phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-06-27 18:50:41 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: accept MCP authorization headers for MCP toolgroups (#2230)

Browse source 
The most interesting MCP servers are those with an authorization wall in
front of them. This PR uses the existing `provider_data` mechanism of
passing provider API keys for passing MCP access tokens (in fact,
arbitrary headers in the style of the OpenAI Responses API) from the
client through to the MCP server.

```
class MCPProviderDataValidator(BaseModel):
    # mcp_endpoint => list of headers to send
    mcp_headers: dict[str, list[str]] | None = None
```

Note how we must stuff the headers for all MCP endpoints into a single
"MCPProviderDataValidator". Unlike existing providers (e.g., Together
and Fireworks for inference) where we could name the provider api keys
clearly (`together_api_key`, `fireworks_api_key`), we cannot name these
keys for MCP. We have a single generic MCP provider which can serve
multiple "toolgroups". So we use a dict to combine all the headers for
all MCP endpoints you may want to use in an agentic call.


## Test Plan

See the added integration test for usage.
This commit is contained in:
Ashwin Bharambe 2025-05-23 08:52:18 -07:00 • committed by GitHub
parent 2708312168
commit 51945f1e57
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
9 changed files with 331 additions and 47 deletions
Download patch file Download diff file Expand all files Collapse all files

4
llama_stack/distribution/server/server.py
View file

@ -28,7 +28,7 @@ from fastapi.responses import JSONResponse, StreamingResponse
from openai import BadRequestError
from pydantic import BaseModel, ValidationError
from llama_stack.distribution.datatypes import LoggingConfig, StackRunConfig
from llama_stack.distribution.datatypes import AuthenticationRequiredError, LoggingConfig, StackRunConfig
from llama_stack.distribution.distribution import builtin_automatically_routed_apis
from llama_stack.distribution.request_headers import (
PROVIDER_DATA_VAR,
@ -122,6 +122,8 @@ def translate_exception(exc: Exception) -> HTTPException | RequestValidationErro
return HTTPException(status_code=504, detail=f"Operation timed out: {str(exc)}")
elif isinstance(exc, NotImplementedError):
return HTTPException(status_code=501, detail=f"Not implemented: {str(exc)}")
elif isinstance(exc, AuthenticationRequiredError):
return HTTPException(status_code=401, detail=f"Authentication required: {str(exc)}")
else:
return HTTPException(
status_code=500,