e2e8f0;padding-top:16px;align-items:center}.actions button {min-height:40px;background:#4f46e5;border-color:#4f46e5;font-weight:600;border-radius:12px}.actions button[type=button] {background:#fff;border-color:#dbe3ec;color:#475569}
+    .recovery-settings {border-color:#e2e8f0;border-radius:14px;padding:14px 16px}.recovery-settings summary {font-size:13px}
   `]
 })
 export class ModelSettingsComponent implements OnInit {
+  readonly reasoningValues = REASONING_VALUES;
+  readonly reasoningLabels = REASONING_LABELS;
   private http = inject(HttpClient);
   saved = output<void>();
-  readonly modelPresets = ['gpt-5.6-luna', 'mimo-v2.6-flash'];
+  modelPresets: string[] = MODEL_PRESETS;
   baseUrl = ''; model = ''; apiKey = ''; timeout = 300; retries = 5; retryInterval = 10; reasoning = 'high';
+  rateLimitRetry = 60; toolFallbackModel = 'gpt-6-luna'; toolFallbackReasoning = 'high';
+  toolFallbackRetries = 3; toolFallbackNoProgress = 3;
   busy = signal<'save' | 'test' | null>(null);
   loading = signal(true); keyConfigured = signal(false); envPath = signal('');
   message = signal(''); error = signal('');
 
   ngOnInit(): void {
-    this.http.get<ModelSettings>('/api/system/model-settings').pipe(finalize(() => this.loading.set(false))).subscribe({
+    this.http.get<ModelSettings>('/api/system/model-settings').pipe(timeout(UI_READ_TIMEOUT_MS), finalize(() => this.loading.set(false))).subscribe({
       next: config => this.apply(config),
       error: () => this.error.set('配置加载失败，请刷新页面重试。')
     });
   }
   valid(): boolean {
     return !!this.baseUrl.trim() && !!this.model.trim() && (!!this.apiKey.trim() || this.keyConfigured())
-      && this.timeout > 0 && this.timeout <= 3600 && Number.isInteger(this.retries) && this.retries >= 0 && this.retries <= 10 && this.retryInterval >= 0 && this.retryInterval <= 300;
+      && this.timeout > 0 && this.timeout <= 3600 && Number.isInteger(this.retries) && this.retries >= 0 && this.retries <= 10 && this.retryInterval >= 0 && this.retryInterval <= 300
+      && this.rateLimitRetry >= 0 && this.rateLimitRetry <= 3600 && !!this.toolFallbackModel.trim()
+      && Number.isInteger(this.toolFallbackRetries) && this.toolFallbackRetries >= 0 && this.toolFallbackRetries <= 10
+      && Number.isInteger(this.toolFallbackNoProgress) && this.toolFallbackNoProgress >= 1 && this.toolFallbackNoProgress <= 10;
   }
   private payload() {
     return { base_url: this.baseUrl.trim(), model: this.model.trim(), api_key: this.apiKey.trim() || undefined,
-      timeout_seconds: this.timeout, max_retries: this.retries, retry_interval_seconds: this.retryInterval, reasoning_effort: this.reasoning };
+      timeout_seconds: this.timeout, max_retries: this.retries, retry_interval_seconds: this.retryInterval, reasoning_effort: this.reasoning,
+      rate_limit_retry_seconds: this.rateLimitRetry, tool_fallback_model: this.toolFallbackModel.trim(),
+      tool_fallback_reasoning_effort: this.toolFallbackReasoning, tool_fallback_max_retries: this.toolFallbackRetries,
+      tool_fallback_no_progress_attempts: this.toolFallbackNoProgress };
   }
   private apply(config: ModelSettings): void {
     this.baseUrl = config.base_url; this.model = config.model; this.timeout = config.timeout_seconds;
+    this.modelPresets = config.model_presets;
     this.retries = config.max_retries; this.retryInterval = config.retry_interval_seconds ?? 10; this.reasoning = config.reasoning_effort ?? 'high'; this.keyConfigured.set(config.api_key_configured); this.envPath.set(config.env_path);
+    this.rateLimitRetry = config.rate_limit_retry_seconds ?? 60;
+    this.toolFallbackModel = config.tool_fallback_model ?? 'gpt-6-luna';
+    this.toolFallbackReasoning = config.tool_fallback_reasoning_effort ?? 'high';
+    this.toolFallbackRetries = config.tool_fallback_max_retries ?? 3;
+    this.toolFallbackNoProgress = config.tool_fallback_no_progress_attempts ?? 3;
   }
   save(): void {
     if (this.busy() || !this.valid()) return;

apps/showcase_ui/src/app/components/nav-switcher/nav-switcher.component.scss

@@ -43,6 +43,17 @@
     border-color: rgba(203, 213, 225, 0.9);
   }
 
+  .app-version {
+    padding: 0 10px 0 8px;
+    border-left: 1px solid #e2e8f0;
+    color: #64748b;
+    font-size: 11px;
+    font-weight: 500;
+    line-height: 18px;
+    white-space: nowrap;
+    font-variant-numeric: tabular-nums;
+  }
+
   .nav-tab-btn {
     display: inline-flex;
     align-items: center;

apps/showcase_ui/src/app/components/nav-switcher/nav-switcher.component.ts

@@ -17,6 +17,7 @@
 import { Component, ChangeDetectionStrategy } from '@angular/core';
 
 import { RouterLink, RouterLinkActive } from '@angular/router';
+import { version } from '../../../../package.json';
 
 @Component({
   selector: 'app-nav-switcher',
@@ -43,9 +44,15 @@ import { RouterLink, RouterLinkActive } from '@angular/router';
         <span class="material-symbols-outlined tab-icon">space_dashboard</span>
         <span class="tab-label">工作区</span>
       </a>
+      <a routerLink="/plugins" routerLinkActive="active" class="nav-tab-btn" title="管理扩展插件">
+        <span class="material-symbols-outlined tab-icon">extension</span><span class="tab-label">插件</span>
+      </a>
+      <span class="app-version" [title]="'ARTEMIS 版本 ' + version" [attr.aria-label]="'ARTEMIS 版本 ' + version">v{{version}}</span>
     </nav>
   `,
   changeDetection: ChangeDetectionStrategy.Eager,
   styleUrls: ['./nav-switcher.component.scss']
 })
-export class NavSwitcherComponent {}
+export class NavSwitcherComponent {
+  readonly version = version;
+}

apps/showcase_ui/src/app/components/report-markdown/report-markdown.component.spec.ts

@@ -0,0 +1,36 @@
+import { TestBed } from '@angular/core/testing';
+import { ReportMarkdownComponent } from './report-markdown.component';
+
+describe('报告 Markdown 展示', () => {
+  it('renders report headings, emphasis, lists, tables and code in both preview and full view', () => {
+    const fixture = TestBed.createComponent(ReportMarkdownComponent);
+    fixture.componentRef.setInput('content', '### 双端布局报告\n- **手机**：`/zh`\n\n| 区域 | 结论 |\n| --- | --- |\n| 页底 | 已确认 |\n\n```text\n<原始证据>\n```');
+    fixture.componentRef.setInput('compact', true);
+    fixture.detectChanges();
+    const view: HTMLElement = fixture.nativeElement;
+    expect(view.querySelector('h3')?.textContent).toBe('双端布局报告');
+    expect(view.querySelector('li strong')?.textContent).toBe('手机');
+    expect(view.querySelector('li code')?.textContent).toBe('/zh');
+    expect(view.querySelector('td')?.textContent).toBe('页底');
+    expect(view.querySelector('pre code')?.textContent).toBe('<原始证据>');
+    expect(view.querySelector('.compact')).not.toBeNull();
+    fixture.componentRef.setInput('compact', false);
+    fixture.detectChanges();
+    expect(view.querySelector('.compact')).toBeNull();
+    expect(view.querySelectorAll('td').length).toBe(2);
+    fixture.destroy();
+  });
+
+  it('keeps HTML in model output inert and updates when the selected report changes', () => {
+    const fixture = TestBed.createComponent(ReportMarkdownComponent);
+    fixture.componentRef.setInput('content', '<img src=x onerror=alert(1)>\n<script>alert(1)</script>');
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('img,script')).toBeNull();
+    expect(fixture.nativeElement.textContent).toContain('<img');
+    fixture.componentRef.setInput('content', '## 新报告\n**待确认**');
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('h2').textContent).toBe('新报告');
+    expect(fixture.nativeElement.textContent).not.toContain('alert');
+    fixture.destroy();
+  });
+});

apps/showcase_ui/src/app/components/report-markdown/report-markdown.component.ts

@@ -0,0 +1,39 @@
+import { Component, computed, input } from '@angular/core';
+import { renderMarkdownToHtml } from '../../utils/markdown-parser.util';
+
+/** Shared, sanitized presentation for report prose (not the editable task plan). */
+@Component({
+  selector: 'app-report-markdown',
+  standalone: true,
+  template: `<div class="report-prose" [class.compact]="compact()" [innerHTML]="html()"></div>`,
+  styles: [`
+    :host { display: block; min-width: 0; max-width: 100%; }
+    .report-prose { color: inherit; font-size: 13px; font-weight: 400; line-height: 1.7; white-space: normal; overflow-wrap: anywhere; word-break: normal; }
+    .compact { max-height: 170px; overflow: hidden; }
+    :host ::ng-deep .report-prose {
+      > div, p { margin: 4px 0; }
+      > :first-child { margin-top: 0; }
+      > :last-child { margin-bottom: 0; }
+      h1,h2,h3,h4,h5,h6 { margin: 16px 0 6px; font-weight: 600; line-height: 1.5; color: #0f172a; }
+      h1 { font-size: 18px; } h2 { font-size: 16px; } h3,h4,h5,h6 { font-size: 14px; }
+      strong { font-weight: 600; }
+      ul,ol { margin: 6px 0; padding-left: 24px; }
+      li { margin: 4px 0; }
+      blockquote { margin: 10px 0; padding: 6px 12px; border-left: 3px solid #93c5fd; background: #f8fafc; }
+      code { font-family: Consolas, monospace; font-size: .92em; border-radius: 4px; padding: 1px 4px; background: #f1f5f9; }
+      pre { max-width: 100%; overflow: auto; padding: 10px 12px; border-radius: 6px; background: #f1f5f9; white-space: pre; }
+      pre code { padding: 0; background: none; }
+      .md-table-scroll { max-width: 100%; overflow-x: auto; }
+      table { width: 100%; border-collapse: collapse; font-size: 12px; }
+      th,td { border: 1px solid #cbd5e1; padding: 7px 9px; text-align: left; vertical-align: top; }
+      th { background: #eff6ff; font-weight: 600; }
+      .md-para-gap { height: 4px; margin: 0; }
+      hr { border: 0; border-top: 1px solid #e2e8f0; margin: 12px 0; }
+    }
+  `]
+})
+export class ReportMarkdownComponent {
+  content = input('');
+  compact = input(false);
+  html = computed(() => renderMarkdownToHtml(this.content()));
+}

apps/showcase_ui/src/app/components/setting-slider/setting-slider.component.spec.ts

@@ -0,0 +1,17 @@
+import { TestBed } from '@angular/core/testing';
+import { SettingSliderComponent, REASONING_VALUES, REASONING_LABELS } from './setting-slider.component';
+
+describe('Setting slider', () => {
+  it('maps the selected position to the exact request value and exposes its readable label', () => {
+    const f=TestBed.createComponent(SettingSliderComponent);
+    f.componentRef.setInput('label','思考强度');f.componentRef.setInput('values',REASONING_VALUES);f.componentRef.setInput('labels',REASONING_LABELS);f.componentRef.setInput('value','high');f.detectChanges();
+    const slider=f.nativeElement.querySelector('input');expect(slider.getAttribute('aria-valuetext')).toBe('高 · high');
+    const emit=spyOn(f.componentInstance.valueChange,'emit');slider.value='7';slider.dispatchEvent(new Event('input'));
+    expect(emit).toHaveBeenCalledWith('max');f.destroy();
+  });
+  it('preserves a saved custom time until the slider is actually changed', () => {
+    const f=TestBed.createComponent(SettingSliderComponent);const emit=spyOn(f.componentInstance.valueChange,'emit');
+    f.componentRef.setInput('label','执行时限');f.componentRef.setInput('values',[60,300,1800]);f.componentRef.setInput('value',125);f.detectChanges();
+    expect(f.nativeElement.querySelector('output').textContent).toBe('125 秒');expect(emit).not.toHaveBeenCalled();f.destroy();
+  });
+});

apps/showcase_ui/src/app/components/setting-slider/setting-slider.component.ts

@@ -0,0 +1,46 @@
+import { Component, computed, input, output } from '@angular/core';
+
+@Component({
+  selector: 'app-setting-slider', standalone: true,
+  template: `
+    <div class="slider-heading"><span>{{label()}}</span><output>{{display()}}</output></div>
+    <input type="range" min="0" [max]="values().length - 1" step="1" [value]="index()"
+      [disabled]="disabled()" [attr.aria-label]="label()" [attr.aria-valuetext]="display()"
+      [style.--fill]="(index() / (values().length - 1) * 100) + '%'"
+      (input)="change($any($event.target).value)" />
+    <div class="slider-ends"><span>{{labels()[0] || values()[0]}}</span><span>{{labels().at(-1) || values().at(-1)}}</span></div>
+    @if(hint()){<p>{{hint()}}</p>}
+  `,
+  styles: [`
+    :host{display:block;min-width:0;padding:14px 16px;background:#f8fafc;border:1px solid #e2e8f0;border-radius:14px;font-size:13px;color:#334155}
+    .slider-heading,.slider-ends{display:flex;align-items:center;justify-content:space-between;gap:10px}
+    .slider-heading{font-weight:600}output{color:#4f46e5;font-size:12px;white-space:nowrap}.slider-ends{font-size:11px;color:#64748b}
+    input{appearance:none;display:block;box-sizing:border-box;width:100%;height:6px;border:0;border-radius:10px;margin:20px 0 12px;background:linear-gradient(to right,#6366f1 var(--fill),#dbe3ee var(--fill));cursor:pointer}
+    input::-webkit-slider-thumb{appearance:none;width:18px;height:18px;border-radius:50%;background:#fff;border:2px solid #6366f1;box-shadow:0 2px 6px #33415526}
+    input::-moz-range-thumb{width:14px;height:14px;border-radius:50%;background:#fff;border:2px solid #6366f1}
+    input:focus-visible{outline:2px solid #a5b4fc;outline-offset:5px}input:disabled{opacity:.5;cursor:default}p{margin:10px 0 0;font-size:12px;line-height:1.6;color:#64748b}
+  `]
+})
+export class SettingSliderComponent {
+  label = input.required<string>();
+  value = input.required<string | number>();
+  values = input.required<(string | number)[]>();
+  labels = input<string[]>([]);
+  hint = input('');
+  disabled = input(false);
+  valueChange = output<any>();
+  index = computed(() => {
+    const exact = this.values().indexOf(this.value());
+    if(exact >= 0) return exact;
+    if(typeof this.value() === 'number') return this.values().reduce<number>((best, v, i) => Math.abs(Number(v)-Number(this.value())) < Math.abs(Number(this.values()[best])-Number(this.value())) ? i : best, 0);
+    return 0;
+  });
+  display = computed(() => {
+    const exact = this.values().indexOf(this.value());
+    return exact >= 0 ? this.labels()[exact] || String(this.value()) : String(this.value()) + (typeof this.value() === 'number' ? ' 秒' : '');
+  });
+  change(index: string) { this.valueChange.emit(this.values()[Number(index)]); }
+}
+
+export const REASONING_VALUES = ['default','none','minimal','low','medium','high','xhigh','max'];
+export const REASONING_LABELS = ['模型默认','关闭 · none','最低 · minimal','低 · low','中 · medium','高 · high','极高 · xhigh','最高 · max'];

apps/showcase_ui/src/app/components/task-batch-filter/task-batch-filter.component.spec.ts

@@ -0,0 +1,26 @@
+import { signal } from '@angular/core';
+import { fakeAsync, TestBed, tick } from '@angular/core/testing';
+import { provideHttpClient } from '@angular/common/http';
+import { HttpTestingController, provideHttpClientTesting } from '@angular/common/http/testing';
+import { AgentService } from '../../services/agent.service';
+import { TaskBatchFilterComponent } from './task-batch-filter.component';
+
+describe('队列批次筛选', () => {
+  it('支持全部批量、单独任务及具体批次，可清除筛选', fakeAsync(() => {
+    const agent = {batchFilter:signal('batch-1')};
+    TestBed.configureTestingModule({providers:[provideHttpClient(),provideHttpClientTesting(),{provide:AgentService,useValue:agent}]});
+    const fixture = TestBed.createComponent(TaskBatchFilterComponent);
+    const http = TestBed.inject(HttpTestingController);
+    http.expectOne('/api/automation/batches').flush([{id:'batch-1',name:'第一轮冒烟',created_at:100},{id:'batch-2',name:'第二轮冒烟',created_at:200}]);
+    fixture.detectChanges(); tick(); fixture.detectChanges();
+    const select: HTMLSelectElement = fixture.nativeElement.querySelector('select');
+    expect(select.value).toBe('batch-1'); expect(select.options.length).toBe(6);
+    for(const value of ['__batch__','__single__','__exploration__','batch-2']){
+      select.value=value; select.dispatchEvent(new Event('change')); tick(); fixture.detectChanges();
+      expect(agent.batchFilter()).toBe(value);
+    }
+    fixture.nativeElement.querySelector('button').click(); fixture.detectChanges();
+    expect(agent.batchFilter()).toBe('');
+    fixture.destroy(); http.verify();
+  }));
+});

apps/showcase_ui/src/app/components/task-batch-filter/task-batch-filter.component.ts

@@ -0,0 +1,33 @@
+import { Component, DestroyRef, inject, signal } from '@angular/core';
+import { DatePipe } from '@angular/common';
+import { FormsModule } from '@angular/forms';
+import { HttpClient } from '@angular/common/http';
+import { AgentService } from '../../services/agent.service';
+
+@Component({
+  selector: 'app-task-batch-filter', standalone: true, imports: [FormsModule, DatePipe],
+  template: `<div class="batch-filter"><label>任务范围<select aria-label="按批次筛选任务" [ngModel]="agent.batchFilter()" (ngModelChange)="agent.batchFilter.set($event)">
+    <option value="">全部任务</option><option value="__batch__">全部批量任务</option><option value="__single__">单独任务</option><option value="__exploration__">探索任务</option>
+    @for(batch of batches(); track batch.id){<option [value]="batch.id">{{batch.name}} · {{batch.created_at * 1000 | date:'MM-dd HH:mm'}}</option>}
+    @if(agent.batchFilter() && !['__batch__','__single__','__exploration__'].includes(agent.batchFilter()) && !hasSelectedBatch()){<option [value]="agent.batchFilter()">当前选定批次</option>}
+  </select></label>@if(agent.batchFilter()){<button type="button" (click)="agent.batchFilter.set('')">清除筛选</button>}</div>
+  @if(error()){<div class="filter-error" role="alert">{{error()}} <button type="button" (click)="load()">重试</button></div>}`,
+  styles: [`:host{display:block;margin:0 0 12px}.batch-filter{display:flex;align-items:flex-end;gap:8px;color:#475569;font-size:12px}label{flex:1;min-width:0}select{display:block;box-sizing:border-box;width:100%;height:34px;margin-top:5px;border:1px solid #cbd5e1;border-radius:7px;padding:0 9px;background:#fff;color:#1e293b;font:inherit}option{background:#fff;color:#1e293b}button{height:34px;white-space:nowrap;border:1px solid #cbd5e1;border-radius:7px;padding:0 9px;background:#fff;color:#475569;font:inherit;cursor:pointer}select:focus-visible,button:focus-visible{outline:2px solid #3b82f6;outline-offset:2px}.filter-error{font-size:12px;color:#b91c1c;margin-top:5px}.filter-error button{height:auto;padding:2px 6px}`]
+})
+export class TaskBatchFilterComponent {
+  agent = inject(AgentService); private http = inject(HttpClient);
+  batches = signal<{id: string; name: string; created_at: number}[]>([]); error = signal('');
+  private destroyed = false;
+  constructor() {
+    this.load();
+    const timer = setInterval(() => this.load(), 15000);
+    inject(DestroyRef).onDestroy(() => { this.destroyed = true; clearInterval(timer); });
+  }
+  hasSelectedBatch() { return this.batches().some(batch => batch.id === this.agent.batchFilter()); }
+  load() {
+    this.http.get<{id: string; name: string; created_at: number}[]>('/api/automation/batches').subscribe({
+      next: batches => { if (!this.destroyed) { this.batches.set(batches); this.error.set(''); } },
+      error: () => { if (!this.destroyed) this.error.set('批次列表加载失败'); }
+    });
+  }
+}

