feat: add multi-device concurrency, server lifecycle CLI, and unified execution stream

- Implement DevicePool and DeviceExecutionLock concurrency modes (per-device FIFO vs global serialization)
- Extend ArtemisClient SDK and CLI with explicit device targeting and concurrency strategy configuration
- Add server lifecycle management (artemis restart, stop, status) with cross-platform PID discovery
- Update MCP server tools to support multi-device dispatch, trace tagging, and per-device execution
- Enhance Showcase UI with step-by-step visual replay, playback speed controls, and device serial badges
- Add comprehensive unit test coverage and edge-case verification scripts

artemis/engine/graph_runner.py

@@ -50,7 +50,7 @@ class GraphRunner(BaseRunner):
                     artemis_ctx = ArtemisContext(
                         device=DeviceContext(
                             host_platform=platform.os_type.name,
-                            device_serial=self.driver.device_id,
+                            device_id=self.driver.device_id,
                         )
                     )
                 graph = await get_graph(artemis_ctx)

artemis/engine/pipeline.py

@@ -12,6 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import os
 from typing import Literal
 
 from artemis.config import OutputConfig, load_agent_config
@@ -50,13 +51,16 @@ class Pipeline:
         device_id: str = "default-device",
         output_config: OutputConfig | None = None,
         enable_outputter: bool | None = None,
+        locked_package: str | None = None,
+        **kwargs: Any,
     ) -> ExecutionContextState:
         ctx = ExecutionContext(task_goal=goal, device_id=device_id)
         if not driver:
             driver = MockDeviceDriver(device_id=device_id)
 
+        max_turns = kwargs.get("max_turns", 5)
         runner = cls.create_runner(profile=profile, ctx=ctx, driver=driver)
-        state = await runner.run()
+        state = await runner.run(max_turns=max_turns)
 
         # Check Outputter configuration
         try:

artemis/engine/reactive_runner.py

@@ -59,7 +59,7 @@ class ReactiveRunner(BaseRunner):
                     artemis_ctx = ArtemisContext(
                         device=DeviceContext(
                             host_platform=platform.os_type.name,
-                            device_serial=self.driver.device_id,
+                            device_id=self.driver.device_id,
                         )
                     )
                 flash_runner = FlashRunner(