scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-06 02:48:13 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 103

fix(mcp): address Copilot review and CI gates on request-tags PR - explicit [] no longer falls back to header (is not None), Final on header_value, TQ008 suppressions on new patches, sync stale schema.d.ts soft_budget removal

Browse source
This commit is contained in:
onatozmenn 2026-09-14 00:09:29 +03:00
parent bc63bba9a3
commit 36b1a22f03
No known key found for this signature in database
3 changed files with 35 additions and 18 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
litellm/proxy/_experimental/mcp_server/server.py
View file

@ -213,7 +213,7 @@ def _request_tags_from_raw_headers(
) -> Sequence[str] | None:
"""The caller's tags, parsed by the same helper the LLM routes use so an MCP operation and a
chat completion attribute an identical header identically."""
header_value = _request_tags_header(raw_headers)
header_value: Final = _request_tags_header(raw_headers)
if header_value is None:
return None
return LiteLLMProxyRequestSetup.add_request_tag_to_metadata(
@ -2190,7 +2190,11 @@ if MCP_AVAILABLE:
list_tools_call_id: Final = str(uuid.uuid4())
# Derive trace_id from raw_headers when not explicitly passed (same as A2A / MCP call_tool)
effective_litellm_trace_id: Final = litellm_trace_id or get_chain_id_from_headers(raw_headers)
effective_request_tags: Final = request_tags or _request_tags_from_raw_headers(raw_headers)
# An explicit [] means the caller resolved to no tags; only fall back to the
# header when nothing was passed at all.
effective_request_tags: Final = (
request_tags if request_tags is not None else _request_tags_from_raw_headers(raw_headers)
)
spend_logs_metadata: Final[dict[str, object]] = {
"mcp_operation": "list_tools",
}

43
tests/test_litellm/proxy/_experimental/mcp_server/test_mcp_server.py
View file

@ -5031,26 +5031,26 @@ async def test_get_tools_from_mcp_servers_takes_list_tools_tags_from_x_litellm_t
return dummy_logging_obj, None
with (
patch(
patch( # test-quality-ok: server allowlist is a module-level function; the suite has no injection seam
"litellm.proxy._experimental.mcp_server.server._get_allowed_mcp_servers",
new=AsyncMock(return_value=[server_a]),
),
patch(
patch( # test-quality-ok: header prep is a module-level function; the suite has no injection seam
"litellm.proxy._experimental.mcp_server.server._prepare_mcp_server_headers",
return_value=(None, None),
),
patch(
patch( # test-quality-ok: manager is a module-level singleton; patching it is the suite established seam
"litellm.proxy._experimental.mcp_server.server.global_mcp_server_manager",
) as mock_manager,
patch(
patch( # test-quality-ok: tool filter is a module-level function; the suite has no injection seam
"litellm.proxy._experimental.mcp_server.server.filter_tools_by_allowed_tools",
side_effect=lambda tools, _server: tools,
),
patch(
patch( # test-quality-ok: permission filter is a module-level async function; the suite has no injection seam
"litellm.proxy._experimental.mcp_server.server.filter_tools_by_key_team_permissions",
new=AsyncMock(side_effect=lambda tools, **_: tools),
),
patch(
patch( # test-quality-ok: logging setup is a module-level function; patched to capture spend-log metadata kwargs
"litellm.proxy._experimental.mcp_server.server.function_setup",
side_effect=_capture_function_setup,
),
@ -5073,7 +5073,8 @@ async def test_get_tools_from_mcp_servers_takes_list_tools_tags_from_x_litellm_t
@pytest.mark.asyncio
async def test_get_tools_from_mcp_servers_prefers_explicit_request_tags_over_the_header():
"""`request_tags` is the resolved value a caller passes in; a header must not override it."""
"""`request_tags` is the resolved value a caller passes in; a header must not override it.
An explicit empty list resolves to no tags rather than falling back to the header."""
try:
from litellm.proxy._experimental.mcp_server.server import (
_get_tools_from_mcp_servers,
@ -5105,26 +5106,26 @@ async def test_get_tools_from_mcp_servers_prefers_explicit_request_tags_over_the
return dummy_logging_obj, None
with (
patch(
patch( # test-quality-ok: server allowlist is a module-level function; the suite has no injection seam
"litellm.proxy._experimental.mcp_server.server._get_allowed_mcp_servers",
new=AsyncMock(return_value=[server_a]),
),
patch(
patch( # test-quality-ok: header prep is a module-level function; the suite has no injection seam
"litellm.proxy._experimental.mcp_server.server._prepare_mcp_server_headers",
return_value=(None, None),
),
patch(
patch( # test-quality-ok: manager is a module-level singleton; patching it is the suite established seam
"litellm.proxy._experimental.mcp_server.server.global_mcp_server_manager",
) as mock_manager,
patch(
patch( # test-quality-ok: tool filter is a module-level function; the suite has no injection seam
"litellm.proxy._experimental.mcp_server.server.filter_tools_by_allowed_tools",
side_effect=lambda tools, _server: tools,
),
patch(
patch( # test-quality-ok: permission filter is a module-level async function; the suite has no injection seam
"litellm.proxy._experimental.mcp_server.server.filter_tools_by_key_team_permissions",
new=AsyncMock(side_effect=lambda tools, **_: tools),
),
patch(
patch( # test-quality-ok: logging setup is a module-level function; patched to capture spend-log metadata kwargs
"litellm.proxy._experimental.mcp_server.server.function_setup",
side_effect=_capture_function_setup,
),
@ -5141,8 +5142,22 @@ async def test_get_tools_from_mcp_servers_prefers_explicit_request_tags_over_the
list_tools_log_source="mcp_protocol",
request_tags=["explicit"],
)
explicit_metadata = dict(function_setup_kwargs["metadata"])
assert function_setup_kwargs["metadata"]["tags"] == ["explicit"]
await _get_tools_from_mcp_servers(
user_api_key_auth=user_auth,
mcp_auth_header=None,
mcp_servers=["server_a"],
mcp_server_auth_headers=None,
raw_headers={"x-litellm-tags": "from-header"},
log_list_tools_to_spendlogs=True,
list_tools_log_source="mcp_protocol",
request_tags=[],
)
empty_metadata = dict(function_setup_kwargs["metadata"])
assert explicit_metadata["tags"] == ["explicit"]
assert "tags" not in empty_metadata
@pytest.mark.parametrize(