apps/showcase_ui/src/app/components/task-history/task-history.component.spec.ts

@@ -1,19 +1,127 @@
 import { signal } from '@angular/core';
-import { TestBed } from '@angular/core/testing';
+import { of, throwError } from 'rxjs';
+import { Clipboard } from '@angular/cdk/clipboard';
+import { fakeAsync, TestBed, tick } from '@angular/core/testing';
 import { provideHttpClient } from '@angular/common/http';
 import { HttpTestingController, provideHttpClientTesting } from '@angular/common/http/testing';
 import { AgentService } from '../../services/agent.service';
 import { TaskHistoryComponent } from './task-history.component';
+import { FavoritesService } from '../../services/favorites.service';
 
 describe('历史任务检索', () => {
   let http: HttpTestingController;
   beforeEach(() => {
     TestBed.configureTestingModule({providers: [provideHttpClient(), provideHttpClientTesting(),
-      {provide: AgentService, useValue: {currentSessionId: signal(null), selectSession: jasmine.createSpy()}}]});
+      {provide: AgentService, useValue: {currentSessionId: signal(null), batchFilter: signal(''), resultRevision: signal(0), selectSession: jasmine.createSpy(), deleteSession: jasmine.createSpy().and.returnValue(of({})), canDeleteSession: AgentService.prototype.canDeleteSession, deletionPrompt: (_id: string, kind: string) => kind === 'exploration' ? '删除探索及全部探索记录，复跑与批次保留' : '删除历史任务'}}]});
     http = TestBed.inject(HttpTestingController);
+    TestBed.inject(FavoritesService).loaded.set(true);
   });
   afterEach(() => http.verify());
 
+  it('已终止探索可删除，确认范围后刷新，取消不发请求', () => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    const row = {session_id:'stopped', status:'cancelled', exploration_kind:'exploration', exploration_status:'paused', start_time:100};
+    http.expectOne(r => r.url === '/api/automation/history').flush({total:1,items:[row]});
+    fixture.detectChanges();
+    const button = fixture.nativeElement.querySelector('.delete-history');
+    expect(button).not.toBeNull();
+    const confirm = spyOn(window, 'confirm').and.returnValue(false);
+    button.click();
+    expect(TestBed.inject(AgentService).deleteSession).not.toHaveBeenCalled();
+    expect(confirm.calls.mostRecent().args[0]).toContain('复跑与批次保留');
+    confirm.and.returnValue(true);
+    button.click();
+    expect(TestBed.inject(AgentService).deleteSession).toHaveBeenCalledOnceWith('stopped');
+    http.expectOne(r => r.url === '/api/automation/history').flush({total:0,items:[]});
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.delete-history')).toBeNull();
+    fixture.destroy();
+  });
+
+  it('服务端拒绝删除时保留记录并显示具体原因', () => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    http.expectOne(r => r.url === '/api/automation/history').flush({total:1,items:[
+      {session_id:'stopped',status:'cancelled',exploration_kind:'exploration',start_time:100}
+    ]});
+    spyOn(window, 'confirm').and.returnValue(true);
+    (TestBed.inject(AgentService).deleteSession as jasmine.Spy).and.returnValue(throwError(() => ({error:{detail:'探索正在分析，请稍后删除'}})));
+    fixture.detectChanges();
+    fixture.nativeElement.querySelector('.delete-history').click();
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('[role="alert"]').textContent).toContain('探索正在分析');
+    expect(fixture.nativeElement.querySelector('.delete-history')).not.toBeNull();
+    fixture.destroy();
+  });
+
+  it('完成任务分别显示探索、Flash、Pro，模式不受通过或失败颜色影响', () => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    http.expectOne(r => r.url === '/api/automation/history').flush({total:3,items:[
+      {session_id:'e',status:'completed',start_time:100,profile:'pro',exploration_kind:'exploration'},
+      {session_id:'f',status:'failed',start_time:100,profile:'flash',exploration_kind:'validation',case_version:2},
+      {session_id:'p',status:'completed',start_time:100,profile:'pro'}
+    ]});
+    fixture.detectChanges();
+    const badges = fixture.nativeElement.querySelectorAll('.mode-badge');
+    expect(badges.length).toBe(3);
+    expect(badges[0].textContent).toContain('探索');
+    expect(badges[1].textContent).toContain('Flash');
+    expect(badges[2].textContent).toContain('Pro');
+    expect(fixture.nativeElement.textContent).toContain('用例复跑 · v2');
+    fixture.destroy();
+  });
+
+  it('更换批次时保留搜索条件并回到第一页，复验后立即刷新', () => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    http.expectOne(r => r.url === '/api/automation/history').flush({total:0,items:[]});
+    fixture.detectChanges();
+    fixture.componentInstance.query = '设置'; fixture.componentInstance.page = 3;
+    const agent = TestBed.inject(AgentService);
+    agent.batchFilter.set('batch-2'); fixture.detectChanges();
+    const filtered = http.expectOne(r => r.params.get('batch_id') === 'batch-2');
+    expect(filtered.request.params.get('page')).toBe('1');
+    expect(filtered.request.params.get('q')).toBe('设置');
+    filtered.flush({total:0,items:[]});
+    agent.resultRevision.update(value => value + 1); fixture.detectChanges();
+    http.expectOne(r => r.params.get('batch_id') === 'batch-2').flush({total:0,items:[]});
+    fixture.destroy();
+  });
+
+  it('执行完成的任务仍能显示测试不通过，旧任务不会自动显示通过', () => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    http.expectOne(r => r.url === '/api/automation/history').flush({total: 3, items: [
+      {session_id:'failed-test', name:'预期校验', status:'completed', start_time:100, test_result:{status:'failed', reason:'预期出现成功，实际出现错误', evidence:['步骤 3 显示错误']}},
+      {session_id:'passed-test', name:'已验收', status:'completed', start_time:100, test_result:{status:'passed', reason:'实际与预期一致', evidence:['步骤 3 显示成功']}},
+      {session_id:'old-task', name:'旧任务', status:'completed', start_time:100}
+    ]});
+    fixture.detectChanges();
+    const rows = fixture.nativeElement.querySelectorAll('button.row');
+    expect(rows[0].textContent).toContain('执行完成');
+    expect(rows[0].querySelector('app-test-result-badge').textContent).toContain('测试不通过');
+    expect(rows[0].querySelector('.test-result-badge').title).toContain('实际出现错误');
+    expect(rows[1].querySelector('app-test-result-badge').textContent).toContain('测试通过');
+    expect(rows[2].querySelector('app-test-result-badge').textContent).toContain('未验证');
+    fixture.destroy();
+  });
+
+  it('历史 ID 独立于打开任务按钮，点击复制不会打开任务', async () => {
+    spyOn(navigator.clipboard, 'writeText').and.rejectWith(new Error('blocked'));
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    http.expectOne(r => r.url === '/api/automation/history').flush({total: 1, items: [
+      {session_id:'completed-task', name:'已完成任务', status:'completed', start_time:Date.now()/1000}
+    ]});
+    const copy = spyOn(TestBed.inject(Clipboard), 'copy').and.returnValue(true);
+    fixture.detectChanges();
+    const id: HTMLElement = fixture.nativeElement.querySelector('app-task-id-copy [role="button"]');
+    expect(id.closest('button.row')).toBeNull();
+    id.click();
+    await Promise.resolve();
+    expect(copy).toHaveBeenCalledOnceWith('completed-task');
+    expect(TestBed.inject(AgentService).selectSession).not.toHaveBeenCalled();
+    fixture.nativeElement.querySelector('button.row').click();
+    expect(TestBed.inject(AgentService).selectSession).toHaveBeenCalledWith('completed-task', true);
+    fixture.destroy();
+  });
+
   it('迟到的旧查询不会覆盖最新搜索结果', () => {
     const fixture = TestBed.createComponent(TaskHistoryComponent);
     const first = http.expectOne(r => r.url === '/api/automation/history');
@@ -29,6 +137,142 @@ describe('历史任务检索', () => {
     fixture.destroy();
   });
 
