Sanitizer misclassifies tool results due to whitespace in tool_call_id

Fixes on record

Merged to main. A stamp counts the issues a fix is linked to.

• #55519 fix: normalize tool_call_id whitespace in sanitizer

  merged to main

  files: agent/agent_runtime_helpers.py, run_agent.py, tests/run_agent/test_agent_guardrails.py

  Show the diff

  • agent/agent_runtime_helpers.py

    @@ -2163,7 +2163,7 @@ def sanitize_api_messages(messages: List[Dict[str, Any]]) -> List[Dict[str, Any]
         result_call_ids: set = set()
         for msg in messages:
             if msg.get("role") == "tool":
    -            cid = msg.get("tool_call_id")
    +            cid = (msg.get("tool_call_id") or "").strip()
                 if cid:
                     result_call_ids.add(cid)
     
    @@ -2172,7 +2172,7 @@ def sanitize_api_messages(messages: List[Dict[str, Any]]) -> List[Dict[str, Any]
         if orphaned_results:
             messages = [
                 m for m in messages
    -            if not (m.get("role") == "tool" and m.get("tool_call_id") in orphaned_results)
    +            if not (m.get("role") == "tool" and (m.get("tool_call_id") or "").strip() in orphaned_results)
             ]
             _ra().logger.debug(
                 "Pre-call sanitizer: removed %d orphaned tool result(s)",

  • run_agent.py

    @@ -3429,8 +3429,8 @@ def _build_system_prompt(self, system_message: str = None) -> str:
         def _get_tool_call_id_static(tc) -> str:
             """Extract call ID from a tool_call entry (dict or object)."""
             if isinstance(tc, dict):
    -            return tc.get("call_id", "") or tc.get("id", "") or ""
    -        return getattr(tc, "call_id", "") or getattr(tc, "id", "") or ""
    +            return (tc.get("call_id", "") or tc.get("id", "") or "").strip()
    +        return (getattr(tc, "call_id", "") or getattr(tc, "id", "") or "").strip()
     
         @staticmethod
         def _get_tool_call_name_static(tc) -> str:

  • tests/run_agent/test_agent_guardrails.py

    @@ -108,6 +108,30 @@ def test_sdk_object_tool_calls(self):
             assert len(out) == 2
             assert out[1]["tool_call_id"] == "c6"
     
    +    def test_tool_result_with_leading_whitespace_preserved(self):
    +        """Tool result IDs with leading/trailing whitespace should match assistant call IDs."""
    +        msgs = [
    +            {"role": "assistant", "tool_calls": [assistant_dict_call("functions.cronjob:24")]},
    +            tool_result(" functions.cronjob:24"),  # leading whitespace
    +        ]
    +        out = AIAgent._sanitize_api_messages(msgs)
    +        # Should NOT inject a stub — the tool result is valid after stripping
    +        assert len(out) == 2
    +        assert out[1]["role"] == "tool"
    +        assert out[1]["content"] == "ok"
    +
    +    def test_truly_orphaned_with_whitespace_still_removed(self):
    +        """Truly orphaned tool results with whitespace should still be removed."""
    +        msgs = [
    +            {"role": "assistant", "tool_calls": [assistant_dict_call("c_valid")]},
    +            tool_result(" c_ORPHAN "),  # whitespace + no matching call
    +        ]
    +        out = AIAgent._sanitize_api_messages(msgs)
    +        assert len(out) == 2  # assistant + stub for c_valid, orphan removed
    +        tool_msgs = [m for m in out if m["role"] == "tool"]
    +        assert len(tool_msgs) == 1
    +        assert tool_msgs[0]["tool_call_id"] == "c_valid"
    +
     
     # ---------------------------------------------------------------------------
     # Phase 2a — _cap_delegate_task_calls