modelcontextprotocol / modelcontextprotocol/python-sdk
find_context_parameter() fails to detect Context parameter in callable class instances
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 24.3k
- Forks
- 4k
- Avg merge
- 1d 1h
- Merged PRs (30d)
- 31
Description
Initial Checks
- I confirm that I'm using the latest version of MCP Python SDK
- I confirm that I searched for my issue in https://github.com/modelcontextprotocol/python-sdk/issues before opening this issue
Description
When registering a callable class instance as an MCP tool via FastMCP.add_tool(), the ctx: Context parameter is incorrectly exposed as an externally visible tool parameter instead of being injected by the framework. This is because find_context_parameter() uses typing.get_type_hints() which doesn't introspect the call method of callable class instances.
Suggested Fix
Update find_context_parameter() to handle callable class instances:
def find_context_parameter(fn: Callable[..., Any]) -> str | None:
from mcp.server.fastmcp.server import Context
# Handle callable class instances by using __call__ method
target = fn
if not (inspect.isfunction(fn) or inspect.ismethod(fn)):
if callable(fn) and hasattr(fn, "__call__"):
target = fn.__call__
try:
hints = typing.get_type_hints(target)
except Exception:
return None
# ... rest of function unchanged
Example Code
from typing import Annotated, Any
from mcp.server.fastmcp import Context, FastMCP
from mcp.server.session import ServerSession
from starlette.requests import Request
import pydantic
class MyTool:
def __init__(self, name: str):
self.__name__ = name
async def __call__(
self,
query: Annotated[str, pydantic.Field(description="Search query")],
ctx: Context[ServerSession, Any, Request],
) -> str:
"""Performs a search."""
return f"Results for: {query}"
mcp = FastMCP(name="Test Server")
mcp.add_tool(MyTool(name="my_search"), name="my_search", description="Search tool")
# When listing tools, 'ctx' appears as a required parameter in the schema
Python & MCP Python SDK
1.26.0
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start at find_context_parameter(), which is used while registering tools through FastMCP.add_tool(). Compare its handling of regular functions with callable class instances and verify the example no longer exposes ctx as a tool parameter when the callable instance is registered.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- api, backend
- Issue type
- Bug
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Clearly specified
- Newbie friendliness
- 55/100