+  it('定时刷新不插入加载行或禁用翻页，并保留滚动、焦点、折叠和选中项', fakeAsync(() => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    const now = Date.now()/1000;
+    const items = Array.from({length:8}, (_,i) => ({session_id:`task-${i}`, name:`历史用例 ${i}`, status:'completed', start_time:now}));
+    items.push({session_id:'older', name:'前一天的用例', status:'completed', start_time:now-86400});
+    TestBed.inject(AgentService).currentSessionId.set('task-0');
+    http.expectOne(r => r.url === '/api/automation/history').flush({total:60,items});
+    fixture.detectChanges();
+    tick();
+    const host: HTMLElement = fixture.nativeElement;
+    host.style.cssText = 'display:block;width:360px;height:400px;overflow:auto';
+    const group = host.querySelectorAll('details')[1];
+    group.open = false;
+    group.dispatchEvent(new Event('toggle'));
+    const search = host.querySelector('input')!;
+    search.focus({preventScroll:true});
+    host.scrollTop = 100;
+    const row = host.querySelector('.row.selected')!;
+    const before = {height:host.scrollHeight, scroll:host.scrollTop, top:row.getBoundingClientRect().top};
+    expect(before.scroll).toBeGreaterThan(0);
+
+    tick(15000);
+    const refresh = http.expectOne(r => r.url === '/api/automation/history');
+    fixture.detectChanges();
+    expect(host.textContent).not.toContain('正在查询');
+    expect(host.querySelectorAll<HTMLButtonElement>('.pager button')[1].disabled).toBeFalse();
+    expect(host.scrollHeight).toBe(before.height);
+    expect(host.scrollTop).toBe(before.scroll);
+    expect(row.getBoundingClientRect().top).toBe(before.top);
+    refresh.flush({total:60,items:items.map(item => ({...item,name:item.name.replace('历史','更新')}))});
+    fixture.detectChanges();
+    expect(host.querySelector('.row.selected')).toBe(row);
+    expect(row.textContent).toContain('更新用例 0');
+    expect(host.querySelectorAll('details')[1]).toBe(group);
+    expect(group.open).toBeFalse();
+    expect(document.activeElement).toBe(search);
+    expect(host.scrollTop).toBe(before.scroll);
+    fixture.destroy();
+  }));
+
+  it('空历史在后台刷新期间保持空状态，收到新任务后正常更新', fakeAsync(() => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    http.expectOne(r => r.url === '/api/automation/history').flush({total:0,items:[]});
+    fixture.detectChanges();
+    const empty = fixture.nativeElement.querySelector('p');
+    tick(15000);
+    const refresh = http.expectOne(r => r.url === '/api/automation/history');
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('p')).toBe(empty);
+    expect(fixture.nativeElement.textContent).not.toContain('正在查询');
+    refresh.flush({total:1,items:[{session_id:'new',name:'新完成任务',status:'completed',start_time:Date.now()/1000}]});
+    fixture.detectChanges();
+    expect(fixture.nativeElement.textContent).toContain('新完成任务');
+    expect(fixture.nativeElement.textContent).not.toContain('没有符合条件');
+    fixture.destroy();
+  }));
+
+  it('慢刷新不叠加请求，搜索优先且旧刷新不能覆盖搜索结果', fakeAsync(() => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    const first = http.expectOne(r => r.url === '/api/automation/history');
+    tick(15000);
+    http.expectNone(r => r.url === '/api/automation/history');
+    first.flush({total:0,items:[]});
+    tick(15000);
+    const refresh = http.expectOne(r => r.url === '/api/automation/history');
+    tick(15000);
+    http.expectNone(r => r.url === '/api/automation/history');
+    fixture.componentInstance.query = '设置';
+    fixture.componentInstance.search();
+    tick(250);
+    const search = http.expectOne(r => r.params.get('q') === '设置');
+    search.flush({total:1,items:[{session_id:'settings',name:'设置',status:'completed',start_time:100}]});
+    refresh.flush({total:99,items:[]});
+    expect(fixture.componentInstance.total()).toBe(1);
+    expect(fixture.componentInstance.groups()[0].rows[0].session_id).toBe('settings');
+    fixture.destroy();
+  }));
+
+  it('轮询和结果通知不抢先提交仍在输入的搜索条件', fakeAsync(() => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    http.expectOne(r => r.url === '/api/automation/history').flush({total:0,items:[]});
+    fixture.detectChanges();
+    tick(14900);
+    fixture.componentInstance.query = '新条件';
+    fixture.componentInstance.search();
+    TestBed.inject(AgentService).resultRevision.update(value => value+1);
+    fixture.detectChanges();
+    tick(100);
+    http.expectNone(r => r.url === '/api/automation/history');
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('h3 [role="status"]').textContent).toContain('正在查询');
+    tick(150);
+    http.expectOne(r => r.params.get('q') === '新条件').flush({total:0,items:[]});
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('h3 [role="status"]').textContent).toBe('');
+    fixture.destroy();
+  }));
+
+  it('刷新失败保留列表和错误提示，重试期间不反复移除提示', fakeAsync(() => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    const items = [{session_id:'kept',name:'已有历史',status:'completed',start_time:100}];
+    http.expectOne(r => r.url === '/api/automation/history').flush({total:1,items});
+    fixture.detectChanges();
+    const row = fixture.nativeElement.querySelector('.row');
+    tick(15000);
+    http.expectOne(r => r.url === '/api/automation/history').flush({}, {status:503,statusText:'Unavailable'});
+    fixture.detectChanges();
+    const alert = fixture.nativeElement.querySelector('[role="alert"]');
+    expect(alert.textContent).toContain('历史查询失败');
+    tick(15000);
+    const retry = http.expectOne(r => r.url === '/api/automation/history');
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('[role="alert"]')).toBe(alert);
+    expect(fixture.nativeElement.querySelector('.row')).toBe(row);
+    retry.flush({total:1,items});
+    fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('[role="alert"]')).toBeNull();
+    expect(fixture.nativeElement.querySelector('.row')).toBe(row);
+    fixture.destroy();
+  }));
+
+  it('显示明确的结果、完成时间与总耗时，标题悬停显示完整提示词', () => {
+    const fixture = TestBed.createComponent(TaskHistoryComponent);
+    http.expectOne(r => r.url === '/api/automation/history').flush({total: 1, items: [
+      {session_id:'case-1', name:'1.1 打开并核对设置', initial_goal:'请打开设置首页并核对两个真实分类', status:'failed', start_time:100, end_time:165, error:'任务执行超时（60 秒）'}
+    ]});
+    fixture.detectChanges();
+    const row: HTMLButtonElement = fixture.nativeElement.querySelector('button.row');
+    expect(row.getAttribute('title')).toBe('请打开设置首页并核对两个真实分类');
+    expect(row.textContent).toContain('1.1 打开并核对设置');
+    expect(row.textContent).toContain('总耗时 1分5秒');
+    expect(row.textContent).toContain('任务执行超时');
+    expect(fixture.nativeElement.querySelector('.status-failed .history-status').textContent).toContain('失败');
+    fixture.destroy();
+  });
+
   it('日期筛选用本地零点到次日零点，后端分页参数独立传递', () => {
     const fixture = TestBed.createComponent(TaskHistoryComponent);
     http.expectOne(r => r.url === '/api/automation/history').flush({total:0,items:[]});

apps/showcase_ui/src/app/components/task-history/task-history.component.ts

@@ -1,44 +1,103 @@
-import { Component, DestroyRef, inject, signal } from '@angular/core';
+import { TaskIdCopyComponent } from '../task-id-copy/task-id-copy.component';
+import { FavoriteButtonComponent } from '../favorite-button/favorite-button.component';
+import { TaskModeBadgeComponent } from '../task-mode-badge/task-mode-badge.component';
+import { TestResultBadgeComponent } from '../test-result-badge/test-result-badge.component';
+import { Component, DestroyRef, effect, inject, signal, untracked } from '@angular/core';
 import { CommonModule } from '@angular/common';
 import { FormsModule } from '@angular/forms';
 import { HttpClient, HttpParams } from '@angular/common/http';
 import { AgentService } from '../../services/agent.service';
 
