mirror of
https://github.com/diegosouzapw/OmniRoute.git
synced 2026-09-20 13:52:28 +03:00
2af688a94f3becba669f19c1cc6ef37ca04ea9f3
1627 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Tiangao
|
fda9ef78b1 |
feat(proxylogs): show registry proxy name in proxy log columns (#12814)
* feat(proxylogs): show registry proxy name in proxy log columns Registry resolution already attaches name to the runtime proxy object; the name was dropped at the persistence boundary (ProxyInfo had no name field) and never rendered. Add proxy_name column (base schema + ALTER heal for existing DBs), persist/hydrate it, render it in the ProxyLogger table and ProxyLogDetail pane with host:port fallback, and search by name. Local-only (PMO City): not submitted upstream. Re-apply after upgrades via patch file (see pmo-city-builds omniroute/Operator/runbooks/upgrade.md). * test(proxylogs): flush batched writes before asserting persisted row The v3.8.50 rebase kept upstream's batched proxy-log persistence (enqueueProxyLogs/flushProxyLogsSync); logProxyEvent no longer writes synchronously, so the persist+hydrate test closed the DB before the row was flushed. Flush explicitly first. * test(proxylogs): drain batched queue in resetStorage to avoid cross-test row bleed * docs(changelog): add changelog fragment for proxy registry name in proxy logs Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: Tiangao (hermes) <montigaud@aikumi.pro> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Davide Baraldo
|
1b2349de22 |
feat(sse): Claude OAuth lower-priority lane + weekly session-limit reset (#13074)
* feat(sse): Claude OAuth lower-priority lane + weekly session-limit reset
Mirror Claude Code's /low-priority and /limit-reset for OmniRoute-managed
Claude subscription accounts (wire contract captured from Claude Code 2.1.263).
Both are opt-in per connection (providerSpecificData.lowPriorityMode /
autoLimitReset, Edit connection -> Claude section, default off) and only act
on the 5-hour usage wall: a 429 carrying
anthropic-ratelimit-unified-status: rejected and, when eligible,
anthropic-ratelimit-unified-slow-offer: treatment. Nothing is sent before
that first wall 429.
- Lower-priority lane: on the wall the executor retries the SAME account
with `anthropic-usage-limit: slow` and keeps the header on every request
until anthropic-ratelimit-unified-reset (+60s). The intercepted 429 never
reaches chatCore, so the connection is not cooled down or rotated away.
slot_busy (429) / 529 wait slow-retry-after (20s default, 5-600s, +-30%
jitter) bounded by slow-max-wait (20min default, 1min-6h), then end +
10min cool-off. weekly_limit / budget_exhausted / off / ineligible, a
5h-window rollover, or ineligible + overage-in-use end the lane and let
the response flow to the normal cooldown path.
- Session-limit reset: GET /api/oauth/usage?at_wall=1&skip_spend=1 ->
juniper_tide block; when arm=reset and available, POST
/api/organizations/{org}/reset_rate_limits {program: "juniper_tide"} and
retry at full speed. already_used / not offered memoise next_available_at.
- State is in-memory per connection; the executor owns the abort-aware
sleep; the pure state machine and the HTTP client are separate modules
with unit tests; an executor-level test proves the header/retry wiring
end to end with a mocked upstream.
* fix(sse): make the Claude usage-wall handling race-safe for parallel requests
Two requests on the same Claude OAuth connection can hit the 5-hour wall in
the same instant.
- Lower-priority lane: the executor now tells the decider whether THIS
request carried `anthropic-usage-limit: slow`. A sibling built while the
lane was still idle (no header) whose 429 lands after the lane activated
is re-sent on the lane instead of being misread as a "wall" verdict that
would end it; its 2xx is not counted as lane telemetry either.
- Session-limit reset: concurrent wall hits share one in-flight status+claim
round trip (no duplicate POST reset_rate_limits), and for 60s after a
granted reset stale sibling walls are answered "reset" without touching
the network, so they retry at full speed instead of re-claiming or
falling into the slow lane.
Tests cover both races.
* fix(sse): address adversarial review of the Claude usage-wall handling
Three defects found by a 3-lens review of the two previous commits.
1. Lane wait could outlive the request (high). The slot_busy/529 sleep shares
the request's AbortSignal with chatCore's upstream-start timeout (10 min by
default), while the lane's own max-wait defaults to 20 min and can reach 6h
from the server header. A long slot_busy streak was therefore killed
mid-sleep with a TimeoutError instead of ending gracefully as max_wait with
its cool-off. The decision now takes a waitCeilingMs — what is left of the
executor's own timeout, minus a 5s margin — which caps the effective
max-wait and clamps each individual sleep.
2. A wall 429 surfacing only after a 400-driven intra-attempt retry was missed
(medium). The context-editing / thinking-budget / effort / auto-learn
fallbacks all re-fetch and REASSIGN `response`, and the wall check ran
before them, so such a 429 fell through to the generic path and cooled the
connection down. The check now runs after those retries, on the final
response of the attempt.
3. `ineligible` + `overage-in-use: true` ended the lane as plain `ineligible`
on a 429 (medium) because the status mapping ran first; only the non-429
tail produced `extra_usage`. Overage takeover now wins on every status.
Also bounds the module-level per-connection maps with the same FIFO policy as
the identity caches in claudeIdentity.ts: the state key falls back to the
access token when a connection id is absent, and OAuth tokens rotate on every
refresh, so the maps could grow for the process lifetime.
Tests cover all three fixes, including an executor-level regression for the
400-then-wall ordering.
* fix(i18n): add the Claude usage-wall toggle strings to pt-BR
`tests/unit/i18n-pt-br.test.ts` (#6695) requires pt-BR.json to carry every key
present in en.json; the four new `providers.claude{LowPriorityMode,AutoLimitReset}*`
keys were only added to en and it, so the gate failed on this branch.
* refactor(sse): keep the usage-wall change inside the frozen quality budgets
The three ratchets this PR tripped were all its own, not inherited:
- file-size (frozen, may only shrink): open-sse/executors/base.ts 1857 > 1751
and EditConnectionModal.tsx 1653 > 1631.
- complexity / cognitive-complexity (new-code mode): three functions over the
15 threshold — runClaudeLimitResetAttempt (27), handleClaudeUsageLimitResponse
(19 / cognitive 23) and observeClaudeLowPriorityResponse (17 / 17).
Extractions, all behavior-preserving:
- New open-sse/executors/claudeUsageLimit.ts owns the executor-side glue (header
injection, wait accounting, abort-aware sleep, timeout-derived wait ceiling and
the decision logging) behind a ClaudeUsageLimitGuard, so base.ts keeps a
three-line call site instead of ~100 lines of mechanics.
- Three long-standing Claude blocks leave base.ts for the modules they belong to:
mergeCcHeaders + applyStainlessHeaders into config/anthropicHeaders.ts and
stripClaudeSystemPrefixBlocks into executors/claudeIdentity.ts. base.ts is back
at its frozen 1750 lines.
- The modal's Claude section becomes ClaudeConnectionFields.tsx (mirroring
CcCompatibleRequestDefaultsFields) plus a claudeConnectionFields.ts helper that
de-duplicates the field defaults across the modal's two init sites; the file
drops to 1622, below its frozen 1631.
- The three over-threshold functions are split into focused helpers
(observeErrorResponse / observeSuccessResponse, shouldClaimLimitReset,
resolveLimitResetOffer / runLimitResetClaim / memoiseNotBefore).
Gates now: file-size OK, complexity 0 new violations, cognitive 0 new,
fetch-targets / error-helper / build-scope / deps OK, typecheck clean, ESLint 0,
Prettier clean, 155 unit tests green across the feature and its neighbours.
Still failing and NOT this branch's: pack-policy (unexpected
@omniroute/opencode-plugin-v2 files in the npm artifact) and
mutation-test-coverage (stryker tap.testFiles missing entries for
circuitBreaker.ts and comboStructure.ts) — both reproduce on the untouched base.
---------
Co-authored-by: davidebaraldo <davidebaraldo@users.noreply.github.com>
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
|
||
|
Sean Ford
|
0fbd8854c5 |
fix(api): adapt TEI/Infinity request and response shapes on the /v1/rerank node path (#13733)
* feat(api): route /v1/rerank to remote provider nodes behind RERANK_REMOTE_PROVIDER_NODES POST /v1/rerank only ever dispatched to provider nodes whose base URL hostname was localhost, 127.0.0.1, or 172.16.0.0/12 — a filter hardcoded in the route. A rerank node on any other host (a LAN box or Tailscale peer running TEI, Infinity, vLLM, …) was silently dropped and the request fell through to "Invalid rerank model", even though the same node served /v1/embeddings without complaint and had already passed the provider outbound URL policy at creation time. The memory engine's rerank step calls this route over loopback, so `rerankProviderModel` could not reach such a node either. Mirror the audio routes (#3963): loopback nodes stay always-eligible and unchanged; remote nodes are opt-in via a new `RERANK_REMOTE_PROVIDER_NODES` feature flag (default off — routing to a remote host changes egress identity) AND must pass the provider outbound URL policy (`getProviderOutboundGuard()`, `public-only` deployments never route to private hosts. - src/shared/network/loopbackNodeHost.ts: one pure definition of the loopback host set, replacing three copies (rerank route, audioRegistry, localHealthCheck). The shared version also rejects `user@host` URLs, which the audio copy did not. - src/shared/network/providerNodeHost.ts: policy-aware remote-node eligibility that mirrors guardProviderNodeBaseUrl() on the creation path. - src/app/api/v1/_shared/rerankProviderNodes.ts: pure, testable selection step + loader, modelled on audioProviderNodes.ts. - Feature flag definition, FEATURE_FLAGS.md / ENVIRONMENT.md / .env.example rows, API_REFERENCE.md and MEMORY.md notes, changelog fragment. - tests/unit/rerank-remote-provider-nodes.test.ts covers the host classification, the three policy modes, the selection step, and the route end-to-end (flag off → 400 without contacting the node; flag on → forwarded to <base>/v1/rerank with the node credential; flag on + strict policy → still excluded). Feature-flag count test bumped to 56. * chore(changelog): name the #13732 fragment * fix(api): adapt TEI/Infinity request and response shapes on the /v1/rerank node path The provider-node branch of POST /v1/rerank already fell back from <base>/v1/rerank to <base>/rerank on 404 "for Infinity / TEI", but it kept sending the Cohere body and returned the upstream JSON verbatim. Against Hugging Face text-embeddings-inference that could never work: TEI requires the candidate list as `texts` (HTTP 422 otherwise), takes `return_text`, and answers a bare `[{index, score, text?}]` with no `results` envelope and `score` instead of `relevance_score`. Thin gateways in front of TEI/Infinity commonly emit `score` too. Either way the memory engine's applyRerank(), which reads `results[].relevance_score`, ended up with undefined scores. Add two pure adapters in src/app/api/v1/_shared/rerankLocalNodeShapes.ts: - buildLocalRerankRequestBody(): one upstream body carrying both spellings (`documents` + `texts`, `return_documents` + `return_text`). TEI's request struct is not deny_unknown_fields and the OpenAI-shaped servers (vLLM, llama.cpp, Infinity, oMLX) ignore extras, so a single body serves all. - normalizeLocalRerankResponse(): folds `{results:[…]}`, Voyage-style `{data:[…]}`, and TEI's bare array into the Cohere envelope, backfilling `relevance_score` from `score`, sorting by score, honouring `top_n`, attaching `document.text` when requested, dropping malformed entries, and preserving other top-level fields (`model`, `usage`, …). The route now uses both on the primary and fallback fetch. Cloud registry providers are untouched (they go through open-sse/handlers/rerank.ts). tests/unit/rerank-local-node-shapes.test.ts covers the adapters and the route end-to-end: 404 → /rerank with `texts`, bare TEI array normalized and top_n-capped; `score`-only gateway → `relevance_score` for the client. * chore(changelog): name the #13733 fragment * refactor(api): split the local rerank response normalizer into per-entry helpers The complexity ratchet (new-code mode) flagged normalizeLocalRerankResponse at 18/15 on both metrics. Pull the per-entry validation and the document resolution into toCohereResult() / resolveResultDocument(); behaviour and tests are unchanged. --------- Co-authored-by: seanford <seanford@users.noreply.github.com> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