2
ui/litellm-dashboard/src/lib/http/schema.d.ts generated vendored
View file

@ -16781,7 +16781,6 @@ export interface paths {
* - permissions: Optional[dict] - [Not Implemented Yet] User-specific permissions, eg. turning off pii masking.
* - metadata: Optional[dict] - Metadata for user, store information for user. Example metadata = {"team": "core-infra", "app": "app2", "email": "ishaan@berri.ai" }
* - max_parallel_requests: Optional[int] - Rate limit a user based on the number of parallel requests. Raises 429 error, if user's parallel requests > x.
* - soft_budget: Optional[float] - Get alerts when user crosses given budget, doesn't block requests.
* - model_max_budget: Optional[dict] - Model-specific max budget for user. [Docs](https://docs.litellm.ai/docs/proxy/users#add-model-specific-budgets-to-keys)
* - budget_fallbacks: Optional[Dict[str, List[str]]] - Per-model fallback chain tried in order when that model's own `model_max_budget` is exceeded, e.g. {"gpt-4o": ["gpt-4o-mini"]}.
* - model_rpm_limit: Optional[float] - Model-specific rpm limit for user. [Docs](https://docs.litellm.ai/docs/proxy/users#add-model-specific-limits-to-keys)
@ -16887,7 +16886,6 @@ export interface paths {
* - permissions: Optional[dict] - [Not Implemented Yet] User-specific permissions, eg. turning off pii masking.
* - metadata: Optional[dict] - Metadata for user, store information for user. Example metadata = {"team": "core-infra", "app": "app2", "email": "ishaan@berri.ai" }
* - max_parallel_requests: Optional[int] - Rate limit a user based on the number of parallel requests. Raises 429 error, if user's parallel requests > x.
* - soft_budget: Optional[float] - Get alerts when user crosses given budget, doesn't block requests.
* - model_max_budget: Optional[dict] - Model-specific max budget for user. [Docs](https://docs.litellm.ai/docs/proxy/users#add-model-specific-budgets-to-keys)
* - budget_fallbacks: Optional[Dict[str, List[str]]] - Per-model fallback chain tried in order when that model's own `model_max_budget` is exceeded, e.g. {"gpt-4o": ["gpt-4o-mini"]}.
* - model_rpm_limit: Optional[float] - Model-specific rpm limit for user. [Docs](https://docs.litellm.ai/docs/proxy/users#add-model-specific-limits-to-keys)