 @Component({
-  selector:'app-task-history', standalone:true, imports:[CommonModule,FormsModule],
+  selector:'app-task-history', standalone:true, imports:[CommonModule,FormsModule,TaskIdCopyComponent,TestResultBadgeComponent,TaskModeBadgeComponent,FavoriteButtonComponent],
-  template:`<section aria-label="历史任务"><h3>历史任务 <small>{{total()}} 条</small></h3>
+  template:`<section aria-label="历史任务"><h3>历史任务 <small>{{total()}} 条</small><small class="query-status" role="status">{{loading() ? '正在查询…' : ''}}</small></h3>
     <input aria-label="搜索历史任务" placeholder="搜索名称、目标、任务编号或标签" [(ngModel)]="query" (ngModelChange)="search()">
     <div class="filters"><label>起始日期<input type="date" [(ngModel)]="start" (ngModelChange)="search()"></label><label>结束日期<input type="date" [(ngModel)]="end" (ngModelChange)="search()"></label><label>状态<select [(ngModel)]="status" (ngModelChange)="search()"><option value="">全部状态</option><option value="completed">执行完成</option><option value="failed">失败</option><option value="cancelled">已取消</option><option value="interrupted">中断</option><option value="skipped">跳过</option></select></label></div>
     @if(error()){<p role="alert">{{error()}}</p>}
-    @if(loading()){<small>正在查询…</small>}
     @for(group of groups(); track group.key){<details [open]="!collapsed.has(group.key)" (toggle)="toggle(group.key,$event)"><summary>{{group.title}} <small>本页 {{group.rows.length}} 条</small></summary>
-      @for(row of group.rows; track row.session_id){<div class="history-item"><button class="row" [class.selected]="agent.currentSessionId() === row.session_id" (click)="agent.selectSession(row.session_id,true)"><strong>{{row.name || row.initial_goal}}</strong><span>{{label(row.status)}} · {{row.start_time*1000 | date:'HH:mm:ss'}} · {{row.tags}}</span><small>{{row.session_id}}</small></button><button class="delete-history" (click)="remove(row.session_id)" aria-label="删除此历史任务"><span class="material-symbols-outlined">delete</span></button></div>}
+      @for(row of group.rows; track row.session_id){<div class="history-item" [class]="'history-item status-'+row.status"><div class="history-heading"><app-favorite-button [sourceId]="row.session_id" /><button class="row" [class.selected]="agent.currentSessionId() === row.session_id" (click)="agent.selectSession(row.session_id,true)" [title]="row.initial_goal || ''"><strong>{{row.name || row.initial_goal}}</strong><span class="history-meta"><app-task-mode-badge [session]="row" />@if(row.exploration_kind === "validation"){<span>用例复跑 · v{{row.case_version}}</span>}<b class="history-status">{{label(row.status)}}</b><app-test-result-badge [result]="row.test_result" [executionStatus]="row.status" [compact]="true" /> <span>{{row.end_time ? '完成于 '+(row.end_time*1000 | date:'MM-dd HH:mm:ss') : '开始于 '+(row.start_time*1000 | date:'MM-dd HH:mm:ss')}}</span><span>总耗时 {{duration(row.start_time,row.end_time)}}</span></span>@if(row.error){<span class="history-error" [title]="row.error">{{row.error}}</span>}</button></div><app-task-id-copy [taskId]="row.session_id" />@if(agent.canDeleteSession(row)){<button class="delete-history" (click)="remove(row.session_id, row.exploration_kind)" aria-label="删除此历史任务"><span class="material-symbols-outlined">delete</span></button>}</div>}
     </details>}
     @if(!loading() && !total()){<p>没有符合条件的历史任务。</p>}
     <div class="pager"><button [disabled]="page === 1 || loading()" (click)="page=page-1;load()">上一页</button><span>{{page}} / {{pages()}}</span><button [disabled]="page >= pages() || loading()" (click)="page=page+1;load()">下一页</button></div>
   </section>`,
-  styles:[`:host{display:block;font-size:12px;color:inherit}h3{font-size:14px;margin:14px 0 8px}small{opacity:.6}input,select,button{font:inherit;color:inherit;border:1px solid #64748b55;border-radius:6px;padding:8px;background:transparent;box-sizing:border-box}input{width:100%}select option{background:white;color:#1e293b}.filters{display:flex;gap:6px;flex-wrap:wrap;margin:8px 0}.filters label{flex:1;min-width:110px}.filters input,.filters select{display:block;width:100%}summary{padding:10px 0;cursor:pointer}.row{display:flex;flex-direction:column;gap:4px;width:100%;text-align:left;cursor:pointer;margin:5px 0;overflow-wrap:anywhere}.row strong{font-weight:500}.row span{opacity:.75}.row.selected{border-color:#60a5fa;background:#3b82f615}.history-item{position:relative}.history-item .row{padding-right:35px}.delete-history{position:absolute;right:5px;top:8px;border:0;padding:3px;color:#64748b;cursor:pointer;font-size:12px}.delete-history .material-symbols-outlined{font-size:16px}.delete-history:hover{color:#b91c1c}.pager{display:flex;align-items:center;justify-content:space-between;gap:8px;margin:12px 0}button:disabled{opacity:.4}p{opacity:.7}`]
+  styles:[`
+    :host{display:block;font-size:12px;color:inherit}
+    h3{display:flex;align-items:center;gap:6px;font-size:14px;margin:14px 0 8px}
+    .query-status{margin-left:auto;font-size:11px;font-weight:400;white-space:nowrap}
+    small{opacity:.6}
+    input,select,button{font:inherit;color:inherit;border:1px solid #64748b55;border-radius:6px;padding:8px;background:transparent;box-sizing:border-box}
+    input{width:100%}
+    select option{background:white;color:#1e293b}
+    .filters{display:flex;gap:6px;flex-wrap:wrap;margin:8px 0}
+    .filters label{flex:1;min-width:110px}
+    .filters input,.filters select{display:block;width:100%}
+    summary{padding:10px 0;cursor:pointer}
+    .row{display:flex;flex-direction:column;gap:6px;width:100%;text-align:left;cursor:pointer;margin:5px 0;overflow-wrap:anywhere}
+    .row strong{font-weight:600;display:flex;align-items:center;min-height:28px;padding-left:33px}
+    .row.selected{border-color:#60a5fa;background:#3b82f615}
+    .history-item{position:relative;border:1px solid #cbd5e1;border-left:3px solid #94a3b8;border-radius:8px;margin:6px 0;padding:8px;background:#fff}
+    .history-item.status-completed,.history-item.status-success{border-left-color:#16a34a}
+    .history-item.status-failed{border-left-color:#dc2626;background:#fffafa}
+    .history-item.status-cancelled,.history-item.status-interrupted{border-left-color:#d97706;background:#fffdf8}
+    .history-item.status-skipped{border-left-color:#64748b}
+    .history-item:has(.row.selected){border-color:#60a5fa;background:#eff6ff}
+    .history-item .row{border:0;padding:0 35px 0 0;margin:0 0 4px}
+    .history-item .row.selected{background:transparent}
+    .history-heading{position:relative;min-width:0}
+    .history-heading>app-favorite-button{position:absolute;left:0;top:0;z-index:1}
+    .history-heading>.row{min-width:0;width:100%}
+    .history-meta{display:flex;gap:5px;align-items:center;flex-wrap:wrap;color:#475569;font-size:11px}
+    .history-status{display:inline-flex;align-items:center;box-sizing:border-box;height:20px;line-height:14px;border-radius:5px;padding:0 6px;background:#f1f5f9;color:#334155}
+    .status-completed .history-status,.status-success .history-status{background:#dcfce7;color:#166534}
+    .status-failed .history-status{background:#fee2e2;color:#991b1b}
+    .status-cancelled .history-status,.status-interrupted .history-status{background:#fef3c7;color:#92400e}
+    .history-error{font-size:11px;color:#b91c1c;line-height:1.4}
+    .delete-history{position:absolute;right:5px;top:8px;border:0;padding:3px;color:#64748b;cursor:pointer;font-size:12px}
+    .delete-history .material-symbols-outlined{font-size:16px}
+    .delete-history:hover{color:#b91c1c}
+    .pager{display:flex;align-items:center;justify-content:space-between;gap:8px;margin:12px 0}
+    button:disabled{opacity:.4}
+    p{opacity:.7}
+  `]
 })
 export class TaskHistoryComponent {
   agent=inject(AgentService); private http=inject(HttpClient); query='';start='';end='';status='';page=1;
   total=signal(0);groups=signal<any[]>([]);loading=signal(false);error=signal('');collapsed=new Set<string>();
-  private revision=0;private debounce: ReturnType<typeof setTimeout>|undefined;
+  private revision=0;private pendingRequest=0;private debounce: ReturnType<typeof setTimeout>|undefined;
-  constructor(){this.load();const timer=setInterval(()=>this.load(),15000);inject(DestroyRef).onDestroy(()=>{this.revision++;clearInterval(timer);clearTimeout(this.debounce);});}
+  constructor(){
-  remove(id:string){if(!confirm("确定删除此历史任务？此操作无法撤销。"))return;this.agent.deleteSession(id).subscribe({next:()=>this.load(),error:()=>this.error.set("删除失败，请稍后重试")});}
+    let filter=this.agent.batchFilter(), resultRevision=this.agent.resultRevision();
+    this.load();
+    effect(()=>{
+      const nextFilter=this.agent.batchFilter(), nextRevision=this.agent.resultRevision();
+      if(nextFilter!==filter || nextRevision!==resultRevision){
+        const filterChanged=nextFilter!==filter;
+        if(filterChanged)this.page=1;
+        filter=nextFilter;resultRevision=nextRevision;untracked(()=>this.load(!filterChanged));
+      }
+    });
+    const timer=setInterval(()=>this.load(true),15000);inject(DestroyRef).onDestroy(()=>{this.revision++;clearInterval(timer);clearTimeout(this.debounce);});
+  }
+  remove(id:string, kind?:string){if(!confirm(this.agent.deletionPrompt(id,kind)))return;this.agent.deleteSession(id).subscribe({next:()=>this.load(),error:e=>this.error.set(e.error?.detail || "删除失败，请稍后重试")});}
   pages(){return Math.max(1,Math.ceil(this.total()/30));}
-  label(s:string){return ({completed:'执行完成',success:'执行完成',failed:'失败',cancelled:'已取消',interrupted:'中断',skipped:'跳过'} as Record<string,string>)[s] || s;}
+  label(s:string){return ({completed:'执行完成',success:'执行完成',failed:'失败',cancelled:'已终止',interrupted:'中断',skipped:'跳过'} as Record<string,string>)[s] || s;}
+  duration(start:number,end?:number){if(!start||!end)return '—';const seconds=Math.max(0,Math.floor(end-start));const hours=Math.floor(seconds/3600);const minutes=Math.floor((seconds%3600)/60);return hours?`${hours}时${minutes}分`:minutes?`${minutes}分${seconds%60}秒`:`${seconds}秒`;}
   toggle(key:string,event:Event){if((event.target as HTMLDetailsElement).open)this.collapsed.delete(key);else this.collapsed.add(key);}
-  search(){this.page=1;this.revision++;clearTimeout(this.debounce);this.debounce=setTimeout(()=>this.load(),250);}
+  search(){this.page=1;this.revision++;this.loading.set(true);clearTimeout(this.debounce);this.debounce=setTimeout(()=>this.load(),250);}
-  load(){
+  load(background=false){
-    const revision=++this.revision;this.loading.set(true);this.error.set('');
+    // Polling must not displace rows, toggle pagination, or supersede a user's query.
+    if(background && (this.pendingRequest || this.debounce!==undefined))return;
+    if(!background){clearTimeout(this.debounce);this.debounce=undefined;this.loading.set(true);}
+    const revision=++this.revision;this.pendingRequest=revision;
     let params=new HttpParams().set('q',this.query).set('page',this.page).set('status',this.status);
+    if(this.agent.batchFilter())params=params.set('batch_id',this.agent.batchFilter());
     if(this.start)params=params.set('start',new Date(this.start+'T00:00:00').getTime()/1000);
     if(this.end){const d=new Date(this.end+'T00:00:00');d.setDate(d.getDate()+1);params=params.set('end',d.getTime()/1000);}
     this.http.get<any>('/api/automation/history',{params}).subscribe({next:r=>{
       if(revision!==this.revision)return;
+      this.pendingRequest=0;this.error.set('');
       this.total.set(r.total);const map=new Map<string,any>();const today=new Date();const yesterday=new Date();yesterday.setDate(today.getDate()-1);
       for(const row of r.items){const d=new Date(row.start_time*1000);const key=d.toDateString();if(!map.has(key))map.set(key,{key,title:key===today.toDateString()?'今天':key===yesterday.toDateString()?'昨天':d.toLocaleDateString('zh-CN'),rows:[]});map.get(key).rows.push(row);}
       this.groups.set([...map.values()]);this.loading.set(false);
-    },error:e=>{if(revision===this.revision){this.loading.set(false);this.error.set(e.error?.detail || '历史查询失败，请稍后重试');}}});
+    },error:e=>{if(revision===this.revision){this.pendingRequest=0;this.loading.set(false);this.error.set(e.error?.detail || '历史查询失败，请稍后重试');}}});
   }
 }

