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/tests/unit/core/agentscope/runtime/test_runner.py

@@ -7,6 +7,7 @@ from ag_ui.core import RunAgentInput
 import core.agentscope.runtime.runner as runner_module
 from core.agentscope.runtime.runner import AgentScopeRunner
 from schemas.agent.runtime_models import (
+    RunStatus,
     RouterAgentOutput,
     WorkerAgentOutputLite,
 )
@@ -60,6 +61,31 @@ def test_build_worker_input_messages_only_contains_router_contract() -> None:
     assert "[RouterAgentOutput]" in str(input_messages[0].content)
 
 
+def test_build_agent_sets_worker_max_iters(
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    captured: dict[str, object] = {}
+
+    class _FakeJsonReActAgent:
+        def __init__(self, **kwargs: object) -> None:
+            captured.update(kwargs)
+
+    monkeypatch.setattr(runner_module, "JsonReActAgent", _FakeJsonReActAgent)
+
+    runner = AgentScopeRunner()
+    model = runner_module.TrackingChatModel(object())
+
+    agent = runner._build_agent(
+        agent_name="worker",
+        system_prompt="test",
+        toolkit=object(),
+        model=model,
+    )
+
+    assert isinstance(agent, _FakeJsonReActAgent)
+    assert captured["max_iters"] == 7
+
+
 def test_build_router_messages_injects_user_input_when_context_last_not_user() -> None:
     runner = AgentScopeRunner()
     run_input = _run_input()
@@ -119,6 +145,45 @@ def test_build_router_messages_appends_user_input_to_context_tail() -> None:
     assert messages[0].content == "上一轮回复"
 
 
+def test_enforce_tool_evidence_contract_keeps_success_when_tool_succeeds() -> None:
+    runner = AgentScopeRunner()
+
+    worker_output = runner._enforce_tool_evidence_contract(
+        worker_output=WorkerAgentOutputLite(
+            status=RunStatus.SUCCESS,
+            answer="今天没有日程",
+            suggested_actions=["查明天"],
+        ),
+        requires_tool_evidence=True,
+        has_successful_tool_result=True,
+    )
+
+    assert worker_output.status == RunStatus.SUCCESS
+    assert worker_output.answer == "今天没有日程"
+    assert worker_output.suggested_actions == ["查明天"]
+    assert worker_output.error is None
+
+
+def test_enforce_tool_evidence_contract_forces_failure_without_successful_tool() -> None:
+    runner = AgentScopeRunner()
+
+    worker_output = runner._enforce_tool_evidence_contract(
+        worker_output=WorkerAgentOutputLite(
+            status=RunStatus.SUCCESS,
+            answer="今天没有日程",
+            suggested_actions=["查明天"],
+        ),
+        requires_tool_evidence=True,
+        has_successful_tool_result=False,
+    )
+
+    assert worker_output.status == RunStatus.FAILED
+    assert worker_output.answer == "无法确认结果：所需工具调用未成功完成。"
+    assert worker_output.suggested_actions == []
+    assert worker_output.error is not None
+    assert worker_output.error.code == "TOOL_EVIDENCE_MISSING"
+
+
 def test_build_model_omits_none_generate_kwargs(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None: