feat(agent): redesign project_cli with module/method/input protocol

- Replace command/subcommand/args with module/method/input envelope
- Calendar handler uses discriminated union (mode) for read operations
- Strict Pydantic models with extra='forbid' for all calendar methods
- Worker max_iters=7, router prompt simplified (removed project_cli_defaults)
- Skill index cards + per-action files for progressive disclosure
- Frontend/AG-UI aligned to module/method dispatch
- Protocol docs updated to module/method/input contract

WIP: action cards need envelope fix, 2 tests need update, memory
handler needs Pydantic models.

backend/src/core/agentscope/runtime/runner.py

@@ -39,7 +39,9 @@ from schemas.agent.forwarded_props import (
     parse_forwarded_props_runtime_mode,
 )
 from schemas.agent.runtime_models import (
+    ErrorInfo,
     RouterAgentOutput,
+    RunStatus,
     WorkerAgentOutputLite,
 )
 from schemas.agent.skill_config import ProjectCliCommand, SkillName
@@ -74,6 +76,8 @@ class AgentScopeRunner:
         self._active_agent: JsonReActAgent | None = None
         self._active_agent_lock = asyncio.Lock()
 
+    _WORKER_MAX_ITERS = 7
+
     async def execute(
         self,
         *,
@@ -442,6 +446,11 @@ class AgentScopeRunner:
                     if self._active_agent is agent:
                         self._active_agent = None
             worker_payload = worker_output_model.model_validate(response_msg.metadata or {})
+            worker_payload = self._enforce_tool_evidence_contract(
+                worker_output=worker_payload,
+                requires_tool_evidence=requires_tool_evidence,
+                has_successful_tool_result=emitter.has_successful_tool_result,
+            )
             response_metadata = self._llm_pricing_service.build_usage_metadata(
                 model=stage_config.model_code,
                 usage_summary=tracking_model.usage_summary(),
@@ -458,6 +467,28 @@ class AgentScopeRunner:
         finally:
             reset_tool_credential(credential_token)
 
+    @staticmethod
+    def _enforce_tool_evidence_contract(
+        *,
+        worker_output: WorkerAgentOutputLite,
+        requires_tool_evidence: bool,
+        has_successful_tool_result: bool,
+    ) -> WorkerAgentOutputLite:
+        if not requires_tool_evidence or has_successful_tool_result:
+            return worker_output
+        return worker_output.model_copy(
+            update={
+                "status": RunStatus.FAILED,
+                "answer": "无法确认结果：所需工具调用未成功完成。",
+                "suggested_actions": [],
+                "error": ErrorInfo(
+                    code="TOOL_EVIDENCE_MISSING",
+                    message="requires_tool_evidence=true but no tool call completed successfully in this run",
+                    retryable=False,
+                ),
+            }
+        )
+
     def _build_worker_input_messages(
         self,
         *,
@@ -501,6 +532,7 @@ class AgentScopeRunner:
         model: TrackingChatModel,
         emitter: PipelineStageEmitter | None = None,
         force_tool_on_first_reasoning: bool = False,
+        max_iters: int = _WORKER_MAX_ITERS,
     ) -> JsonReActAgent:
         return JsonReActAgent(
             name=agent_name,
@@ -511,6 +543,7 @@ class AgentScopeRunner:
             memory=InMemoryMemory(),
             emitter=emitter,
             force_tool_on_first_reasoning=force_tool_on_first_reasoning,
+            max_iters=max_iters,
         )
 
     async def _emit_step_event(