apps/showcase_ui/src/app/components/task-id-copy/task-id-copy.component.spec.ts

@@ -0,0 +1,83 @@
+import { Clipboard } from '@angular/cdk/clipboard';
+import { TestBed } from '@angular/core/testing';
+import { TaskIdCopyComponent } from './task-id-copy.component';
+
+describe('任务 ID 复制', () => {
+  let clipboard: jasmine.SpyObj<Clipboard>;
+  beforeEach(() => {
+    spyOn(navigator.clipboard, 'writeText').and.rejectWith(new Error('blocked'));
+    clipboard = jasmine.createSpyObj('Clipboard', ['copy']);
+    TestBed.configureTestingModule({ providers: [{ provide: Clipboard, useValue: clipboard }] });
+  });
+  afterEach(() => window.getSelection()?.removeAllRanges());
+
+  function setup() {
+    const fixture = TestBed.createComponent(TaskIdCopyComponent);
+    fixture.componentRef.setInput('taskId', 'task-123');
+    fixture.detectChanges();
+    const target: HTMLElement = fixture.nativeElement.querySelector('[role="button"]');
+    return { fixture, target };
+  }
+
+  it('点击复制完整 ID，并提示成功，不触发父卡片', async () => {
+    clipboard.copy.and.returnValue(true);
+    const { fixture, target } = setup();
+    const parentClick = jasmine.createSpy('parentClick');
+    fixture.nativeElement.addEventListener('click', parentClick);
+    target.click();
+    await Promise.resolve();
+    fixture.detectChanges();
+    expect(clipboard.copy).toHaveBeenCalledOnceWith('task-123');
+    expect(fixture.nativeElement.querySelector('[role="status"]').textContent).toContain('已复制');
+    expect(parentClick).not.toHaveBeenCalled();
+    fixture.destroy();
+  });
+
+  it('优先使用浏览器剪贴板，并在写入完成后提示成功', async () => {
+    const write = navigator.clipboard.writeText as jasmine.Spy;
+    let complete!: () => void;
+    write.and.returnValue(new Promise<void>(resolve => complete = resolve));
+    const { fixture, target } = setup();
+    target.click();
+    expect(fixture.componentInstance.feedback()).toBe('');
+    complete();
+    await Promise.resolve();
+    expect(write).toHaveBeenCalledOnceWith('task-123');
+    expect(clipboard.copy).not.toHaveBeenCalled();
+    expect(fixture.componentInstance.feedback()).toBe('任务 ID 已复制');
+    fixture.destroy();
+  });
+
+  it('复制失败不会提示成功，ID 仍可手动选择', async () => {
+    clipboard.copy.and.returnValue(false);
+    const { fixture, target } = setup();
+    target.click();
+    await Promise.resolve();
+    fixture.detectChanges();
+    expect(fixture.nativeElement.textContent).toContain('复制失败，请选中任务 ID 手动复制');
+    expect(getComputedStyle(target.querySelector('.value')!).userSelect).toBe('text');
+    fixture.destroy();
+  });
+
+  it('拖选 ID 后的点击不会覆盖剪贴板或清空选区', () => {
+    const { fixture, target } = setup();
+    const range = document.createRange();
+    range.selectNodeContents(target.querySelector('.value')!);
+    const selection = window.getSelection()!;
+    selection.removeAllRanges();
+    selection.addRange(range);
+    target.click();
+    expect(clipboard.copy).not.toHaveBeenCalled();
+    expect(selection.toString()).toContain('task-123');
+    fixture.destroy();
+  });
+
+  it('支持键盘 Enter 复制', async () => {
+    clipboard.copy.and.returnValue(true);
+    const { fixture, target } = setup();
+    target.dispatchEvent(new KeyboardEvent('keydown', { key: 'Enter', bubbles: true }));
+    await Promise.resolve();
+    expect(clipboard.copy).toHaveBeenCalledOnceWith('task-123');
+    fixture.destroy();
+  });
+});

apps/showcase_ui/src/app/components/task-id-copy/task-id-copy.component.ts

@@ -0,0 +1,66 @@
+import { Clipboard } from '@angular/cdk/clipboard';
+import { Component, DestroyRef, Input, inject, signal } from '@angular/core';
+
+@Component({
+  selector: 'app-task-id-copy',
+  standalone: true,
+  template: `
+    <span class="copy-target" role="button" tabindex="0"
+      [attr.aria-label]="'复制任务 ID：' + taskId" title="点击复制任务 ID，也可选中文字复制"
+      (click)="copy($event)" (keydown.enter)="copy($event)" (keydown.space)="copy($event)">
+      <span class="value">ID: {{ taskId }}</span>
+      <span class="material-symbols-outlined" aria-hidden="true">content_copy</span>
+    </span>
+    <span class="feedback" role="status" aria-live="polite">{{ feedback() }}</span>
+  `,
+  styles: [`
+    :host { display: block; min-width: 0; font-size: 11px; color: inherit; user-select: text; }
+    .copy-target { display: inline-flex; align-items: center; gap: 5px; max-width: 100%;
+      padding: 3px 0; cursor: pointer; user-select: text; border-radius: 3px; }
+    .value { min-width: 0; overflow-wrap: anywhere; user-select: text; }
+    .material-symbols-outlined { flex-shrink: 0; font-size: 14px; user-select: none; }
+    .copy-target:hover { color: #2563eb; }
+    .copy-target:focus-visible { outline: 2px solid #60a5fa; outline-offset: 2px; }
+    .feedback { display: block; line-height: 1.6; user-select: text; }
+    .feedback:empty { display: none; }
+  `]
+})
+export class TaskIdCopyComponent {
+  @Input({ required: true }) taskId = '';
+  readonly feedback = signal('');
+  private readonly clipboard = inject(Clipboard);
+  private readonly destroyRef = inject(DestroyRef);
+  private timer: ReturnType<typeof setTimeout> | undefined;
+
+  constructor() {
+    this.destroyRef.onDestroy(() => clearTimeout(this.timer));
+  }
+
+  async copy(event: Event): Promise<void> {
+    event.stopPropagation();
+    // A drag selection must remain available for the browser's normal copy action.
+    const selection = window.getSelection();
+    if (event.type === 'click' && selection && !selection.isCollapsed
+      && (event.currentTarget as HTMLElement).contains(selection.anchorNode)) return;
+    event.preventDefault();
+    let copied = false;
+    try {
+      if (navigator.clipboard?.writeText) {
+        await navigator.clipboard.writeText(this.taskId);
+        copied = true;
+      } else {
+        copied = this.clipboard.copy(this.taskId);
+      }
+    } catch {
+      try {
+        copied = this.clipboard.copy(this.taskId);
+      } catch {
+        // Keep manual selection available when the browser blocks both clipboard APIs.
+      }
+    }
+    if (this.destroyRef.destroyed) return;
+    this.feedback.set(copied ? '任务 ID 已复制' : '复制失败，请选中任务 ID 手动复制');
+    clearTimeout(this.timer);
+    this.timer = setTimeout(() => this.feedback.set(''), 3000);
+  }
+}

apps/showcase_ui/src/app/components/task-mode-badge/task-mode-badge.component.spec.ts

@@ -0,0 +1,21 @@
+import { TestBed } from '@angular/core/testing';
+import { TaskModeBadgeComponent } from './task-mode-badge.component';
+
+describe('执行模式徽标', () => {
+  it('探索优先于底层 Pro，复跑显示实际使用的 Flash / Pro', () => {
+    const fixture = TestBed.createComponent(TaskModeBadgeComponent);
+    const check = (session: object, mode: string, text: string) => {
+      fixture.componentRef.setInput('session', session);
+      fixture.detectChanges();
+      const badge = fixture.nativeElement.querySelector('.mode-badge');
+      expect(badge.classList.contains(mode)).toBeTrue();
+      expect(badge.textContent).toContain(text);
+      expect(badge.title).toBe('执行模式：' + text);
+    };
+    check({exploration_kind:'exploration',profile:'pro'}, 'explore', '探索');
+    check({exploration_kind:'validation',profile:'flash'}, 'flash', 'Flash');
+    check({profile:'pro'}, 'pro', 'Pro');
+    check({model_info:{architecture:'ARTEMIS Flash'}}, 'flash', 'Flash');
+    check({model_info:{id:'mimo-v2.6-pro'}}, 'unknown', '模式未记录');
+  });
+});

apps/showcase_ui/src/app/components/task-mode-badge/task-mode-badge.component.ts

@@ -0,0 +1,27 @@
+import { Component, computed, input } from '@angular/core';
+import { Session } from '../../core/models/session.model';
+
+@Component({
+  selector: 'app-task-mode-badge',
+  standalone: true,
+  template: `<span class="mode-badge" [class]="'mode-badge ' + mode()" [title]="'执行模式：' + label()"><span class="material-symbols-outlined" aria-hidden="true">{{icon()}}</span>{{label()}}</span>`,
+  styles: [`
+    :host{display:inline-flex;vertical-align:middle}
+    .mode-badge{display:inline-flex;align-items:center;box-sizing:border-box;height:20px;gap:3px;padding:0 6px;border:1px solid #e2e8f0;border-radius:5px;background:#f8fafc;color:#475569;font-size:11px;font-weight:600;line-height:14px;white-space:nowrap}
+    .material-symbols-outlined{font-size:13px;line-height:1}
+    .explore{background:#eef2ff;border-color:#c7d2fe;color:#4338ca}
+    .flash{background:#fff7ed;border-color:#fed7aa;color:#9a3412}
+    .pro{background:#eff6ff;border-color:#bfdbfe;color:#1d4ed8}
+  `]
+})
+export class TaskModeBadgeComponent {
+  session = input.required<Partial<Session>>();
+  mode = computed<'explore' | 'flash' | 'pro' | 'unknown'>(() => {
+    const s = this.session();
+    if (s.exploration_kind === 'exploration') return 'explore';
+    const profile = (s.profile || s.model_info?.architecture || '').toLowerCase().replace(/^artemis\s+/, '');
+    return profile === 'explore' || profile === 'flash' || profile === 'pro' ? profile : 'unknown';
+  });
+  label = computed(() => ({explore:'探索',flash:'Flash',pro:'Pro',unknown:'模式未记录'})[this.mode()]);
+  icon = computed(() => ({explore:'explore',flash:'bolt',pro:'psychology',unknown:'help_outline'})[this.mode()]);
+}