birdleandro-bit
|
1e8c913ca2 |
feat(services): add open-wa as a 6th embedded service (#13222)
* feat(services): add open-wa as a 6th embedded service Adds @open-wa/wa-automate (WhatsApp Web automation via headless Chromium) following the existing embedded-service pattern, mirroring Mux's lifecycle-managed-only shape (no Layer 4 executor — this is not a routing target). Flags/env verified directly against the installed 4.76.0 source rather than trusted from web docs, which mix this stable v4 line with an unreleased v5 alpha CLI surface. healthIntervalMs is set to 60s (vs. the usual 5s) for this service: open-wa's HTTP server does not start listening until the full WhatsApp handshake resolves, which blocks on a human scanning the pairing QR code on first pairing. At the default 5s interval the supervisor's 3-consecutive-failure threshold would declare "error" ~15s into every legitimate start. This is a local, single-service config change — a proper fix (a startup-grace knob distinct from the steady-state poll interval) belongs in ServiceSupervisor/HealthChecker as a follow-up affecting all embedded services. * fix(db): renumber open-wa seed migration to avoid collision with 163 Migration 163 was already taken by 163_radar_feed_cache_generated_at.sql on release/v3.8.51 (tip is at 179). Renumbered to 180, the next free slot. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> * fix(db): renumber open-wa seed migration to avoid collision with 180 Slot 180 was reused by 180_memory_fts_au_conditional_memory_id.sql (merged 2026-09-16), so this PR's seed migration moves to the owner's assigned slot 185. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> Co-authored-by: birdleandro-bit <birdleandro-bit@users.noreply.github.com> |
||
|
voidstack
|
e400cf9ac7 |
fix(db): resolve backup retention from persisted setting on health-check path (#13773)
* fix(db): resolve backup retention from persisted setting on health-check path #13404 fixed the missing prune call after health-check-repair backups but only resolved maxFiles/retentionDays from env vars, so the persisted Storage-page setting (honored for manual/API/auto backups via getDbBackupMaxFiles/getDbBackupRetentionDays) was silently ignored on this path. Extract that env->persisted->default precedence into resolveDbBackupRetention() in backupRetention.ts and share it between backup.ts and core.ts's createManagedDbBackup(). * docs: add changelog fragment for #13308 persisted-setting follow-up * fix(db): re-point backup retention fix at managedBackup.ts's prune call The base drifted since this branch was opened: the health-check-repair backup path (createManagedDbBackup) moved from core.ts into managedBackup.ts (writeManagedDbBackup), taking its env-only maxFiles/retentionDays resolution along with it. This branch's resolveDbBackupRetention() extraction and backup.ts delegation were already correct and unaffected; only the wiring that used to live in core.ts needed to move to managedBackup.ts's prune call so the persisted Storage-page setting is honored on this path too (#13308). Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
jbovard2016
|
d574826f37 |
fix(usage): detach completed request previews (#13623)
* fix(usage): bound completed request retention Completed request previews used V8 sliced strings that kept multi-megabyte request backing stores alive. Detach and byte-bound cached details, and add a credential-free profiler with cleanup and physical-retention assertions. * fix(usage): give the JON-562 memory-profile canary realistic timeouts The 100k-token worker step alone takes ~230s (tsx/esm boot of the full route/handler module graph plus the real request lifecycle), well past the driver's hardcoded 180s spawnSync timeout — the resulting SIGKILL surfaces as `worker.status === null`, indistinguishable from a real crash. Bump the worker timeout to 300s and the test's own outer/inner timeouts to match the measured ~230-330s real runtime. Also make git-branch provenance detached-HEAD safe: `git branch --show-current` is empty on a detached HEAD (the normal state for a CI PR checkout, and for this fix worktree itself), which made the canary throw "git branch is empty" deterministically outside a regular branch checkout. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Ryan
|
70eebe9adb |
fix(arena+analytics): atomic ELO sync (fetch-first) + flatRateAsZero in compression writer (#13446)
* fix(analytics+arena): flatRateAsZero in compression writer; atomic arena sync redesign * docs(changelog): add fragments for arena ELO sync and compression flat-rate fixes Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: CrashCartCapital <crashcartcapital@gmail.com> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Domenico Massafra
|
fbe195d69e |
fix(opencode): preserve catalog display names (#13168)
* fix(opencode): preserve catalog display names * test(cli): cover OpenCode catalog display-name precedence Adds the automated unit test the PR body's manual smoke check (Auto Chat / DeepSeek V4 Pro) was standing in for, covering all four name-precedence branches: existing custom name, catalog display_name, native catalog name with owned_by prefix stripped, and the auto/* readable fallback. Also adds the changelog.d/fixes/ fragment. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: ginettododo <117327638+ginettododo@users.noreply.github.com> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
KeelTrace
|
95d3b164a5 |
fix(models): preserve free-model metadata from discovery (#12763)
* fix(models): preserve live free economics in synced discovery * docs(changelog): add fragment for free-model metadata discovery fix Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Kareem Jalal
|
ad633c8440 |
fix(memory): word/sentence-boundary aware fact truncation (#12383)
* fix(memory): word/sentence-boundary aware truncation in extraction sanitizeMatch() and capExtractionText() previously did raw character-offset slices (slice(0, MAX_FACT_LENGTH) / slice(-MAX_EXTRACTION_TEXT_LENGTH)) with no boundary awareness, producing garbled mid-word/mid-clause fragments that get injected into LLM context as memory facts. - sanitizeMatch() now backs the cut off to the nearest sentence-ending punctuation (. ! ?) within a lookback window, falling back to a plain whitespace boundary, falling back to the original hard cut only when no boundary exists nearby. - capExtractionText() applies the equivalent boundary-aware trim on the front edge of the kept tail. Mirrors the boundary-aware truncation pattern already used by open-sse/services/compression/lite.ts (#8169) for tool-result truncation. Adds tests/unit/memory-extraction-boundary-truncation.test.ts covering word-boundary cuts, sentence-boundary preference, short-string passthrough, the no-boundary-available fallback, and capExtractionText's tail behavior. * docs(changelog): add fragment for word/sentence-boundary fact truncation Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Paijo
|
023a57476f |
feat(chat-admission): expose admission tunables via dashboard settings (#12038)
* feat(chat-admission): add settings store for admission tunables * fix(chat-admission): extract parseEnvNumber to reduce cyclomatic complexity * fix(chat-admission): repair settings store write path and add coverage The settings store could not persist anything: `updateChatAdmissionSettings` targeted an `updated_at` column that `key_value` does not have (the schema is namespace/key/value — src/lib/db/core.ts), so every write threw `table key_value has no column named updated_at`. Also fixes, found while adding the tests: - `getChatAdmissionSettingsSource` returned a partial map (only the keys whose layer differed from the default) and dropped the unset keys entirely, so a dashboard reading it could not render a complete row. - env parsing used `parseFloat` for the shed ratio, so `"0.5x"` was silently accepted as 0.5 while `chatBodyAdmission.ts` rejects that same input — both paths now share one per-field predicate table. - DB reads validated `typeof === "number"` but not integrality/range, so a hand-edited row could serve `2.5` or `-1` to the admission controller. - writes persisted unvalidated input. - malformed, non-object, and partial rows are now tolerated per field. Adds tests/unit/db-chat-admission-settings.test.ts (17 cases) covering CRUD round-trips, namespace isolation, reset, env parsing/validation boundaries, env-over-DB precedence, provenance, normalization on write, and malformed-row tolerance, per Hard Rule #8. Verification: eslint clean; `npm run typecheck:core` clean; the new suite plus the two sibling settings suites pass 63/63; check-complexity-ratchets reports complexityNewCode=0; check-db-rules OK; check-env-doc-sync OK (all three vars are already documented in .env.example). --------- Co-authored-by: oyi77 <oyi77@users.noreply.github.com> |
||
|
voidstack
|
c07cebbab7 |
fix(db): close failed initialization connections (#13342)
* fix(db): close failed initialization connections * docs: add changelog fragment for #13303 db handle-leak fix --------- Co-authored-by: voidstackloop <voidstackloop@users.noreply.github.com> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Amirreza Kimiyaei
|
04cc8aab67 |
fix(cache): fold the response output contract into the semantic cache signature (#12309)
* fix(cache): fold the response output contract into the semantic cache signature
The signature hashed only {model, messages, temperature, top_p}, so two temp=0
requests with identical messages but different response_format shared a cache
key: the second was served the first's stored body under a 200, violating the
schema it asked for. tools/tool_choice had the same exposure.
generateSignature now takes an optional output contract — response_format,
text.format, tools, tool_choice, collected by outputContractOf() — and folds it
into the digest only when present, so plain-chat signatures (and every cache
entry already written for them) are unchanged. All three call sites pass it;
read/write symmetry is preserved because bodyForCacheWrite snapshots the same
body object the read path hashed (#cache-signature-asymmetry).
Closes #12307
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* docs(changelog): add fragment for #12307 semantic-cache output-contract fix
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
* fix(cache): populate both constraint spellings in outputContractOf
The merge with #12734 left generateSignature reading the camelCase
constraints (toolChoice/responseFormat) with a snake_case fallback, but
outputContractOf only filled the snake_case keys, so the #12734
"signature is called with tool_choice/tools/response_format from body"
store tests failed on the merged branch. Set both spellings so either
caller shape reads the value it expects.
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
Co-authored-by: amirrezakm <amirrezakm@users.noreply.github.com>
|
||
|
ducphamtien-fonos
|
128f06d645 |
fix(translator): support Responses custom tool choice (#13128)
* fix(translator): support Responses custom tool choice * fix(translator): preserve custom tools across response paths * docs(changelog): add fragment for Responses custom tool choice fix Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: Pham Tien Duc <phamtienduceng@gmail.com> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> Co-authored-by: ducphamtien-fonos <ducphamtien-fonos@users.noreply.github.com> |
||
|
James
|
3ebea07278 |
fix(sse): replay reasoning for Responses-API targets on plain turns and Anthropic clients (#13031)
* fix(sse): replay reasoning for Responses-API targets on plain turns and Anthropic clients
DeepSeek thinking mode requires the reasoning of every prior assistant turn
to be passed back once the request carries `tools`, including turns that
made no tool call. Since #10540 routed opencode-go/deepseek-v4-* to
`/responses`, the reasoning replay cache had two gaps on Responses-API
targets, and clients that drop `reasoning_content` hit intermittent
`400 The reasoning_text in the thinking mode must be passed back`.
