When a tool uses from __future__ import annotations and has a Context or Depends() parameter, without_injected_parameters creates a wrapper function to strip those params before Pydantic processes the signature. But the wrapper inherited __module__ = 'fastmcp.server.dependencies' instead of the user's module. Pydantic resolves string annotations (which is what future annotations produces) via sys.modules[fn.__module__].__dict__ — so it was looking in the wrong namespace and failing with NameError for any type not imported by dependencies.py itself (Annotated, Literal, user-defined types, etc.).

The fix sets wrapper.__module__ and wrapper.__qualname__ from the original function so Pydantic resolves annotations against the user's module where the names actually exist.

from __future__ import annotations
from typing import Annotated
from pydantic import Field
from fastmcp import FastMCP, Context

mcp = FastMCP("example")

@mcp.tool()
def my_tool(query: Annotated[str, Field(description="Search query")], ctx: Context) -> str:
    return f"Result for: {query}"
# Previously: NameError: name 'Annotated' is not defined
# Now: works correctly

Closes #3238, closes #905