* feat(skills): semantic search over the LiteLLM-hosted skill registry
Adds GET /v1/skills?query= (custom_llm_provider=litellm_proxy) and a
skill_search MCP virtual tool, ranking the caller's accessible skills by
semantic similarity, mirroring the A2A agent registry search (LIT-6309).
Also fixes a pre-existing bug where create_skill() dropped description and
instructions for the litellm_proxy provider, which left every LiteLLM-hosted
skill with no searchable text.
* fix(mcp): coerce skill_search top_k instead of raising 500 on malformed input
The MCP-REST skill_search dispatch validated raw tool arguments through a
pydantic model directly, so a non-numeric top_k raised a ValidationError
that the endpoint's catch-all turned into an HTTP 500. Mirrors the
agent_search branch's tolerant coerce_top_k handling instead.
* fix(skills): enforce key limits on search embeddings and bound the semantic index
Semantic search embeddings now run the same pre_call_hook the /embeddings
route runs, so key rate limits, budgets and guardrails apply before the
embedding model is called. The shared SemanticTextIndex caps cached vectors
and evicts the least recently searched entries, and each skill's embedded
text is capped so one skill cannot inflate the embedding batch
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(skills): surface proxy 429s from search embeddings instead of a 503
ProxyRateLimitError is also an OpenAIError, so the search engine was folding
a key rate limit into skill_search_unavailable. Proxy HTTPExceptions now
propagate so the caller gets the same 429 the /embeddings route returns
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(skills): import assert_never from typing_extensions for Python 3.10
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(skills): embed the request as the pre-call hooks returned it, not the original text
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* refactor(skills): keep the litellm_proxy provider check for GET /v1/skills?query= inside llms/
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(skills): move the GET /v1/skills?query= endpoint tests under tests/test_litellm/proxy
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* feat(ui): add key-scoped auto-router usage tab
GET /auto_router/benchmarks takes an optional api_key filter, applied in the
rollup aggregate on the primary key's leading column. Proxy admins get a
separate Auto-router usage tab on key detail pages with spend, baseline,
savings, tier routing, cache metrics and the existing router selector
* fix(ui): share key analytics date range
* feat(proxy): resolve root_path per request from SERVER_ROOT_PATHS
One deployment can encode exactly one client-visible URL path prefix
today: SERVER_ROOT_PATH is a scalar stamped onto the app at startup, so
a pod fronting several ingress prefixes 404s every prefix but one before
any handler runs, and MCP OAuth discovery can emit only one prefix's
URLs (RFC 9728 section 3 exact-match fails for the rest).
Add an opt-in outermost ASGI middleware that matches the request path
against a configured prefix list (SERVER_ROOT_PATHS, comma-separated) on
a segment boundary and sets scope["root_path"] for that request only.
Everything downstream is stock Starlette: route matching strips
root_path so routes stay registered root-relative, and request.base_url
re-includes it, so the discovery documents' resource and the 401
challenges' resource_metadata land under the prefix the client actually
called — with no discovery-builder changes.
LazyFeatureMiddleware now strips the scope root_path (falling back to
the cached SERVER_ROOT_PATH scalar) before feature prefix matching, so
lazily-registered routers — the MCP OAuth discovery router among them —
load under per-request prefixes.
Follow-up to the routing discussion on #35226; composes with, but does
not depend on, #35576.
* fix(proxy): import Sequence from collections.abc (ruff UP035 strict-budget gate)
* review(greptile): trim implementation commentary; fixture-own MCP registry state in tests
Addresses both P2s from the first Greptile pass:
- per_request_root_path_middleware.py (and the related _lazy_features /
proxy_server comments) cut down to the constraints the code cannot
express, per repo comment guidance
- the new discovery tests no longer clear/repopulate the shared MCP
registry inline; a fixture snapshots it, hands the test an empty
registry, and restores it afterwards so no state leaks between cases
* fix(lint): mutable-ok marker on the prefix accumulator (LIT002 type-discipline gate)
* fix(proxy): tie 401 challenges and get_custom_url to the per-request root_path
The per-request root_path middleware sets scope["root_path"] to the
prefix the client actually called, but the OAuth 401 challenges
(raise_user_oauth_challenge / raise_token_exchange_challenge) still
built their resource_metadata from SERVER_ROOT_PATH. On a pod fronting
several prefixes, the challenge advertised a discovery URL under a
different prefix than the discovery document served — the two
disagreed on where the resource metadata lives, and a strict RFC 9728
client refused the challenge. Route the challenges through a small
ContextVar the middleware populates so they read the same effective
root_path Starlette resolves the request under.
The same accessor fixes get_custom_url: when a request lives under a
SERVER_ROOT_PATHS-matched prefix, request.base_url already carries it,
so appending the SERVER_ROOT_PATH scalar on top produced e.g.
/tenant-a/legacy/sso/callback — a path that does not exist. Reading
the per-request prefix instead (and relying on join_paths's tail-dedup)
keeps SSO login/callback URLs under one prefix — the one the request
actually arrived on.