1. Plain (non-tool-call) turns are keyed on a digest of the normalized
OpenAI transcript. Both capture sites used `translatedBody.messages` as
the history, which a Responses body (`input`) does not carry, so the
write-time digest never matched the read side. translateRequest now
reports the pivot transcript it digested via `onReasoningReplayHistory`,
and the streaming / non-streaming capture sites digest that transcript.
2. The Responses replay pass was gated on `sourceFormat === "openai"`, so
Anthropic Messages clients (Claude -> OpenAI -> Responses) got no replay
at all. The pass now runs on the OpenAI pivot for every source format,
right before the Responses conversion discards `messages`.
The reported transcript is a shallow snapshot of the digested fields only
and travels through a callback, not the body, so nothing new reaches the
upstream payload.
* docs(changelog): add fragment for #13031
* fix(sse): guard the Responses capture sites and skip plain-turn writes with no history
Review follow-ups for #13031:
- Add tests/unit/chatcore-reasoning-cache-write-guard-responses.test.ts:
runs the real handleChatCore against a mocked opencode-go/deepseek-v4-flash
Responses upstream (JSON and SSE), then asserts the next turn's upstream
body carries the replayed `reasoning` input item. Removing either capture
site fallback turns both cases red.
- Project the reported transcript down to the digested fields only
(tool_calls keep type/name/arguments, ids are dropped) and document that
`content` is shared by reference.
- Skip the plain-turn cache write when the history is empty: a real request
always has a prior user turn, so an empty history means the transcript
could not be recovered and a one-message digest can never match.
- Changelog wording: the pre-fix write digested only the assistant message.
* test(sse): select the /responses dispatch by URL in the Responses replay guard
Review follow-ups for #13031: the guard picks the upstream body by URL
(`/responses`) and asserts exactly one such dispatch per turn instead of
taking the last fetch, the streaming case asserts the same body shape as the
non-streaming one, and the `historyMessages` doc on
NonStreamingClientTranslateInput names the Responses-shaped fallback.
* docs(routing): name the replay-history hand-off without tripping the hook heuristic
The fabricated-docs gate treats any `onXxx` token in prose as a plugin hook
name and flagged `onReasoningReplayHistory` (a translateRequest option, not a
hook). Point at the option's home file instead.
* chore(quality): freeze chatCore.ts at 6159 for the Responses replay wiring
check:file-size in PR mode caps a frozen file at max(frozen, base). The rebase onto
the v3.8.51 tip (
|
||
|
Wu Shuwen
|
10bb627576 |
fix(evals): mark eval-runner requests as self-managed so cases measure the model (#13139) (#13206)
* fix(evals): mark eval-runner requests as self-managed so cases measure the model executeEvalCase() built its request with only Content-Type and Authorization, so every graded case picked up the chat path's contextual injections: a selected output style was prepended as a system message (gated on `x-omniroute-compression`) and, once the request carried an API key, retrieved memory plus the built-in `memory_*` tools were appended (gated on `x-omniroute-no-memory`). An evaluation therefore measured the operator's injected context as much as the model, and passing an API key to a run made its score worse, because the key is what gives the request a memory owner (Refs #13139). Both are documented request-header opt-outs, so the runner now sets them on every case. Request construction moves to an exported buildEvalCaseRequest() so the header contract is testable without invoking the chat route. * docs(changelog): add the eval-runner self-managed-context fragment (#13206) --------- Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
IAMBOBJIM
|
d1b26bcb62 |
fix(sse): aggregate findInsensitive collision warning into one line per build (#12972)
modelMetadataRegistry's findInsensitive() warned once per colliding key while building its lowercase index. On a real catalog that is hundreds of lines per rebuild: a production log carried 27,296 of these in a single file — 40% of all lines, in ~500/sec bursts — driving 52 MB log rotations and ~466 MB of logs on disk. The warning itself is worth keeping: a case-insensitive collision is a genuine upstream data-quality signal (models.dev returning both "OpenAI" and "openai" as distinct provider keys), and first-match-wins silently discards the later value. Only the volume was wrong. Collisions are now collected during the index build and reported as a single line carrying the total count plus the first 5 keys, so the diagnostic survives at 1/N the volume. No behavior change: the index, the first-match-wins resolution, and the WeakMap identity cache are untouched. Validated by TDD (Hard Rule #18): tests/unit/model-metadata-registry-collision-log.test.ts fails on the old implementation (3 collisions -> 3 warnings, 50 -> 50) and passes after (always 1). Also covers the no-collision case emitting nothing, and asserts the aggregated line still names colliding keys. Note for reviewers: the test fixture deliberately spells the provider key "OpenAI" rather than "openai". findInsensitive short-circuits on `if (key in obj) return obj[key]` before the index is ever built, so a fixture containing the literal lookup key produces zero warnings and proves nothing. Gates: eslint clean on both changed files. typecheck:core reports 9 pre-existing errors in open-sse/services/compression/omniglyph* — unrelated to this change (those files are byte-identical to origin/release/v3.8.50) and caused by a local stale node_modules carrying omniglyph 1.3.1 against the required ^1.4.0. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Pixma
|
01b2467d61 |
feat(sse): add LLM Gateway DevPass quota tracking (#12462)
* feat(sse): add LLM Gateway DevPass quota tracking Surface the LLM Gateway DevPass allowance (GET /v1/key) in OmniRoute's quota telemetry, mirroring the OpenRouter API-key fetcher pattern. - llmgatewayQuotaFetcher.ts: fetch + parse the DevPass /v1/key response (decimal-string USD values), exposing two windows — monthly plan credits and the 7-day premium-model window — with a 45s TTL cache. Pay-as-you-go keys (devPlan "none") and 401/403 fail open (no quota). - Register in chat.ts before registerGenericQuotaFetchers + register the named windows for the dashboard cutoff modal. - usage/llmgateway.ts leaf + usage.ts dispatch case so the Limits page renders the monthly + weekly premium rows. - Add "llmgateway" to USAGE_FETCHER_PROVIDERS, USAGE_SUPPORTED_PROVIDERS, PROVIDER_LIMITS_APIKEY_PROVIDERS, and the dashboard label/order map. - tests: 21 cases covering the parser, auth fail-open, cache TTL, window exhaustion, preflight proceed/block, registration, and the usage leaf. * docs(sse): add changelog fragment + codebase-doc entry for llmgateway quota * refactor(sse): register llmgateway quota via quotaTrackersBatch Move the LLM Gateway fetcher registration out of chat.ts (a frozen file-size-baseline chokepoint) into quotaTrackersBatch.ts, the dedicated side-effect module that exists precisely so new fetchers don't grow chat.ts. The batch import runs at module load, before registerGenericQuotaFetchers(), so the bespoke fetcher still wins over the generic path. Fixes the file-size gate (chat.ts must not grow). --------- Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Diego Rodrigues de Sa e Souza
|
f3ab24b8c7 |
fix(db): rotate proxy pools on the chat path like the registry does (#13575) (#14044)
resolveProxyForConnection cached a scope pool's first resolution result for the life of the per-connection cache, so a chat-path request never saw the pool's round-robin/sticky/random strategy advance again — only the narrow #13578 set-aside escape hatch could break the freeze. resolveProxyForScopeFromRegistry (used directly by every existing rotation test) always re-ran the strategy and rotated correctly. The cache now treats a registry-sourced pool result as due for re-resolution on every call (falling through to the same cascade the direct registry callers use), except for the two populations that need a stable egress across requests: EGRESS_BUCKETED_LOCK_PROVIDERS (opencode's quota is bucketed by egress IP) and grok-web (its cf_clearance cookie is pinned to the IP/UA/TLS fingerprint that earned it). Regression test: tests/unit/proxy-pool-chat-path-rotation-13575.test.ts, RED before the fix (resolveProxyForConnection returned the same host 6/6 times for a 3-member pool), GREEN after. Updated tests/unit/proxy-pool-skips-refused-member.test.ts's three assertions that encoded the frozen-cache contract to the corrected always-rotates-except-pinned contract; all other cases in that file and in tests/unit/proxy-pool-rotation-6365.test.ts pass unchanged. |
||
|
Diego Rodrigues de Sa e Souza
|
0d31fd3d24 |
fix(quota): resolve plan from pool's primary connection (#13876) (#14042)
Multi-connection Quota Sharing pools resolve a DIFFERENT provider plan depending on which member connection actually served a request (write path, enforceQuotaShare/recordConsumption) vs. the pool's primary connection (dashboard read path, /api/quota/pools/[id]/usage). The wizard's "Limite" step PUTs a manual plan override only to the primary connection, so any other pool member fell back to a different (catalog/empty) plan shape. Since the quota_consumption dimension key is poolId:unit:window, a different unit/window meant recordConsumption wrote to a bucket the dashboard never read, so real traffic served via a non-primary connection never appeared as "consumed". Fix: resolve the plan from the pool's canonical primary connection (pool.connectionId) in both enforceQuotaShare and recordConsumption, matching the dashboard's read path. recordConsumption now keeps the matched pool object (not just its id) so it can reach connectionId. getSaturation(input.connectionId, ...) is untouched — that signal is legitimately per-connection. |
||
|
Diego Rodrigues de Sa e Souza
|
a96c8381f7 |
fix(db): serve getPricingForModel() from the pricing cache (#13891) (#14040)
getPricingForModel() called the uncached getPricing() on every invocation instead of the existing getCachedPricing() helper (30s TTL, readCache.ts), so usageStats.getUsageStats() re-ran a 3-SELECT + JSON.parse + merge cycle against key_value once per GROUP BY row -- up to 531 times on a large usage_history table -- blocking the event loop for several seconds on /api/usage/history. Every known pricing writer (updatePricing, LiteLLM/models.dev sync) already invalidates this cache via touchPricing()/invalidateDbCache, so a write remains immediately visible; added a regression test that proves both the cache hit path and the invalidation path. |
||
|
Diego Rodrigues de Sa e Souza
|
b14ef5c7e5 |
fix(guardrails): resolve nested combo-ref hops before vision-bridge decision (#13927) (#14038)
getComboVisionBridgeDecision() treated any top-level combo-ref step as an unconditional "process", without ever resolving the referenced combo's real leaf models. A pass-through combo whose only member is a combo-ref to an all-vision-capable inner combo was wrongly routed through the describe-and-replace path, and with no describer model configured every image was replaced with the literal stub text. Recursively resolve combo-ref steps to their real leaf models (depth-guarded by the same MAX_COMBO_DEPTH used by the flatten dispatch path, plus a visited-set cycle guard) and fold their vision capability into the same accumulation used for direct model steps. An unresolvable combo-ref (not found / empty / circular / depth-exceeded) is conservatively treated as a single non-vision-capable leaf instead of forcing the whole combo to "process". |
||
|
Diego Rodrigues de Sa e Souza
|
3b535968c4 |
fix(providers): detect Lemonade labels[] vision capability (#13918) (#14023)
detectVisionInput() only recognized supportsVision, architecture.input_modalities, top-level input_modalities, and architecture/modality string shapes. Lemonade Server's GET /v1/models exposes capabilities only through a labels[] string array (e.g. ["chat", "vision", "reasoning", "tool-calling"]), so a vision-labelled Lemonade model imported with supportsVision unset and was advertised as text-only. Add a fifth branch that does a case-insensitive, trimmed EXACT membership test for "vision" in record.labels[] (not a substring match, per the prior false-positive lesson with bare gemma id-fragment matching). Purely additive - all four existing shapes stay byte-identical, proven by a new regression test that exercises the architecture.modality path unchanged. |
