feat: add shared test acceptance and release 1.5.0

artemis/agents/checker/checker.py

@@ -46,6 +46,7 @@ from pydantic import BaseModel, Field, ValidationError
 
 from artemis.core.tool_failure import ToolFailure, is_tool_failure
 from artemis.constants import CHECKER_MAX_ITERATIONS
+from artemis.runtime.case_result import CaseTestResult, normalize_case_result
 from artemis.context import ArtemisContext
 from artemis.data_engine.context_vars import CURRENT_TRACE_ID
 from artemis.data_engine.trace import trace, trace_langchain_tool
@@ -98,6 +99,10 @@ class CheckVerdict(BaseModel):
 
 class CheckReport(BaseModel):
     verdicts: list[CheckVerdict] = Field(default_factory=list)
+    test_result: CaseTestResult | None = Field(
+        default=None,
+        description="Final review only: overall test verdict against the user's prompt and guidance, independent of execution completion.",
+    )
     unmet_subgoals: list[str] = Field(
         default_factory=list,
         description=(
@@ -287,6 +292,7 @@ def assemble_checker_prompt_segments(
         segments.append(prompts.get("anchor_guide", ""))
     else:
         segments.append(prompts.get("final_guide", ""))
+        segments.append(prompts.get("case_result_guide", ""))
     if probe_tool_registered:
         segments.append(prompts.get("probe_guide", ""))
     return "\n\n".join(s for s in segments if s)
@@ -458,7 +464,11 @@ def _normalize_report(report: CheckReport, check_items: list) -> CheckReport:
                     evidence="no verdict produced for this item",
                 )
             )
-    return CheckReport(verdicts=verdicts, unmet_subgoals=list(report.unmet_subgoals))
+    return CheckReport(
+        verdicts=verdicts,
+        unmet_subgoals=list(report.unmet_subgoals),
+        test_result=normalize_case_result(report.test_result) if report.test_result else None,
+    )
 
 
 async def _run_check_loop(
@@ -796,6 +806,7 @@ async def run_final_check(
     ledger: list[dict],
     check_items,
     attempt_id: str | None = None,
+    execution_report: dict | str | None = None,
 ) -> CheckReport:
     """Final review at task exit: audits the user's original goal and all
     declared check items, with the final screen state available."""
@@ -828,6 +839,8 @@ async def run_final_check(
     human_text = (
         f"# User's Original Goal\n{goal}\n\n"
         f"# User guidance received during the run\n{_format_user_guidance(ctx, steps)}\n\n"
+        f"# Executor's Final Report (claims to verify, not independent evidence)\n"
+        f"{json.dumps(execution_report, ensure_ascii=False, default=str) if execution_report is not None else '(not provided)'}\n\n"
         f"# Final Task Plan\n{plan_text or '(no plan)'}\n\n"
         f"# Declared Check Items\n{_format_check_items(items)}\n\n"
         f"# Verdict Ledger (checkpoint results, authoritative for on_complete"
@@ -852,7 +865,15 @@ async def run_final_check(
     ]
     with CheckerStreamCapture(ctx) as capture:
         try:
-            return await _run_check_loop(ctx, messages, tools, items)
+            report = await _run_check_loop(ctx, messages, tools, items)
+            report.test_result = (
+                normalize_case_result(report.test_result)
+                if report.test_result
+                else CaseTestResult(
+                    status="inconclusive", reason="验收未生成测试结论，不能确认用例是否通过。"
+                )
+            )
+            return report
         finally:
             _persist_stream(
                 ctx, capture, attempt_id=attempt_id, phase="final", checkpoint_id="final"