Fallback: outside a request (module-load-time UI URL builders,
background tasks) the ContextVar is unset and the accessor reads
SERVER_ROOT_PATH, matching get_server_root_path() so scalar-only
deployments are byte-identical.
* fix(mcp): challenge URL under per-request prefix must route, and mock parity
Two follow-ups to the review fix that made the 401 challenge use the
per-request root_path:
1. oauth_protected_resource_path must pick the URL structure that
actually routes for the mechanism in use:
- The scalar SERVER_ROOT_PATH deployment registers the well-known
routes with the prefix INSERTED (via well_known_root_suffix at
import time), matching RFC 8414 §3. The challenge URL must use the
same insertion or a client fetching it 404s.
- The per-request SERVER_ROOT_PATHS deployment can't register routes
per prefix; PerRequestRootPathMiddleware strips the prefix from
scope["path"] and the router matches the un-inserted route. The
URL must place the prefix BEFORE .well-known so the strip leaves a
matching path.
The previous fix used the insertion form for both, which 404'd the
discovery fetch on the per-request path — the discovery doc and the
challenge would then disagree on where the resource metadata lives,
the very failure the review flagged. End-to-end verified: the URL
the challenge advertises routes and the doc's `resource` field
equals the URL the client originally called (RFC 9728 §3).
2. get_request_root_path now delegates its fallback through
get_server_root_path() instead of reading the env directly, so every
existing `monkeypatch.setattr("litellm.proxy.utils.get_server_root_path"`
test override keeps working. This unstubbed the mock on the /v2/login
test that failed on the last CI run.
Plus the lint budget: annotate the local accumulator Final, tag the
scope["root_path"] rewrite as an intentional ASGI-contract mutation,
tag the reused `path`/`root_path` rebinds in LazyFeatureMiddleware, and
add reason strings to the two new PLC0415 lazy-import noqas.
* test(mcp): pin the reviewer's expected end-state — challenge URL routes, resource matches called URL
End-to-end regression test that mounts the discoverable router + the
per-request root_path middleware, hits an MCP endpoint that raises
raise_user_oauth_challenge, fetches the resource_metadata URL the
challenge advertises, and checks the returned document's `resource`
equals the URL the client originally called (RFC 9728 §3 exact match).
Covers /tenant-a, /tenant-b, and the unprefixed path on the same app so
a regression on any prefix — challenge URL 404s, or doc emits a
different prefix than the client called — fails at this test rather
than in a strict MCP client's discovery.
---------
Co-authored-by: gym-cmd <186399764+gym-cmd@users.noreply.github.com>
* fix(spend_logs): keep partition DDL transactions alive for their statement timeout
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* style: ruff format changed files
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(lint): avoid dict-literal kwargs and keep cast-ok on the cast line
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(lint): cast at the call site instead of widening PrismaClient.tx
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(spend_logs): require partition tx timeout to strictly exceed statement bound
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: jesus <jesus@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(mcp): apply key and team guardrails to MCP tool calls
Guardrails attached to a virtual key or team were only enforced on LLM
routes. The synthetic request built for MCP tool call guardrail hooks
carried no guardrails in its metadata, so a guardrail with default_on
false never ran on tools/call even when the key explicitly listed it.
Resolve key, team, and project guardrails onto the synthetic request
with the same helper the chat path uses.
* fix(mcp): pass project metadata through without a mutable default
* fix(mcp): mark the request dict parameter mutable-ok with a reason
* test(mcp): explain the premium_user patch and tighten the helper docstring
The router code coverage gate reads every function defined in router.py
and fails when no test file names it. _as_retry_skipped_deployment_ids
was only reached indirectly through the retry path, so the gate went red
on this PR's tip.
Test it directly instead: a tuple of strings survives, non-string items
inside the tuple are dropped, and every other shape a caller could send
narrows to an empty skip list.
SPEND_LOGS_URL only diverts spend logs when db_writer_client is set, and nothing in the proxy ever assigns that global, so the queued copy was only ever skipped as a duplicate by the local insert.
The retry skip travels as a request kwarg, and the router forwards keys it
does not recognize, so a client can put _retry_skipped_deployment_ids in its
own request body. The value went straight into a pydantic TypeAdapter and
then into a set(), so an int or an object raised TypeError and a string, a
list, or a dict raised a ValidationError, each of them replacing the 400 the
provider had actually returned.
Every read now goes through one narrowing function that keeps a tuple of
strings and skips nothing otherwise, so a forged value costs the caller
nothing beyond the retry landing on the same deployment again.
disable_spend_logs has to keep meaning that no request gets logged, and the row
that makes a batch chargeable exactly once is the one row it cannot drop, so with
logging off that row now carries only what tells the retrieves apart. SPEND_LOGS_URL
deployments get their copy back too: the claim writes straight to this table, so the
row is queued as well when an external writer is the one that takes the spend logs.
A fake-streamed provider hands the adapter one chunk carrying both the
delta payload and the finish_reason, which is exactly what the combined
chunk splitter exists for, but its content check never listed the refusal.