||
|
Dizzle
|
fc6b4587ad |
feat(proxy): support multiple local core endpoints, one per line (#13923)
Co-authored-by: Max <maxmad64@gmail.com> |
||
|
Abhishek Sharma
|
b45e0c69e6 |
feat(security): warn at boot when the inference server is exposed anonymously (#13820)
* feat(security): warn at boot when the inference server is exposed anonymously `GET /v1/models` follows the dashboard login posture (`isAuthRequired()` / `requireAuthForModels`) while the inference routes follow `REQUIRE_API_KEY`. On an instance with an admin password set and `REQUIRE_API_KEY=false`, `/v1/models` answers 401 while `/v1/responses` is open to anyone who can reach the port — so the most natural probe an operator runs reports the opposite of the truth. #12568 added a boot warning for exactly this combination, but wired it only into the API bridge and the live dashboard WebSocket. The Next server that actually answers `/v1/chat/completions` and `/v1/responses` never reached it, and it is the one that binds every interface by default (`process.env.HOST || "0.0.0.0"`). Wire the existing guard into the Next boot hook, and document the split. Resolving the bound host needed care: two entrypoints bind that server and they read different variables. `run-next.mjs` honours `HOST`; the Docker entrypoint delegates to Next's generated `server.js`, which reads `HOSTNAME`. `run-next.mjs` now publishes what it actually binds as `OMNIROUTE_BOUND_HOST`, and the guard reads that, then `HOSTNAME`, then the shared `0.0.0.0` default. `HOST` is deliberately absent from the chain: the standalone server ignores it, so consulting it there would warn about an interface the server is not on — and one false warning teaches an operator to ignore the next one. Closes #13695 * docs(changelog): add changelog.d entry for #13820 |
||
|
Sean Ford
|
010250cf08 |
fix(memory): list provider-node models in the Embedding and Rerank selectors (#13740)
* fix(api): type compatible-provider-node models in /v1/models by the node's apiType Model rows discovered from an OpenAI-compatible provider node rarely carry endpoint metadata — a TEI / Infinity / vLLM `/v1/models` listing is just ids — and the catalog defaulted such rows to `["chat"]`. An `embeddings`-typed node exposing `bge-m3` and a `rerank`-typed node exposing `bge-reranker-v2-m3` therefore both surfaced in GET /v1/models as untyped chat models: clients that build their picker from `type: "embedding"` / `type: "rerank"` never saw them, and chat pickers listed models that 400 on chat. - src/shared/constants/modelSupportedEndpoints.ts: add defaultEndpointsForProviderNodeApiType(apiType) — embeddings → ["embeddings"], rerank → ["rerank"], audio-* → themselves, images-generations → ["images"], chat/responses/unknown → ["chat"] (unchanged default). - src/app/api/v1/models/catalog.ts: build a node-id → apiType map next to the existing node-id → type map; the synced-model and custom-model loops fall back to the node's modality instead of ["chat"] when a row has no supportedEndpoints; the custom-overlay merge path also classifies `type`/`subtype` from the overlay's explicit supportedEndpoints, so a manual `["rerank"]` row layered on a discovered chat-default row is re-typed. Explicit supportedEndpoints on any row still take precedence, and chat / responses nodes keep the historical behavior. tests/unit/catalog-provider-node-apitype-endpoints.test.ts covers the helper and the catalog end-to-end for embeddings, rerank, mixed, chat, and overlay cases via getUnifiedModelsResponse(). * chore(changelog): name the #13734 fragment * refactor(api): keep the provider-node modality helpers out of catalog.ts catalog.ts is frozen by the file-size gate (must not grow past 2075 lines) and the apiType fallback pushed it to 2093. Move the node apiType index, the endpoint fallback and the overlay type/subtype fields into catalogNodeModality.ts so catalog.ts ends one line shorter than the base; behaviour and tests are unchanged. * fix(api): give nodeModelEndpoints a string[] return so the catalog classifier typechecks The API-route typecheck gate flagged TS2345 at both classifyModelSupportedEndpoints() call sites: the helper returned `ModelSupportedEndpoint[] | unknown[]`, and unknown[] is not a readonly string[]. The base code only passed because the synced row's supportedEndpoints was untyped. Same pass-through cast overlayEndpoints() already uses; no behaviour change. * fix(memory): list provider-node models in the embedding and rerank selectors GET /api/memory/embedding-providers and GET /api/memory/rerank-providers appended local provider nodes by apiType alone and always with models: []. A node typed "embeddings" that also serves a rerank model — one TEI / Infinity / vLLM box hosting both bge-m3 and bge-reranker-v2-m3 is the common self-hosted layout — was filtered out of the Rerank selector entirely (apiType not in chat/responses/rerank) and showed up in the Embedding selector as a provider with nothing to pick. Typing prefix/model by hand worked because the request path resolves it directly; only the convenience layer was blind. Add src/lib/memory/embedding/nodeModalityListings.ts, which builds the listing from the node's synced + custom model rows, typing each row the way /v1/models does (explicit supportedEndpoints wins, otherwise the node's apiType via defaultEndpointsForProviderNodeApiType; a custom overlay re-types a discovered row). A node is listed for a modality when its apiType matches, when it is a generic chat/responses node (historical behaviour, kept so catalog-less nodes still appear), or when any of its rows is typed for the modality. Both endpoints use it; the curated registries stay first and win on prefix collisions. * chore(changelog): name the #13740 fragment |
||
|
小妍儿 ✨
|
6d585625f0 |
fix(providers): honor Alibaba workspace embedding and rerank endpoints (#13293)
* fix(providers): honor Alibaba workspace embedding and rerank endpoints * docs(changelog): add Alibaba workspace endpoint fix * refactor(rerank): keep Alibaba response adapter focused --------- Co-authored-by: 千乘妍 (Xiaoyaner) <xiaoyaner0201@users.noreply.github.com> |
||
|
Wu Shuwen
|
36493a6270 |
fix(evals): fail a case whose model call errored instead of scoring it passed (#13201)
* fix(evals): fail a case whose model call errored instead of scoring it passed
runSuite() attached caseMetrics[id].error to the graded result but never forced
`passed` to false. executeEvalCase() returns a failed call as an ordinary output
string ("[ERROR] <message>"), so any expected pattern that happened to match that
text was recorded as a pass. That inflates the reported pass rate, and reports a
non-zero score for a run in which no model was ever reached.
Built-in codex-comparison case codex-07 reproduces it: its pattern is
"try|catch|throw|error|Error" and the provider-resolution failure text ends with
"...added as a combo entry.", so the `try` alternative matches and the case is
scored as passed while carrying a non-empty error.
A case that never reached a model has no measured behaviour to grade, so a
failure is the only honest score.
Refs #13137
* docs(changelog): add fragment for the errored-eval-case fix (#13201)
|
||
|
Tony Yu
|
3d3f71f514 |
fix(oauth): read pollToken body once on non-JSON upstream responses (kimi-coding, github) (#13046)
* fix(oauth): read pollToken body once on non-JSON upstream responses
The device-flow pollToken handlers for kimi-coding and github tried
response.json() first and fell back to response.text() in the catch.
Once .json() rejects on a non-JSON body the stream is already consumed,
so the .text() fallback always throws TypeError (Body is unusable) and
pollToken rejects, surfacing as a generic 500 on /api/oauth/<provider>/poll
instead of the intended graceful { error: "invalid_response" } payload.
Non-JSON responses are realistic when the OAuth upstream sits behind a
CDN/anti-bot HTML error page or a proxy interstitial (auth.kimi.com in
particular).
Read the body once as text, then JSON.parse it, preserving the original
invalid_response fallback. Adds a regression test that drives both
providers with a stubbed fetch returning an HTML error page and a JSON
error body. Prunes the two now-unused no-unused-vars suppressions for the
removed catch bindings.
* docs(changelog): fragment for #13046
|
||
|
Bob.Hou
|
9956f13b35 |
fix(db): never TRUNCATE-checkpoint a live WAL (SIGBUS under traffic) (#14005)
* fix(db): never TRUNCATE-checkpoint a live WAL A live TRUNCATE checkpoint rewrites the shared wal-index while other processes hold it mapped; dereferencing the stale mapping SIGBUSes the process. Two production crashes six hours apart, coredump stack in better-sqlite3 native memcpy (issue #13973). Remove the periodic TRUNCATE scheduler. Runtime checkpoints are PASSIVE-only, which move pages without changing the wal-index geometry, while TRUNCATE stays on the shutdown path where reclaiming the file is safe. The 256MB size guard now warns instead of escalating to a live TRUNCATE, busy PASSIVE ticks feed the persisted busy telemetry that the TRUNCATE tick used to carry, and a positive OMNIROUTE_WAL_TRUNCATE_INTERVAL_MS logs a one-time deprecation warning. Signed-off-by: Minxi Hou <houminxi@gmail.com> * fix(db): RESTART the WAL when it exceeds the size guard The 256MB size guard only warned, so a live WAL could keep growing until the next restart. wal_checkpoint(RESTART) starts a new WAL file without rewriting the mapped wal-index, which is what SIGBUS'd the process when we used TRUNCATE under traffic. Related to #13973. Signed-off-by: Minxi Hou <houminxi@gmail.com> * docs: drop a fake TRUNCATE env name from the WAL guard row Backticks around TRUNCATE made the env/docs checker treat it as a variable. The VACUUM rows next to it were never part of this change and are not in the base docs. Signed-off-by: Minxi Hou <houminxi@gmail.com> --------- Signed-off-by: Minxi Hou <houminxi@gmail.com> Co-authored-by: Diego Rodrigues de Sa e Souza <diegosouza.pw@gmail.com> |
||
|
Diego Rodrigues de Sa e Souza
|
24fc202d9f |
fix(ci): repair the API Route Typecheck base-red blocking every PR (#14079)
The gate fails on the pure release/v3.8.51 tip (3 files above the frozen baseline), so every PR against the release is born red on it. None of the three is a PR defect — they are drift from merged work: - src/lib/usage/glmResetCards.ts: entered the gate's scope when #12754 added the route that imports it. runWithProxyContext is an untyped async helper (Promise<any>), so runWithConnectionFetch<T> could not return T. Every call site passes an async callback and awaits it — declare that contract: fn: () => Promise<T> → Promise<T>. - src/sse/handlers/chat.ts: handleSingleModelChat had no return annotation, so runWithTransientBackendRetry<T extends ResponseLike> fell back to the constraint and the value could no longer feed withSessionHeader(Response). Annotated Promise<Response> (every return path builds a Response). - src/app/api/internal/codex-responses-ws/route.ts: the bridge helpers return either { error: Response } or a payload, but as unannotated object-literal unions TypeScript synthesised error?: undefined on the success member, and every "error" in x guard stopped narrowing (TS2339 x5 on the destructure). Explicit return types keep the discriminant real; the ApiKeyMetadata alias now points at the policy shape (the wider one both sources are assignable to), which clears the TS2740 self-mismatch; logger.warn → log.warn (the module logger factory has no .warn). check-api-typecheck.mjs: OK — 283 errors, baseline ratcheted DOWN from 294 (codex-responses-ws 7→4 TS2339, TS2740 1→0; combos/test and keys/[id] 1→0). typecheck:core clean; 43/43 unit tests in the touched areas green. |
||
|
Fouad Salkini
|
176d632a2d |
feat(api): add per-key allowAutoCombos to gate the built-in auto/* combos (#13670)
* feat(api): add per-key allowAutoCombos to gate the built-in auto/* combos
`auto/*` combos currently bypass per-key authorization entirely. They are
virtual — synthesised in the catalog, never stored as combo rows — so
`resolveRequestedComboName()` returns null for them and
`isComboAllowedForKey()` fails open:
const comboName = await resolveRequestedComboName(modelStr);
if (!comboName) return { allowed: true, comboName: null };
`validateModelAccess()` then sets `requestedComboName = modelStr` for any
`auto/` id and returns before `isModelAllowedForKey()` runs, so
`allowedModels` and `blockedModels` are skipped for those ids too.
