Repository navigation
Revert "test(mcp): verify scoped execution and OAuth credential isolation" - #42039
Conversation
 |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| @@ -1311,27 +1311,6 @@ | |||
| ], | |||
| "tests/integration/cost_calculation/test_cost_tracking.py::test_case_bills_expected_cost[us.anthropic.claude-opus-5-v1:0-stream_full_usage]": [ | |||
| "quota_management.spend_tracking.scripted_wire.logs_cost" | |||
There was a problem hiding this comment.
Rule Used: What: Flag any modifications to existing tests and verify they don't weaken test coverage or mask regressions. Why: Developers may alter tests to make failing code pass rather than fix the actual bug, hiding regressions. Good: ``` // Test updated t... (source)
Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!
| class MockProxyLogging: | ||
| """Mock proxy logging object for testing MCP guardrails""" | ||
|
|
||
| def __init__(self, guardrails: Optional[list] = None): | ||
| self.guardrails = guardrails if guardrails is not None else [] | ||
| self.call_details = {"user_api_key_cache": DualCache()} | ||
| self.dynamic_success_callbacks = [] | ||
| self.call_count = 0 | ||
|
|
||
| def get_combined_callback_list(self, dynamic_success_callbacks, global_callbacks): | ||
| """Return the guardrails for testing""" | ||
| return self.guardrails | ||
|
|
||
| def _convert_mcp_to_llm_format(self, request_obj, kwargs: dict) -> dict: | ||
| """Convert MCP tool call to LLM message format""" | ||
| tool_call_content = ( | ||
| f"Tool: {request_obj.tool_name}\nArguments: {request_obj.arguments}" | ||
| ) | ||
|
|
||
| return { | ||
| "messages": [{"role": "user", "content": tool_call_content}], | ||
| "model": kwargs.get("model", "mcp-tool-call"), | ||
| "user_api_key_user_id": kwargs.get("user_api_key_user_id"), | ||
| "user_api_key_team_id": kwargs.get("user_api_key_team_id"), | ||
| } | ||
|
|
||
| def _convert_llm_result_to_mcp_response(self, llm_result, request_obj): | ||
| """Convert LLM result back to MCP response format""" | ||
| return None # For testing, we don't need to convert back | ||
|
|
||
| def _parse_pre_mcp_call_hook_response(self, response, original_request): | ||
| """Parse pre MCP call hook response""" | ||
| return response | ||
|
|
||
| async def async_pre_mcp_tool_call_hook( | ||
| self, | ||
| kwargs: dict, | ||
| request_obj: Any, | ||
| start_time: datetime, | ||
| end_time: datetime, | ||
| ) -> Optional[Any]: |
There was a problem hiding this comment.
Context Used: AGENTS.md (source)
Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!
| async def async_pre_mcp_tool_call_hook( | ||
| self, kwargs, request_obj: MCPPreCallRequestObject, start_time, end_time | ||
| ) -> Optional[MCPPreCallResponseObject]: |
There was a problem hiding this comment.
Any, violating the requirement for strongly typed function parameters. Fix this before merging
Context Used: AGENTS.md (source)
| """ | ||
| Test file for MCP Hook Architecture | ||
|
|
||
| This file demonstrates the new MCP hook system with comprehensive examples | ||
| and validation tests. | ||
| """ |
There was a problem hiding this comment.
Context Used: AGENTS.md (source)
Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
2 participants
Reverts #41731