The translation short-circuits on finish_reason, so that refusal text was
dropped and the client got `stop_reason: refusal` over an empty content
array, the symptom this PR set out to fix.
Both refusal accumulators also drop their `mutable-ok` lists for a plain
string attribute
Before excluding the deployment that just refused, the retry-skip guard asked
whether another one could still answer. It asked by re-running a single routing
filter, the order filter, while deployment selection also applies cooldowns, the
context-window pre-call check, tag routing, and routing plugins.
Any filter the guard did not replicate made it answer yes while the real pick was
left with nothing. A group narrowed to one deployment by tag routing turned the
provider's own 400 into a no-deployments 429.
The skip now runs where every filter has already been applied, and it keeps the
deployments untouched when skipping would leave none. The caller gets the
provider's error either way, and a group with one eligible deployment retries in
place as it did before.
The first-delta guard read `delta.refusal` directly, while the translation
three lines later goes through `openai_chat_refusal_text`, which also reads
the `provider_specific_fields` LiteLLM parks unrecognized fields in. A
provider that sends the refusal that way had its only refusal delta skipped
as blank, so the client got `stop_reason: refusal` over an empty content
array, which is the symptom this PR set out to fix
The /v1/messages adapter lowers a tier the entry does not accept, so dropping max from the astra
rows moves that path from Foundry's 400 to a request at xhigh. Nothing pinned that, and the guard
test's docstring named gpt-6-astra as the only gpt-5 name with an azure_ai row, which 11 rows
contradict.
The retry-skip guard checks that some other deployment could still answer
before it excludes the one that just refused, so a single-deployment group
keeps the old retry-in-place behavior. It asked that question at the group's
minimum order, but the router picks the retry's deployment at the order the
request has already escalated to.
So a group with a primary at order 1 and a backup at order 2 answered "yes,
order 1 still has a candidate" while the retry was pinned to order 2, and the
exclusion left order 2 with nothing. The caller got a no-deployments error in
place of the provider's own 400.
The helper now takes the active target order and filters by it, which is the
same value async_get_healthy_deployments reads off the request.
`vertex_ai/lyria-3-clip-preview` and `vertex_ai/lyria-3-pro-preview` were
registered with `supports_vision`, `supports_image_input`, and an `image`
modality, which contradicts their `gemini/lyria-3-*` siblings and makes
/model/info advertise image input on text-to-music models.
The new Lyria passthrough branch runs before the image-generation branch
and keys on the same `predictions[0].bytesBase64Encoded` shape imagen
returns, so only the cost-map lookup separates them. Cover an imagen
predict response end to end so a future change that drops that lookup
fails here instead of misbilling images as audio.
The takeover of a $0 row an older proxy left behind used to charge the batch when
the update could not reach the database. That leaves the row still reading $0, so
every later retrieve finds the same row and charges the batch again, which is the
repeat charging this PR exists to stop. The retrieve that does take the row over
is the one that charges, and a batch nobody retrieves again after that failure is
never charged, the same as one whose proxy died inside the write window.
CI's router_code_coverage gate wants every function in router.py called by
name from a test file with "router" in its name, and the new helper had no
direct caller, so the check-quality job failed on the first tip.
Covering it directly also turned up a hole. litellm._should_retry compares
the status code to 500, so a provider exception carrying a string status code
raises TypeError instead of answering. should_retry_this_error has the same
call, but the retry policy path skips it, which is exactly the path this
change enables, so the helper was the first to touch that value. Narrowing to
int leaves those exceptions on the old retry-in-place behavior.
Google prices Lyria per generated clip, so every Vertex Lyria entry in the
price map now carries a single output_cost_per_image and both the speech
and the passthrough cost paths read that one field. The old
output_cost_per_second and audio_seconds_per_prediction pair assumed a
30 second clip, which does not match the 32.768 second WAV Vertex returns,
and no other model in the map priced audio that way
Drops max_audio_length_hours and max_audio_per_prompt from the price map,
its schema, the generator, and ModelInfo, since nothing reads them, and
drops the audio_mime_type hidden param for the same reason: the response
already carries the resolved content type on its own header
Folds the per-model bundled catalog lookups into one cached parse of the
local cost map, validated with a TypeAdapter over a ReadOnly TypedDict
A proxy without this fix wrote the batch's cost row on every poll while the batch
was still running, so that row reads $0 and the insert that claims the charge has
nowhere to land. The retrieve that charges the batch now writes its own payload
over that row under a where clause that still names spend 0.0, so exactly one
retrieve takes it over and every later one reads the charge and charges nothing
gpt-6-astra is the only gpt-5-family name with an azure_ai row. Prefixing the rest
cost them every effort flag, since get_llm_provider sends an azure_ai name down the
azure provider when a global AZURE_AI_API_BASE points at an openai.azure.com host and
azure/<model> is not a key either, which turned temperature, top_p and logprobs on
azure_ai/gpt-5.1-chat-latest from accepted into an UnsupportedParamsError.