The effect is that `allowedCombos` does not constrain `auto/*`: a key
scoped to a single cheap lane can still send `auto/best-coding` and reach
every model on the gateway. `blockedModels: ["auto/*"]` only unadvertises
the ids — it cannot deny them.
Add an explicit per-key flag instead of tightening the fail-open, which
would silently revoke `auto/*` from every key whose `allowedCombos` lacks
an entry for it. `allow_auto_combos` is NOT NULL DEFAULT 1 and the row
parser treats anything but an explicit falsy value as allowed, so every
existing key keeps working and opting out is deliberate.
When set to false:
- `validateModelAccess()` rejects `auto/*` for that key;
- the catalog skips the `auto/*` synthesis loop for it, reusing the
existing `hideAuto` break so the key is not offered ids it cannot use.
Settable via PATCH /api/keys/[id]. The create path and the dashboard
toggle are deliberately left for a follow-up: the API Manager control
needs UI strings across all message catalogs, which does not belong in
the same change as the policy fix.
* feat(dashboard): add the Auto Combos toggle to API key permissions
Exposes the `allowAutoCombos` flag in the API Manager permissions modal so
the per-key gate can be managed from the dashboard rather than only over
the API.
The control mirrors the prompt-compression toggle: a small dedicated
component, a `role="switch"` button, and labels from the `settings`
message namespace.
Defaults to ON. State reads `apiKey?.allowAutoCombos !== false` — using
`!== false` rather than `=== true` so a key that predates the column, or
one that has never been configured, renders as enabled and matches the
`NOT NULL DEFAULT 1` column.
The field is threaded through all three positional lists (the save
handler signature, the modal prop type and the onSave call) plus the
PATCH payload, so no later argument shifts position.
UI strings are added to en.json and to vi.json. Vietnamese is translated
rather than left as a sync placeholder because
tests/unit/i18n-vi-completeness.test.ts asserts key parity with English
and bans `__MISSING__` markers in that locale. The remaining locales fall
back to English at runtime; `i18n:check-ui-coverage` still passes well
clear of its threshold. They are deliberately not mass-synced here: a
full `i18n:sync-ui` run also replicates ~844 unrelated pre-existing gaps
across all 50 catalogs, which does not belong in this change.
* feat(api): advertise the combo description in /v1/models
A combo's description is stored on its record and returned by
GET /api/combos, but the catalog row never carried it, so no client could
show it.
Claude Code's gateway model discovery reads exactly `id`, `display_name`
and `description` from each entry in the /v1/models `data` array and
renders the description in the /model picker — an entry without one reads
"From gateway" instead. Other OpenAI-compatible clients surface it too.
Emit it only when the combo actually has one, so rows for combos without
a description are byte-identical to before. The value is typeof-narrowed
and trimmed because ComboRecord is Record<string, unknown>, and
`comboMetadata` still spreads last so context and capability metadata
keep precedence.
`display_name` is deliberately not sent: a combo's id is already its
human-chosen name, and the field is only consulted when it differs from
the id.
Ref: https://code.claude.com/docs/en/llm-gateway-protocol.md#model-discovery
* fix(api): list a key's allowed combos in /v1/models
`allowedCombos` gates combos; `modelAccessMode`, `allowedModels` and
`blockedModels` gate provider models. The catalog consulted only the
latter, so a key with `modelAccessMode: "restricted"` and an empty
`allowedModels` received an empty catalog — zero rows — while every combo
in its `allowedCombos` dispatched normally. The catalog contradicted the
key.
Observed on a live gateway: a key with 24 entries in `allowedCombos` and
`restricted` + `allowedModels: []` returned {"object":"list","data":[]},
yet `claude-orchestrate` answered 200 on that same key.
Gate combo rows on `allowedCombos` instead of hiding them. Listing a
combo the key can already dispatch grants no new access, so this is a
consistency fix rather than a relaxation, and it needs no opt-in: the
rule is simply that a key's catalog shows what that key can use.
auto/* rows are exempt. They fail open at dispatch — they resolve to no
stored combo — and their synthesis is already gated by allowAutoCombos,
so gating them here would make the catalog stricter than dispatch.
The decision lives in a new exported helper, isComboNameAllowedForKey(),
which wraps the existing matchesComboAccessRule. An absent list means no
combo restriction, matching validateComboAccess, which skips the check
when allowedCombos is not an array; an empty list allows nothing.
Also advertise `display_name` on combo rows from an operator-set
`displayName` field. Claude Code uses it as the picker entry's name when
it differs from the id, which lets a combo carry a discovery-compatible
id and still read cleanly. It is never derived from the combo name — an
unset field advertises nothing.
* fix(api): accept displayName on the combo schemas
The previous commit advertises `display_name` in /v1/models from a
combo's `displayName`, but neither createComboSchema nor
updateComboSchema declared the field, so Zod stripped it from every
request body and the value could never be set. The endpoint would have
answered 200 and written nothing — the feature was unreachable.
This is the same silent no-op that made `blockedModels` unsettable on
API keys: a field plumbed through the route and the store, missing only
its schema declaration.
Declare it on both schemas and count it in updateComboSchema's "no valid
fields" guard, so a body carrying only `displayName` is a valid update
rather than being rejected as empty. Nullable on update so a label can be
cleared.
* feat(api): add per-key catalogScope to scope what /v1/models advertises
A key had no way to say which kinds of thing its catalog should list. It
always advertised whatever the key's model and combo policies permitted,
mixed together. A client that builds its model picker from /v1/models —
Claude Code's gateway discovery, for one — then sees provider models
alongside the curated combos it was meant to offer.
Add a three-way per-key setting: "all" (default), "combos", "models".
This is a listing preference, not an access control: narrowing it never
changes what the key may dispatch, which the model policy and
allowedCombos continue to decide. That is why it is an explicit setting
rather than implied behaviour — unlike gating combo rows on
allowedCombos, which was a correctness fix and needed no opt-in.
Defaults to "all" everywhere: the column, the parser, the metadata and
the UI state, so every existing key is unchanged. The parser widens to
"all" on an unrecognised value rather than narrowing, so a bad value can
never silently hide rows an operator expects to see.
The dashboard control is a segmented radio group beside the Auto Combos
toggle. UI strings are added to en.json and vi.json; the remaining
locales fall back to English, and vi is translated rather than left as a
sync placeholder because tests/unit/i18n-vi-completeness.test.ts asserts
key parity and bans markers there.
* fix(api): invalidate the model catalog on key visibility changes
updateApiKeyPermissions already advances the unified /v1/models catalog
generation for the fields that change what a key may dispatch, but the two
fields this branch introduces -- allowAutoCombos and catalogScope -- were
missing from that predicate. Both change what the catalog advertises, so a
PATCH toggling either one left the request-shaped catalog cache serving the
previous listing until its TTL expired, and the dashboard's API-key screen
could show a catalog that disagreed with the key it had just written.
Add the two fields to the existing predicate -- no new cache machinery. The
call still runs only after a successful write, so a no-op or failed update
does not invalidate, and unrelated metadata edits (isActive, rate limits)
still leave the catalog cached.
Observed on a live deployment before the fix: PATCH catalogScope="combos"
returned 200 and the column read back "combos", yet GET /v1/models kept
returning the previous mixed rows until a process restart, after which the
same key correctly returned combo-only rows.
* docs(changelog): add fragment for per-key allowAutoCombos and catalogScope
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
* chore(quality): rebaseline the two ceilings this PR's own growth moved
src/app/api/v1/models/catalog.ts 2075 -> 2117 and src/lib/db/apiKeys.ts
1625 -> 1659. Measured on the clean tip first: catalog.ts sits at 2074 (under
its 2075 ceiling) and apiKeys.ts at 1620 (under 1625), so none of this is
inherited — it is the feature itself. Gating the built-in auto/* combos per key
means the permission field has to be read, validated and carried all the way to
the catalog filter, and each of those is an explicit call site rather than
something extractable without hiding the gate.
Covered by the PR's 25 tests. The other violations in this tree (chatHelpers.ts,
chatCore.ts, chatcore-translation-paths.test.ts) are inherited base-reds and were
left untouched.
---------
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
|
||
|
Aaron Scherer
|
7d69b02a29 |
fix(db): skip integrity scans during health polling (#13149)
* fix(db): skip integrity scans during health polling * test(db): update health error fixture for scan-free polling * docs(changelog): add fragment for health poll integrity skip * fix(db): keep the #13149 dashboard skip inside the #13717 managed health check Merge fallout only: runManagedDbHealthCheck moved behind the health coordinator on the release tip, so the per-call skipIntegrityCheck now travels through it. A waived integrity scan is part of the job identity, so it is never replayed from the 60s diagnosis cache to a caller that asked for the full scan. Co-authored-by: cryptiklemur <cryptiklemur@users.noreply.github.com> --------- Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> Co-authored-by: cryptiklemur <cryptiklemur@users.noreply.github.com> |
||
|
Bob.Hou
|
5f9e153971 |
fix(mcp): fall back when better-sqlite3 export is not callable (#13903)
* mcp/audit: fall back when better-sqlite3 export is not callable
Dashboard MCP status polls reopen a failed native sqlite load every 30s
because a minified TypeError ("a is not a function") was not treated as
a native load failure and a failed open was not cached. Classify that
shape, fall back to node:sqlite, cache the miss, and refuse to ship a
Docker image without better_sqlite3.node.
Signed-off-by: Minxi Hou <houminxi@gmail.com>
* mcp/audit: force native better-sqlite3 compile in Docker
better-sqlite3 13 ships a linux prebuild. Bare `node-gyp rebuild`
then only TOUCHes stamp files and never writes
build/Release/better_sqlite3.node, so the new test -f gate fails the
image build. Pass --force_build=1, matching the package's own
build-release script.
Signed-off-by: Minxi Hou <houminxi@gmail.com>
* db/core: keep native-load classification under the file-size cap
The audit fallback added two TypeError fingerprints in core.ts and
crossed the frozen 1788-line cap. Move the classifier into
sqliteLoadError.ts and re-export it so existing importers stay stable.
Signed-off-by: Minxi Hou <houminxi@gmail.com>
* build/bootstrap: keep the encrypted-credentials probe narrow
The native-load classifier was copied into scripts/build/bootstrap-env.mjs
alongside the runtime one, but the two files consume its verdict in opposite
directions. In src/lib/db/sqliteLoadError.ts a true verdict means "the driver
is unusable, cascade to node:sqlite", so treating a non-callable export as a
load failure is what we want. In the bootstrap the verdict feeds
hasEncryptedCredentials, where true means "no encrypted credentials found" and
clears the way to generate a fresh STORAGE_ENCRYPTION_KEY.
With the TypeError patterns in the bootstrap copy, a binding that loads but
exports something non-callable over a database full of enc:v1: rows reads as an
empty database, and the operator silently loses access to every stored
credential. Drop those two patterns from the bootstrap copy only, and note in
both files why the pair is deliberately not identical.
A corrupt binding still fails loudly there, now with the database path, the
underlying message, and a rebuild hint, so the narrower classifier does not
cost any diagnosability.
Signed-off-by: Minxi Hou <houminxi@gmail.com>
* fix(mcp): keep audit logging recoverable when the database is created later
getDb() cached a null for the "storage.sqlite does not exist yet" branch, and
closeAuditDb() returns before clearing a falsy cache — so an MCP server started
before the app created the database stayed without audit logging for the whole
process lifetime. Only a genuine driver-load failure is cached now; the
not-found branch retries, which is how it recovers when the file appears.
Covered by a new test that fails without the change.
Also replace the fabricated minified TypeError text ("a is not a function")
thrown by the loader with "better-sqlite3 export is not a function": the
operator sees a diagnosable message and isNativeSqliteLoadError() still
classifies it (it matches on "is not a function").
---------
Signed-off-by: Minxi Hou <houminxi@gmail.com>
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
|
||
|
Bob.Hou
|
ec780ef2dc |
feat(compression): make Lite tool-result truncation length configurable (#13915)
* feat(compression): make Lite tool-result truncation length configurable Lite truncated tool results at a hardcoded 2000 characters. Coding-agent payloads (file reads, crash dumps) lost the middle of the content with no supported way to raise the cap. Honor lite.maxToolLength from settings, then OMNIROUTE_LITE_MAX_TOOL_LENGTH, then 2000. Existing installs keep the old length. Related to #13178. Signed-off-by: Minxi Hou <houminxi@gmail.com> #13178 stays open. * compression/lite: keep a stored cap when a step or toggle write is incomplete An out-of-range step maxToolLength was still a number, so it hid a valid global cap and fell through to env. A toggle-only settings PUT replaced the whole lite row and dropped the stored cap. Save treated an out-of-range number like a cleared field. Reject the bad Save, merge omitted caps, and use null to clear. Related to #13178. Signed-off-by: Minxi Hou <houminxi@gmail.com> * compression/lite: stop dashboard copy from hard-coding a 2000-char cap The page overlays schema descriptions from i18n. Updating only LITE_SCHEMA left operators seeing "over 2,000 characters" after the cap became configurable. Also assert the Save error string, not the Save button. Related to #13178. Signed-off-by: Minxi Hou <houminxi@gmail.com> * chore(changelog): move #13915 entry to a changelog.d fragment --------- Signed-off-by: Minxi Hou <houminxi@gmail.com> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Innokentiy Solntsev
|
e7872e57c3 |
fix(db): reclaim freed pages incrementally instead of a blocking VACUUM in the cleanup scheduler (#12821) (#12830)
* fix(db): reclaim freed pages incrementally instead of a blocking VACUUM in the cleanup scheduler (#12821) startCleanupScheduler() ran a synchronous whole-database VACUUM on the event loop whenever a cleanup pass deleted at least one row - 30 s after every start and every 6 h. With node:sqlite that blocks every route (/healthz included) for the duration: 7 min 55 s on a 540 MB storage.sqlite to reclaim six rows. It also bypassed vacuumScheduler, the app-level owner of full VACUUMs and the operator's scheduledVacuum / vacuumHour settings. cleanup.ts no longer issues a full VACUUM. After each pass reclaimFreedPages() branches on PRAGMA auto_vacuum: - INCREMENTAL: drain the freelist with PRAGMA incremental_vacuum(N) in ~1 MiB batches (N from page_size), pausing between batches for as long as the last one took (<=250 ms), PASSIVE checkpoint every 64 batches and a TRUNCATE checkpoint at the end so the main file shrinks in WAL mode; hard caps of 2048 batches / 30 s per pass, the remainder waits for the next pass. - FULL: nothing to do, SQLite reclaims on commit. - NONE: incremental_vacuum is a no-op, so record a request via the new vacuumScheduler.requestFullVacuum(); the rebuild runs in the configured window (or via the Storage page button). scheduledVacuum=never is honored. vacuumScheduler persists fullVacuumRequestedAt / fullVacuumRequestReason, clears them on the next successful runNow(), and hydrates from key_value before an early request so it cannot clobber a persisted lastRunAt. Loop robustness: db.exec() rather than pragma() (bun:sqlite's all() steps a zero-column pragma once), SQLITE_BUSY/LOCKED and a handle closed under the pass stop it quietly, other errors stop it with partial progress logged. Also drops the duplicate cleanupProxyLogs() call in the scheduled pass - runAutoCleanup() already covers proxy_logs. Tests: new tests/unit/db/cleanup-reclaim-freed-pages.test.ts (INCREMENTAL drain/pause/checkpoint, page_size-derived batch, caps, FULL no-op, NONE defers and leaves page_count untouched, runScheduledCleanupPass() path); vacuum-scheduler.test.ts covers requestFullVacuum persistence, restart survival and clearing; cleanup-column-fix.test.mjs now asserts incremental_vacuum and the absence of a full VACUUM statement. * chore(changelog): name the #12821 fragment after its PR (#12830) * fix(db): extract reclaimFreedPages into its own module and fix full-suite regressions Split the #12821 incremental-vacuum reclamation logic out of cleanup.ts into src/lib/db/reclaimFreedPages.ts (re-exported for callers/tests) so cleanup.ts stays under the file-size cap after the #13011 reconciliation merge grew it past the 1200-line threshold. Also fixes two full-suite failures surfaced by running the cleanup/vacuumScheduler/db-health suite post-merge (not just this PR's own 3 test files, per the plan-file's mandatory item): - tests/unit/cleanup-column-fix.test.mjs scanned cleanup.ts's raw source for the PRAGMA incremental_vacuum invariant, which now lives in the extracted module — updated to scan both files. - tests/unit/db/cleanup-reclaim-freed-pages.test.ts asserted the freelist count is byte-for-byte unchanged when auto_vacuum=NONE. The tip's runAutoCleanup() now also runs cleanupCompressionRunTelemetry(), which lazily creates its table on first use (ensureCompressionRunTelemetryTable) — a legitimate one-time page cost from a freshly migrated DB, unrelated to reclaimFreedPages()'s own behavior. Loosened the assertion to a small tolerance while keeping the page_count assertion that actually guards against a full rebuild. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> * chore(db): drop the reclaimable-bytes VACUUM gate test superseded by incremental reclaim tests/unit/vacuum-reclaimable-threshold.test.ts pinned cleanup.ts's vacuumAfterCleanup()/getReclaimableBytes()/getVacuumMinReclaimableBytes() (#13079). This branch removes the inline post-cleanup full VACUUM entirely in favour of reclaimFreedPages() (#12821), which reads the same freelist_count / page_size signal and defers a full VACUUM to the vacuum scheduler when auto_vacuum=NONE. With those three exports gone the file cannot compile, and the behaviour it guarded no longer exists. --------- Co-authored-by: insoln <is@careerum.com> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Innokentiy Solntsev
|
241e63bfea |
feat(usage): redeem GLM Coding Plan Reset Cards from Provider Limits (#12754)
* feat(usage): redeem GLM Coding Plan Reset Cards from Provider Limits z.ai sells Reset Cards that clear an exhausted GLM coding-plan window (5-hour or weekly) ahead of its natural rollover, but OmniRoute only ever read the passive nextResetTime, so redeeming one meant leaving the dashboard. Add the wire layer for z.ai's two reset endpoints (/api/biz/customer-package-reset/list and /use), which authenticate with the same Bearer API key as /api/monitor/usage/quota/limit and report failures inside an HTTP-200 envelope, so callers must inspect success/code rather than the status line. The banked count rides along with the quota poll - only for keys that actually report a resettable window, and strictly best-effort so a card-less account or a transient failure still renders its quotas. The existing reset-credit card, picker and confirmation flow, until now gated to Codex, now also drive glm/glm-cn/glmt/zai through the new /api/usage/glm-reset-card route, reusing z.ai's requestId as the idempotency key so a retry cannot burn two cards. * test(usage): cover GLM reset-card edge cases * test(dashboard): require GLM reset-card copy * fix(usage): harden GLM reset-card redemption * fix(usage): treat missing GLM key as empty * fix(usage): fence GLM reset-card operations and coalesce lease-window duplicates - Acquire a synthetic 60s exclusive-connection lease around each list/use wire operation; release in finally so a competing lease can acquire immediately after success or failure. - Coalesce same-key duplicates that arrive after lease acquisition by checking the in-flight attempt before loading the connection. - Run the post-commit quota refresh outside the lease (redemption is already committed; the refresh is auxiliary and failure-tolerant). - Do not discard a retained ambiguous attempt on a lease-conflict 409. - Harden transport error mapping: static messages for proxy transport failures, keep explicit direct routing for unproxied connections through list, use, and refresh. * fix(i18n): sync GLM reset-card keys to pt-BR and vi locales --------- Co-authored-by: insoln <is@careerum.com> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Nguyen Thanh Dat
|
ac0a63117e |
fix(providers): read the vLLM context window from max_model_len (#12897)
* fix(providers): read the vLLM context window from max_model_len normalizeDiscoveredModels resolved the window from inputTokenLimit, context_length, contextLength and top_provider.context_length. vLLM reports it as max_model_len and nothing else, so a synced vLLM model carried no inputTokenLimit and the resolver fell back to the 128K default - half the window on a 250K deployment. The native vllm provider and the OpenAI/Anthropic-compatible custom providers all pass raw records through this function, so one chain entry covers the three connection shapes. Closes #12858 * docs(changelog): fragment for #12897 --------- Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Lance Woodson
|
8074e3d596 |
fix(resilience): honor declared effort vocabulary in reasoning rule gate (#12686)
* fix(resilience): honor declared effort vocabulary in reasoning rule gate The reasoning-routing rule capabilityFor() hardcoded a gpt-5.6-(sol|terra|luna) whitelist for forced max/ultra, rejecting every other thinking-capable model even when the model's resolved capabilities declare the requested tier (synced supportedThinkingEfforts or an operator Model Overrides reasoning_efforts override). This 400'd direct calls with "Reasoning effort 'max' is not supported by the configured target" for models like Merge Gateway zai/glm-5.3-flash, which natively accepts low|high|max. The gate now treats a declared vocabulary containing the requested tier as authoritative, mirroring the dispatch-time sanitizer (open-sse/executors/base/reasoningEffort.ts) which already forwards declared tiers verbatim. Undeclared models keep the legacy gpt-5.6 regex verdicts and the unknown passthrough. * fix(resilience): gate forced max against the static registry the sanitizer clamps with Adversarial review finding: the gate read supportedThinkingEfforts from getResolvedModelCapabilities, which prefers the DB override over the registry. For a registered model with a narrow registry vocabulary and a widening operator override, the gate passed forced max but the dispatch-time sanitizer (executors/base/reasoningEffort.ts) clamps against the STATIC registry and would silently downgrade max to the registry ceiling — converting a loud 400 into a silent wrong-effort request. Order of precedence in the gate now: 1. static registry vocabulary (authoritative — matches sanitizer clamping) 2. declared/overridden vocabulary for unregistered providers (#8057 path) 3. legacy gpt-5.6 regex, then unknown/unsupported verdicts Also pins the test fixture to a synthetic model id so a future models.dev sync row cannot flip the unknown-precondition assertion. * fix(resilience): gate registry lookup mirrors the dispatch sanitizer exactly Review findings on the forced max/ultra gate: - resolve the registry through getProviderModels (id->alias namespace) and match entry aliases, mirroring reasoningEffort.ts — a raw provider id or alias-spelled model no longer skips the registry branch and diverges from dispatch clamping - treat an empty declared vocabulary as no declaration (falls through), matching the sanitizer's declaredRanked.length>0 guard — before, a model declaring [] was gated to unsupported while dispatch forwarded verbatim - an operator-declared vocabulary that excludes the forced tier is terminal; the legacy gpt-5.6 regex can no longer resurrect a tier the override narrowed away - rewrite the registry-outranks-override test: create the matching rule so the decision is non-null, assert unconditionally, pin gpt-5.6 narrowing, alias namespace parity, and use the deterministic xai/grok-4.6 fixture * docs(changelog): clarify override scope for registry-declared models * test: drop placeholder issue reference from test names * chore(changelog): name fragment after PR #12686 |
||
|
Lance Woodson
|
821d02ba13 |
fix(providers): parse per-vendor-route reasoning.effort_values in discovery (#12730)
OpenAI-compatible model discovery does not recognize per-vendor-route reasoning vocabularies declared under vendors.<vendor>.capabilities.reasoning in GET /v1/models (Merge Gateway's documented catalog shape), so synced models carry no supportedThinkingEfforts/defaultThinkingEffort and operator effort data resets on every model sync; models whose upstream accepts a native max tier cannot be used with forced-max reasoning rules. Parse the shape into the existing supportedThinkingEfforts pipeline, intersected across vendor routes: the same canonical model declares different vocabularies per route and unpinned requests self-narrow to a route honoring the requested level, so a synced tier must be honored on every route the model can land on. Routes without effort_values declare no effort control and are excluded; disjoint vocabularies produce an authoritative empty list (no fall-through to generic tier shapes). detectDefaultThinkingEffort falls back to the intersection's highest tier ranked by the canonical effort order — only when the vendors shape is the record's winning vocabulary source, never escaping a flat or nested declared list. Detection is shape-gated, not provider-gated; Zod-validated (Hard Rule #7) with malformed vendor and tier entries dropped individually (discarding a whole route would widen the intersection, fail-open). Precedence: flat field > reasoning.supported_efforts / metadata (#7694) > vendor-route intersection > capabilities.effort_tiers (#9160) / supported_reasoning_levels / thinking.levels (#8347). |
||
|
Aaron Scherer
|
46730700f1 |
feat(usage): show separate Fable weekly limits (#13266)
* feat(usage): show separate Fable weekly limits * docs(changelog): add fragment for fable weekly usage |
||
|
initguru
|
b7192b72e2 |
fix(thinking): parse/scrub DSML tool-call markers and recognize adaptive thinking (#12905)
* fix(thinking): recognize adaptive thinking + parse/scrub DSML tool-call markers
Two defects combined to break DeepSeek-V4-Flash turns and raise 502
empty_response on Claude Code autocompact.
Defect 1 — DSML tool-call markers leaked as visible content:
DeepSeek-V4-Flash occasionally emits tool calls in a non-standard DSML
text format using full-width pipes instead of the OpenAI tool_calls JSON.
Two shapes appear in production call logs:
- complete block: <｜DSML｜:Read><path>...</path></｜DSML｜:Read>
- stray closers (truncated call): </｜DSML｜parameter></｜DSML｜invoke>
</｜DSML｜tool_calls>, sometimes trailing a system-prompt echo
The openai-compatible path never parsed these, so the markers leaked to
the client as visible content and the turn ended incomplete.
Fix: add open-sse/utils/dsmlToolCalls.ts — parseDsmlToolCalls() converts
complete DSML blocks into OpenAI tool_calls and strips stray closing
markers from content (streaming-safe via a holdback for partial openers).
Wire it into the response translator before extractXmlInvokeBlocks so
DSML and XML invoke tool calls share the same pending queue.
Defect 2 — adaptive thinking silently suppressed:
A prior inline === 'enabled' check on body.thinking.type silently
suppressed adaptive (the intent Claude Code actually sends), so
reasoning was dropped. The model then emitted DSML tool-call markers
as plain text, producing an incomplete stop finish. Fix: use
hasActiveClaudeThinking() (which recognizes enabled AND adaptive) to
set requestedThinking, thread it through stream.ts and translator
state, and gate thinking block emission on state.requestedThinking
so upstream reasoning_content only relays when the client opted in.
Tests: 29/29 (6 dsml-tool-calls, 5 thinking-active-claude-adapter,
3 translator-resp-dsml-integration, 15 translator-resp-openai-to-claude
incl. requestedThinking suppression regression). typecheck:core clean.
* fix(sse): strip echoed system-prompt preamble + preserve large analysis/summary blocks
DeepSeek-V4 and similar models echo the OMNIROUTE_SYSTEM_INSTRUCTION_APPEND
directive (appended to the system tail by claude-to-openai.ts) and whole chunks
of the system prompt (<analysis>/<system-reminder>/<summary> blocks, prose
reproductions of the superpowers skill section) verbatim at the START of their
reply — the 'system message leak' persisting after the request-side fix.
Add two streaming-safe preamble strippers in directivePreambleStripper.ts:
- createDirectivePreambleStripper(directive): drops a leading reproduction of
the exact configured directive across arbitrary SSE chunk boundaries.
- createSystemPreambleStripper(): removes <analysis>/<system-reminder>/
<summary> echo blocks and known prose heads (Phase B) from the very start
of a stream, only while the stream is still a preamble.
Wire both into openai-to-claude.ts content-delta path: chain the exact-directive
stripper then the system-echo stripper before DSML/XML-invoke parsing, so a
leading system echo is dropped before it reaches the client.
Preserve large blocks (>= SYSTEM_ECHO_THRESHOLD=1000 chars) and blocks with no
trailing content — these are the model's real response (e.g. a Claude Code
autocompact summary), not a short system-echo. Stops the autocompact
empty-response regression where a whole-summary <analysis> block was stripped
to empty (3a8515).
Regression: origin's markdown-boundary feature (bufferedPrefix /
splitMarkdownBoundary, commit
|
||
|
Koosha Paridehpour
|
9c5d60027e |
fix(api): stream /api/logs/export with row cap to prevent V8 heap OOM (#13123) (#13428)
* fix(api): stream /api/logs/export with row cap to prevent V8 heap OOM (#13123) Fixes #13123 GET /api/logs/export buffered every matching row into a single JSON.stringify call with pretty-printing (null,2), roughly doubling the string size. On tables with tens of thousands of rows this crashed the Node process with a V8 heap OOM, taking the gateway down for minutes. Changes: - Stream the response via ReadableStream, serializing one row at a time so peak memory stays bounded regardless of table size. - Add a configurable row cap (limit query param, default 10000, max 50000) so callers cannot accidentally request unbounded exports. - Remove pretty-printing (callers can pretty-print client-side). - Include cap metadata (capped, limit, totalAvailable) when the cap fires so callers know they received a truncated result. - Preserve backward-compatible response envelope: { count, hours, type, logs, ... }. * fix(api): push the /api/logs/export row cap down into the DB layer (#13123) The route-layer streaming + cap from the previous pass still called exportCallLogsSince()/exportProxyLogsSince(), which hydrated and buffered EVERY matching row (including rows beyond the limit) before the cap was ever applied — peak V8 heap was essentially unchanged. Adds countCallLogsSince()/countProxyLogsSince() (cheap COUNT(*), no row hydration, used for totalAvailable) and iterateCallLogsSince()/ iterateProxyLogsSince() that bound the query with SQL LIMIT and yield/hydrate one row at a time: a generator over a LIMIT-bounded id list for call_logs, and fixed-size LIMIT/OFFSET pages for proxy_logs (the shared SqliteAdapter only exposes run/get/all, not a `.iterate()` cursor, so LIMIT/OFFSET pagination is the available cursor-equivalent without widening that interface across all 4 driver adapters). The route now streams from these instead, so the full matching row set is never buffered. Also moves capped/limit/totalAvailable into the response header instead of only the trailer, so a client consuming the stream incrementally learns about truncation before processing every row. Rewrote the test to call the real route.GET handler against a seeded test database instead of a local reimplementation of the stream builder, so a regression in the route or its DB-layer delegates is actually caught. Documents the pre-existing (now more clearly load-bearing) breaking change in a changelog fragment: `limit` defaults to 10,000 rows, so exports that previously returned everything are silently truncated. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: Koosha Pari <koosha@phenotype.ai> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Diego Rodrigues de Sa e Souza
|
21772f40f3 |
fix(security): generate a random per-install CLI token salt (#13679) (#13909)
Both src/lib/machineToken.ts::getActiveSalt() and its mirror in
bin/cli/utils/cliToken.mjs derived the CLI/management bearer token as
HMAC-SHA256(raw machine-id, salt) with a checked-in literal default salt
("omniroute-cli-auth-v1"). Since /etc/machine-id is commonly world-readable,
any local user who never set OMNIROUTE_CLI_SALT could derive the same
bearer token as the server.
getActiveSalt() now generates a random 64-char-hex salt on first use and
persists it under <DATA_DIR>/cli-token-salt.json (falling back to the
literal only when neither the env override nor a persisted/writable salt
can be established). Both implementations use the same resolution order
and the same wx-flag create-race handling so the CLI and server keep
deriving the same token. OMNIROUTE_CLI_SALT stays the explicit operator
override, unchanged.
Regression test: tests/unit/machine-token-random-salt-13679.test.ts
|
||
|
initguru
|
4c4d5c7fbe |
fix(resilience): allow maxWaitMs=0 as disable sentinel for execution expiration (#12902)
* fix(resilience): allow maxWaitMs=0 as disable sentinel for execution expiration
maxWaitMs normalization clamped the value to min:1, silently rewriting
an operator's 0 ("disable the limiter-managed execution deadline") into
1 — a 1ms expiration that killed every long-running job instantly. This
broke long-running reasoning models (GLM-5.2 with reasoning.effort=max
spends minutes before the first token, exceeding any practical
maxWaitMs; the TTB safety net is FETCH_TIMEOUT_MS, default 600s).
Fix: lower the floor to min:0 so 0 is preserved as the disable sentinel.
Issue #4165 follow-up.
Tests: 7/7 (resilience-normalize-maxwaitms-disable 5 + rate-limit-
maxwaitms-disable-execution 2). typecheck:core clean.
