fix(agent): address CRITICAL/HIGH security and validation issues

- Fix SSE JSON injection: use json.dumps for safe serialization
- Add tool validation to dispatcher with allowlist
- Add field validation to tool_registry with proper error handling
- Add run_id consistency check (409 on mismatch)
- Add RunAgentInput constraints: min_length, extra=forbid
- Fix crewai_flow: use Field(default_factory), prefix unused params

backend/src/v1/agent/router.py

@@ -2,7 +2,7 @@ from __future__ import annotations
 
 from typing import Annotated
 
-from fastapi import APIRouter, Depends
+from fastapi import APIRouter, Depends, HTTPException
 from fastapi.responses import StreamingResponse
 
 from v1.agent.dependencies import get_agent_service
@@ -29,6 +29,11 @@ async def resume_run(
     input_data: RunAgentInput,
     service: Annotated[AgentChatService, Depends(get_agent_service)],
 ) -> StreamingResponse:
+    if input_data.runId != run_id:
+        raise HTTPException(
+            status_code=409,
+            detail=f"run_id mismatch: path={run_id}, body={input_data.runId}",
+        )
     return StreamingResponse(
         service.stream_resume(run_id, input_data),
         media_type="text/event-stream",