apps/showcase_ui/src/app/components/task-plugins/task-plugins.component.html

@@ -0,0 +1,30 @@
+<button class="plugin-toggle" type="button" (click)="toggle()" [attr.aria-expanded]="open()">
+  <span class="material-symbols-outlined" aria-hidden="true">extension</span>
+  <span>扩展插件</span><span class="mode">{{ summary() }}</span>
+  <span class="material-symbols-outlined chevron" aria-hidden="true">{{ open() ? 'expand_less' : 'expand_more' }}</span>
+</button>
+@if (open()) {
+  <section class="plugin-options" aria-label="扩展插件设置">
+    <p>照常描述安卓测试任务。需要操作电脑、浏览器等外部步骤时，AI 会按需调用已启用并授权的插件。</p>
+    <label class="auto-option"><input type="checkbox" [(ngModel)]="automatic" (ngModelChange)="emit()">自动使用可用插件</label>
+    @if (loading()) { <p role="status">正在读取插件…</p> }
+    @else if (error()) { <p class="error" role="alert">{{ error() }} <button type="button" class="text-button" (click)="load()">重试</button></p> }
+    @else if (!automatic) {
+      @for (plugin of available(); track plugin.manifest.id) {
+        <div class="choice">
+          <label><input type="checkbox" [checked]="selected[plugin.manifest.id] !== undefined" (change)="choose(plugin, $event)">{{ plugin.manifest.name }}</label>
+          @if (authorizedProfiles(plugin).length > 1) {
+            <select [disabled]="selected[plugin.manifest.id] === undefined" [ngModel]="selected[plugin.manifest.id] || ''" (ngModelChange)="selected[plugin.manifest.id] = $event; emit()" [attr.aria-label]="plugin.manifest.name + '使用设置'">
+              <option value="" disabled>选择要使用的设置</option>
+              @for (profile of authorizedProfiles(plugin); track profile.name) { <option [value]="profile.name">{{ profile.label }}</option> }
+            </select>
+          }
+        </div>
+      } @empty { <p>还没有可用插件，请先安装、授权并启用。</p> }
+      <p class="hint">不勾选任何插件时，本次任务只使用手机能力。</p>
+    } @else {
+      <p class="hint">{{ available().length ? 'AI 会按需使用可用插件，无需逐次选择。关闭自动使用后，可指定本次允许使用的插件。' : '还没有可用插件，手机测试不受影响。' }}</p>
+    }
+    <a routerLink="/plugins" class="manage-link">管理插件 <span class="material-symbols-outlined" aria-hidden="true">arrow_forward</span></a>
+  </section>
+}

apps/showcase_ui/src/app/components/task-plugins/task-plugins.component.scss

@@ -0,0 +1,20 @@
+:host { display: block; width: 100%; color: #475569; font-size: 12px; }
+.plugin-toggle { display: inline-flex; align-items: center; gap: 6px; border: 1px solid #e2e8f0; border-radius: 20px; padding: 5px 10px; background: #f8fafc; color: #64748b; font: inherit; cursor: pointer; transition: background .15s, border-color .15s; }
+.plugin-toggle:hover, .plugin-toggle[aria-expanded=true] { background: #eff6ff; border-color: #bfdbfe; color: #2563eb; }
+.material-symbols-outlined { font-size: 16px; }
+.mode { color: #2563eb; padding-left: 6px; border-left: 1px solid #dce5ef; }
+.chevron { font-size: 15px; }
+.plugin-options { background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 14px; padding: 14px 16px; margin-top: 8px; display: grid; gap: 12px; }
+p { margin: 0; line-height: 1.65; }
+label { display: flex; align-items: center; gap: 8px; }
+.auto-option { font-weight: 500; color: #334155; }
+input[type=checkbox] { accent-color: #2563eb; margin: 0; width: 14px; height: 14px; flex-shrink: 0; }
+.choice { display: flex; align-items: center; justify-content: space-between; gap: 12px; flex-wrap: wrap; }
+select { font: inherit; color: #334155; background: #fff; border: 1px solid #dce3eb; border-radius: 8px; padding: 6px 10px; max-width: 100%; }
+select:disabled { opacity: .5; }
+.hint { color: #64748b; font-size: 11px; }
+.error { color: #b91c1c; }
+.text-button, .manage-link { color: #2563eb; font: inherit; text-decoration: none; }
+.text-button { background: none; border: 0; padding: 0 6px; cursor: pointer; }
+.manage-link { display: inline-flex; align-items: center; gap: 4px; justify-self: start; }
+:focus-visible { outline: 2px solid #3b82f6; outline-offset: 3px; }

apps/showcase_ui/src/app/components/task-plugins/task-plugins.component.spec.ts

@@ -0,0 +1,78 @@
+import { TestBed } from '@angular/core/testing';
+import { provideRouter } from '@angular/router';
+import { provideHttpClient } from '@angular/common/http';
+import { provideHttpClientTesting, HttpTestingController } from '@angular/common/http/testing';
+import { TaskPluginsComponent } from './task-plugins.component';
+
+describe('TaskPluginsComponent', () => {
+  let http: HttpTestingController;
+  beforeEach(() => {
+    TestBed.configureTestingModule({ imports: [TaskPluginsComponent], providers: [provideRouter([]), provideHttpClient(), provideHttpClientTesting()] });
+    http = TestBed.inject(HttpTestingController);
+  });
+  afterEach(() => http.verify());
+
+  it('starts collapsed and automatic without an execution target selector', async () => {
+    const fixture = TestBed.createComponent(TaskPluginsComponent);
+    fixture.detectChanges();
+    http.expectNone('/api/plugins');
+    expect(fixture.nativeElement.textContent).toContain('自动');
+    expect(fixture.nativeElement.textContent).not.toContain('执行目标');
+    expect(fixture.nativeElement.querySelector('select')).toBeNull();
+    const emitted = jasmine.createSpy('changed');
+    fixture.componentInstance.changed.subscribe(emitted);
+    fixture.nativeElement.querySelector('button').click(); fixture.detectChanges();
+    http.expectOne('/api/plugins').flush({ plugins: [] });
+    await fixture.whenStable(); fixture.detectChanges();
+    fixture.nativeElement.querySelector('input').click(); fixture.detectChanges();
+    expect(emitted).toHaveBeenCalledWith({ plugins: [] });
+    fixture.nativeElement.querySelector('input').click(); fixture.detectChanges();
+    expect(emitted).toHaveBeenCalledWith({});
+  });
+
+  it('uses the single authorized setting without exposing a connection selector', async () => {
+    const fixture = TestBed.createComponent(TaskPluginsComponent);
+    const component = fixture.componentInstance;
+    component.toggle();
+    const plugin = { enabled: true, manifest: { id: 'desktop', name: '电脑控制', targets: ['computer', 'mixed'] }, profiles: [
+      { name: 'default', label: '旧配置', authorized: false },
+      { name: 'lab', label: '实验室', authorized: true, is_default: true }
+    ] };
+    http.expectOne('/api/plugins').flush({ plugins: [plugin, { ...plugin, enabled: false, manifest: { ...plugin.manifest, id: 'disabled' } }] });
+    component.automatic = false; fixture.detectChanges(); await fixture.whenStable();
+    const emitted = jasmine.createSpy('changed'); component.changed.subscribe(emitted);
+    const checkbox = fixture.nativeElement.querySelector('.choice input');
+    checkbox.click(); fixture.detectChanges(); await fixture.whenStable();
+    expect(emitted).toHaveBeenCalledWith({ plugins: [{ id: 'desktop', profile: 'lab' }] });
+    expect(fixture.nativeElement.querySelectorAll('.choice').length).toBe(1);
+    expect(fixture.nativeElement.querySelector('select')).toBeNull();
+    expect(fixture.nativeElement.textContent).not.toContain('连接配置');
+  });
+
+  it('allows choosing among multiple authorized settings and excludes unauthorized ones', async () => {
+    const fixture = TestBed.createComponent(TaskPluginsComponent); const component = fixture.componentInstance;
+    component.toggle();
+    http.expectOne('/api/plugins').flush({ plugins: [{ enabled: true,
+      manifest: { id: 'desktop', name: '电脑控制', targets: ['mixed'] }, profiles: [
+        { name: 'old', label: '旧配置', authorized: false },
+        { name: 'lab', label: '实验室', authorized: true, is_default: true },
+        { name: 'office', label: '办公室', authorized: true }
+      ] }] });
+    component.automatic = false; fixture.detectChanges(); await fixture.whenStable();
+    const emitted = jasmine.createSpy('changed'); component.changed.subscribe(emitted);
+    fixture.nativeElement.querySelector('.choice input').click(); fixture.detectChanges(); await fixture.whenStable();
+    const select: HTMLSelectElement = fixture.nativeElement.querySelector('select');
+    expect(select.value).toBe('lab'); expect(select.textContent).not.toContain('旧配置');
+    select.value = 'office'; select.dispatchEvent(new Event('change')); fixture.detectChanges(); await fixture.whenStable();
+    expect(emitted).toHaveBeenCalledWith({ plugins: [{ id: 'desktop', profile: 'office' }] });
+  });
+
+  it('shows a retryable error and keeps automatic mode on lookup failure', () => {
+    const fixture = TestBed.createComponent(TaskPluginsComponent);
+    fixture.componentInstance.toggle();
+    http.expectOne('/api/plugins').flush({}, { status: 503, statusText: 'Unavailable' });
+    fixture.detectChanges();
+    expect(fixture.nativeElement.textContent).toContain('重试');
+    expect(fixture.componentInstance.automatic).toBeTrue();
+  });
+});

apps/showcase_ui/src/app/components/task-plugins/task-plugins.component.ts

@@ -0,0 +1,55 @@
+import { Component, inject, output, signal } from '@angular/core';
+import { FormsModule } from '@angular/forms';
+import { RouterLink } from '@angular/router';
+import { InstalledPlugin, PluginService } from '../../services/plugin.service';
+
+export interface TaskPluginOptions {
+  plugins?: { id: string; profile: string; version?: string }[];
+}
+
+@Component({
+  selector: 'app-task-plugins', standalone: true, imports: [FormsModule, RouterLink],
+  templateUrl: './task-plugins.component.html', styleUrl: './task-plugins.component.scss'
+})
+export class TaskPluginsComponent {
+  changed = output<TaskPluginOptions>();
+  private api = inject(PluginService);
+  automatic = true;
+  selected: Record<string, string> = {};
+  open = signal(false);
+  loading = signal(false);
+  error = signal('');
+  plugins = signal<InstalledPlugin[]>([]);
+
+  summary() {
+    return this.automatic ? '自动' : Object.keys(this.selected).length ? `已选 ${Object.keys(this.selected).length} 个` : '本次停用';
+  }
+  available() {
+    return this.plugins().filter(p => p.enabled && p.profiles.some(profile => profile.authorized) &&
+      p.manifest.targets.some(target => target === 'android' || target === 'mixed'));
+  }
+  authorizedProfiles(plugin: InstalledPlugin) { return plugin.profiles.filter(p => p.authorized); }
+  toggle() {
+    this.open.update(v => !v);
+    if (this.open()) this.load();
+  }
+  load() {
+    this.loading.set(true); this.error.set('');
+    this.api.list().subscribe({
+      next: value => { this.plugins.set(value.plugins); this.loading.set(false); },
+      error: () => { this.error.set('暂时无法读取插件。'); this.loading.set(false); }
+    });
+  }
+  choose(plugin: InstalledPlugin, event: Event) {
+    if ((event.target as HTMLInputElement).checked) {
+      const profiles = this.authorizedProfiles(plugin);
+      this.selected[plugin.manifest.id] = (profiles.find(p => p.is_default) || profiles[0]).name;
+    } else delete this.selected[plugin.manifest.id];
+    this.emit();
+  }
+  emit() {
+    this.changed.emit(this.automatic ? {} : {
+      plugins: Object.entries(this.selected).map(([id, profile]) => ({ id, profile }))
+    });
+  }
+}

apps/showcase_ui/src/app/components/test-result-badge/test-result-badge.component.ts

@@ -0,0 +1,36 @@
+import { Component, computed, input } from '@angular/core';
+import { CaseTestResult } from '../../core/models/session.model';
+
+export const TEST_RESULT_LABELS: Record<CaseTestResult['status'], string> = {
+  passed: '测试通过', failed: '测试不通过', inconclusive: '未验证', blocked: '测试阻塞'
+};
+
+@Component({
+  selector: 'app-test-result-badge',
+  standalone: true,
+  template: `<span [class]="'test-result-badge ' + status()" [class.compact]="compact()" [class.prominent]="prominent()" [title]="(result()?.source === 'human' ? '人工确认：' : 'AI确认：') + (result()?.reason || '执行状态与测试结论分别记录')"><span class="material-symbols-outlined" aria-hidden="true">{{icon()}}</span>{{label()}}</span>`,
+  styles: [`
+    :host{display:inline-flex;vertical-align:middle;max-width:100%}
+    .test-result-badge{display:inline-flex;align-items:center;gap:4px;white-space:nowrap;border-radius:6px;border:1px solid #cbd5e1;padding:3px 8px;font-size:12px;font-weight:700;line-height:1.5;background:#f1f5f9;color:#475569}
+    .material-symbols-outlined{font-size:16px;line-height:1}
+    .passed{background:#15803d;color:#fff;border-color:#15803d}
+    .failed{background:#b91c1c;color:#fff;border-color:#b91c1c}
+    .inconclusive{background:#fff7ed;color:#9a3412;border-color:#fdba74}
+    .blocked{background:#fef3c7;color:#92400e;border-color:#fbbf24}
+    .compact{box-sizing:border-box;height:20px;padding:0 6px;font-size:11px;line-height:14px;gap:3px;border-radius:5px}
+    .compact .material-symbols-outlined{font-size:13px}
+    .prominent{padding:4px 9px;font-size:14px;gap:5px}
+    .prominent .material-symbols-outlined{font-size:18px}
+  `]
+})
+export class TestResultBadgeComponent {
+  result = input<CaseTestResult | null | undefined>();
+  executionStatus = input('');
+  compact = input(false);
+  prominent = input(false);
+  status = computed(() => this.result()?.status || (['completed', 'success', 'failed', 'cancelled', 'interrupted', 'skipped'].includes(this.executionStatus()) ? 'inconclusive' : 'pending'));
+  label = computed(() => this.result()?.source === 'human'
+    ? (this.status() === 'passed' ? '人工通过' : '人工不通过')
+    : this.status() === 'pending' ? '待判定' : TEST_RESULT_LABELS[this.status() as CaseTestResult['status']]);
+  icon = computed(() => ({passed: 'check_circle', failed: 'cancel', blocked: 'block', inconclusive: 'help', pending: 'hourglass_empty'})[this.status()]);
+}

apps/showcase_ui/src/app/components/test-result-badge/test-result-card.component.spec.ts

@@ -0,0 +1,65 @@
+import { TestBed } from '@angular/core/testing';
+import { provideHttpClient } from '@angular/common/http';
+import { HttpTestingController, provideHttpClientTesting } from '@angular/common/http/testing';
+import { TestResultCardComponent } from './test-result-card.component';
+
+describe('用例测试结果卡片', () => {
+  beforeEach(() => TestBed.configureTestingModule({providers:[provideHttpClient(), provideHttpClientTesting()]}));
+  afterEach(() => TestBed.inject(HttpTestingController).verify());
+  it('显式展示不通过及原因，允许展开证据', () => {
+    const fixture = TestBed.createComponent(TestResultCardComponent);
+    fixture.componentRef.setInput('result', {status: 'failed', reason: '预期英文，实际为中文。', evidence: ['截图显示搜索设置']});
+    fixture.detectChanges();
+    const view: HTMLElement = fixture.nativeElement;
+    expect(view.querySelector('[aria-label="用例测试结果"]')).not.toBeNull();
+    expect(view.querySelector('.prominent')?.textContent).toContain('测试不通过');
+    expect(view.querySelector('.reason')?.textContent).toContain('预期英文，实际为中文。');
+    const details = view.querySelector('details')!;
+    expect(details.open).toBeFalse();
+    details.querySelector('summary')!.click();
+    expect(details.open).toBeTrue();
+    expect(details.textContent).toContain('截图显示搜索设置');
+  });
+
+  it('人工复验可覆盖 AI 判定，保留原始结论并提交备注', () => {
+    const fixture = TestBed.createComponent(TestResultCardComponent);
+    const ai = {status:'failed',reason:'AI发现不一致',evidence:['步骤 3 截图'],source:'ai'};
+    fixture.componentRef.setInput('result', ai);
+    fixture.componentRef.setInput('sessionId', 'task-1'); fixture.componentRef.setInput('executionStatus','completed');
+    const saved = jasmine.createSpy(); fixture.componentInstance.reviewSaved.subscribe(saved);
+    fixture.detectChanges();
+    fixture.nativeElement.querySelector('.review-toolbar button').click(); fixture.detectChanges();
+    fixture.componentInstance.note = '实际符合预期，已人工核对';
+    fixture.nativeElement.querySelector('.confirm-pass').click(); fixture.detectChanges();
+    expect(fixture.nativeElement.querySelector('.confirm-fail').disabled).toBeTrue();
+    const request = TestBed.inject(HttpTestingController).expectOne('/api/automation/reports/task-1/review');
+    expect(request.request.method).toBe('PUT');
+    expect(request.request.body).toEqual({status:'passed',reason:'实际符合预期，已人工核对'});
+    const manual = {status:'passed',reason:'实际符合预期，已人工核对',evidence:[],source:'human',reviewed_at:100,ai_result:ai};
+    request.flush({id:'task-1',test_result:manual,verdict:'测试通过（人工确认）'});
+    fixture.componentRef.setInput('result',manual); fixture.detectChanges();
+    expect(saved).toHaveBeenCalled();
+    expect(fixture.nativeElement.textContent).toContain('人工通过');
+    expect(fixture.nativeElement.textContent).toContain('原 AI 结论');
+    expect(fixture.nativeElement.textContent).toContain('AI发现不一致');
+    expect(fixture.nativeElement.querySelector('.review-form')).toBeNull();
+    fixture.destroy();
+  });
+
+  it('保存失败时保留备注并允许重试，执行中不显示人工复验', () => {
+    const fixture = TestBed.createComponent(TestResultCardComponent);
+    fixture.componentRef.setInput('result',{status:'inconclusive',reason:'尚未验证',evidence:[]});
+    fixture.componentRef.setInput('sessionId','task-2'); fixture.componentRef.setInput('executionStatus','running');
+    fixture.detectChanges(); expect(fixture.nativeElement.querySelector('.review-toolbar')).toBeNull();
+    fixture.componentRef.setInput('executionStatus','failed'); fixture.detectChanges();
+    fixture.nativeElement.querySelector('.review-toolbar button').click(); fixture.detectChanges();
+    fixture.componentInstance.note = '人工发现异常';
+    fixture.nativeElement.querySelector('.confirm-fail').click();
+    TestBed.inject(HttpTestingController).expectOne('/api/automation/reports/task-2/review').flush({detail:'报告正在打包'}, {status:409,statusText:'Conflict'});
+    fixture.detectChanges();
+    expect(fixture.componentInstance.note).toBe('人工发现异常');
+    expect(fixture.nativeElement.querySelector('[role="alert"]').textContent).toContain('报告正在打包');
+    expect(fixture.nativeElement.querySelector('.confirm-fail').disabled).toBeFalse();
+    fixture.destroy();
+  });
+});