* fix(resilience): relax requestQueueSettingsSchema.maxWaitMs to allow 0
normalizeRequestQueueSettings already treats maxWaitMs=0 as an explicit
disable sentinel (queue-wait budget off), but the settings API schema
still rejected 0 with min(1), so an operator could never actually reach
the fix through PATCH /api/resilience. executionMaxWaitMs is untouched
(stays min(1) — separate field, separate decision, see #12902 item 4).
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
* test(resilience): prove maxWaitMs=0 vs #12715's queue-wait gate behavior
Answers the open technical question from #12902's review: does a
GLOBAL maxWaitMs=0 reintroduce the unbounded-queue regression #12715
fixed (a request hanging ~6min until the client aborts)?
Evidence, exercising the real gate chatCore.ts actually calls
(accountSemaphore.acquireMany({ timeoutMs: requestQueue.maxWaitMs }),
not the Bottleneck reservoir the PR's own tests cover) under real
contention (maxConcurrency=1, two concurrent acquires):
- No: it does not hang. setTimeout(reject, 0) fires on the next
tick, so a second contending request is rejected with
SEMAPHORE_TIMEOUT in low milliseconds, never minutes.
- But it is also not a genuine 'no cap' — an operator setting 0
expecting 'wait as long as it takes' instead gets near-zero
tolerance for even momentary contention on any configured
concurrency gate (global/provider/account). This is a real
asymmetry vs. the Bottleneck reservoir path (where 0 truly means
unbounded) left for the maintainer to decide how to resolve —
not something this pass can decide unilaterally.
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
---------
Co-authored-by: Jihyun Son <jihyun.son@sk.com>
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
|
||
|
Koosha Paridehpour
|
28557418db |
fix(proxy): add combo scope to fail-closed proxy guard (#13551)
* fix(proxy): add combo scope to fail-closed proxy guard (fixes #13469) The hasBlockingProxyAssignment guard only checked account, provider, and global scopes. Combo-scoped proxy assignments were not checked, so a fully dead combo pool fell through to direct egress — leaking the host IP. - Add combo scope to the SQL guard query - Add optional comboName parameter to hasBlockingProxyAssignment - A dead combo pool now blocks egress like the other three scopes * fix(proxy): thread comboName through safeResolveProxy to the combo-scope guard (#13469) hasBlockingProxyAssignment() gained a comboName parameter and a combo-scope SQL clause, but its only caller, safeResolveProxy() in chatHelpers.ts, never passed it — the clause always bound NULL and never matched a real combo scope_id, so a fully dead combo-scoped proxy pool still fell through to direct egress. Thread comboName from handleSingleModelChat (where it is already in scope) through safeResolveProxy into the guard, and add tests covering both the guard predicate and the end-to-end wiring. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: Koosha Pari <koosha@phenotype.ai> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Koosha Paridehpour
|
5455740faa |
fix(compression): log warnings for unreadable settings rows (#13522)
* fix(compression): log warnings for unreadable settings rows getCompressionSettings() silently skipped non-string (BLOB) and invalid-JSON settings rows, making it impossible to diagnose config drift between the panel and the runtime. Now logs a warn-level message for each unreadable row, including the key name and a remediation hint (re-save from the Storage panel). Also warns when the 'engines' row exists but yields no valid toggles, so operators know their panel-configured engines map is being silently replaced by the legacy fallback. Fixes #13456 * test(compression): cover getCompressionSettings warnings for unreadable rows The test for #13456 only asserted a stubbed console.warn recorded a message and never called getCompressionSettings(), so it never exercised the production change. Seed a BLOB row, an invalid-JSON row, and an 'engines' row that isn't a usable object, and assert the resulting warnings; also assert a legitimately empty (but valid) 'engines' map does not warn. Also stop warning on a valid-but-empty 'engines' row: parseStoredEnginesMap returns null both for an unreadable row and for a well-formed {} (an operator who deliberately disabled every engine), so only warn when the stored value isn't a usable object at all. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: Koosha Pari <koosha@phenotype.ai> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> |
||
|
Diego Rodrigues de Sa e Souza
|
0d089e7e39 |
fix(quality): clear the release/v3.8.51 base-reds (#13947)
* fix(quality): clear the release/v3.8.51 base-reds
19 failing unit tests plus the API Route Typecheck and mutation-test-coverage
gates, all reproduced on the clean tip before touching anything.
Ten of the failures share one cause. #13452/#13798 made `*-compatible-*`
buildUrl() refuse a connection with no baseUrl instead of quietly defaulting to
the real OpenAI/Anthropic API — which would ship the operator's stored key to a
public third party. The guard is right; three fixtures still built those
connections unhydrated, and one of them put baseUrl at the top level of
credentials, where the chat path never reads it.
The rest:
- modelDiscovery.ts missed the VertexModelMetadataProvenance cast that its
read-path twin in db/models/synced.ts already had — both written by #12471.
- A provider-test regexp carried raw 0x00/0x1f bytes, which makes git, GitHub
and ripgrep treat the file as binary. Same character class, written
with escapes instead of the bytes themselves.
- #13399 (Agnes AI China) adds "agnes-cn" + "agnescn": the only two provider
prefixes since the count was last set (412 -> 414). Everything else added in
that range is model ids.
- The free-tier budget card SVG was stale (443 -> 452 models); regenerated by
its own script.
- Three new tests were missing from stryker.conf.json tap.testFiles, so the
mutants they kill did not count.
Three guards asserted syntax rather than the invariant they protect, and broke
when the source legitimately changed. Each was re-expressed and then verified by
mutating the source back:
- #2331 required modelEffort to head the rawEffort chain; #13556 deliberately
put the server-selected force rule first. The real invariant is relative —
modelEffort outranks the defaults a client injects — and it still trips when
explicitReasoning is moved ahead of it.
- The OAuth loopback guard matched the isLocalhost arm literally; #9944 added
`&& !opts?.manualLoopback`. It now matches the arm whatever guards it, and
still fails when the hint stops being built.
- The i18n scanner flagged dynamically-built keys — t("effort." + mode) reaches
it as a literal prefix, never a string. It now accepts a prefix that resolves
to a namespace holding messages, and still fails when the namespace is gone.
tests/unit/sse-auth.test.ts (#12080) expected a bare null where #13879 now
returns the key-policy diagnostic — the same sentinel shape the terminal-state
path has used since #12441. The assertion was rewritten to the constraint #12080
actually protects: nothing usable comes back and neither connection leaks. The
contract risk that remains — those sentinels are truthy, and executeWebSearch
treats any truthy value as a credential — is filed as #13945 rather than
widened into this PR.
Refs #13866
* fix(quality): clear the second wave of release/v3.8.51 base-reds
The tip moved 13 commits while the first pass was running and brought its own
reds. All reproduced locally on the merged tree first.
vitest 4.1.11 -> 5.0.0 in the #13661 development-group bump is a major, and
vitest 5 moved `vite` from a dependency to a peerDependency. This repo only ever
declared `vite` under `overrides`, which pins a version but installs nothing, so
`npm ci` stopped providing it and the Vitest job died at startup with
ERR_MODULE_NOT_FOUND. Declared as the devDependency it actually is — the same
^8.0.16 the override already pinned, and what @vitejs/plugin-react asks for as a
peer — and regenerated the lockfile: 684 lines added, none changed.
#12909 filtered a mapped array with `toolCall is JsonRecord`, but the element
type is the tool-call literal or null, and a predicate's type has to be
assignable to the parameter's (TS2677). Narrowed by the element's own type
instead; the literal still satisfies JsonRecord at the return.
#12906 added `|| result.errorCode === "empty_response"` to the stream-failure
condition and Prettier rewrapped it, so the #8928 probe — which located the
branch by an exact four-line string — stopped finding it. It now matches on what
the branch tests rather than how it is typeset, and still fails when the
eviction call is removed.
probe-7293 is the visible half of a real conflict, filed as #13948. #7293 merges
a mid-array system into index 0; #12908, landed later, demotes it to "user" in
place instead. Both target the same constraint and only one can win, and the
combination also reorders: the pre-translation hoist moves the turn forward
expecting it to stay a system message, then the demotion converts it where it
now sits, ahead of the conversation. Choosing between the two strategies is a
product call, not a base-red one, so the test was realigned to assert the half
that protects the caller — the instruction survives, as a user turn — and pins
the current ordering with a pointer to the issue, so the eventual decision shows
up as a deliberate test change instead of a silent regression.
Refs #13866, #13948
* fix(quality): allowlist vite, rebaseline tip growth, drop a dead import
Third pass on the release/v3.8.51 base-reds. Declaring `vite` in the previous
commit was correct but incomplete: check-deps is a human review point against
typosquatting, so a newly declared package has to be vouched for by name.
Recorded in dependency-allowlist.json with why it is needed — the official Vite
build tool, already pinned through overrides, and a required peer of both
vitest 5 and @vitejs/plugin-react. That also turns check-deps.test.ts green.
check-file-size went red on nine files. One is mine: sse-auth.test.ts grew when
the #12080 assertion was rewritten. Three of the four assertions I had added
were redundant with the strict deepEqual that follows them, so they are gone and
the file grows by 4 lines instead of 8; the cap absorbs the rest.
The other eight are production and test files this PR does not touch, grown by
other work and never rebaselined — which is the whole reason a base-red drain
exists. Each is attributed to the commit that grew it: #12906 (chat.ts,
chatHelpers.ts, proxyFetch.ts, stream.ts), #12904 + #12910 (chatCore.ts), and
batch_api.test.ts from the same wave. Two of them predate the wave entirely and
were already over cap on
|
||
|
initguru
|
9bc7eb8fd2 |
fix(vision-bridge): nested tool_result images + provider-prefix credential check (#12903)
* fix(vision-bridge): extract/replace images nested inside tool_result content
Claude Code sends tool_result images as {type:"image",source:{base64}}
nested inside a tool_result's content array, not as top-level content
parts. The vision-bridge guardrail's extractImageParts filtered nested
hits out (!p.nested), so these images were silently dropped — a
text-only executor then received a request with no image and returned
HTTP 400.
Port the path-based nested extraction/replace fix:
- MediaPart gains a path field: the key/index chain from
message.content[partIndex] down to the media object itself.
- inspect() tracks the path through recursion; pushPart stamps it.
- extractImageParts drops the !p.nested gate and emits path for nested
hits (extract↔replace contract preserved: same order, every hit
replaceable).
- replaceImageParts rewrites via detectMediaParts: top-level hits swap
their content slot, nested hits walk MediaPart.path via the new
replaceObjectAtPath helper.
- ensureBase64ImagesForClaudeWire skips nested hits (.filter(!p.path))
to keep its sequential index map aligned.
TDD: 7 failing tests (path field, nested extract, nested replace,
document order) → 47/47 pass. typecheck:core clean.
* fix(vision-bridge): resolve provider prefix to node id for credential check
Re-land 932002580 (2026-08-19), which was never merged: it branched off
|
||
|
Diego Rodrigues de Sa e Souza
|
3d5baf13f4 |
fix(providers): strip Vertex doc script blocks whose end tag carries junk (#13936)
The Vertex model-docs HTML is converted to plain text before the table parser reads context-window and token-limit numbers out of the cells. The script/style removal pass required the end tag to be `</script\s*>`, but the HTML spec closes the element on `</script\t\n foo>` too. Such a block survived the pass; the generic `<[^>]+>` strip below then removed both tags and kept the script BODY, so text that only ever existed inside a script became cell text the number parser trusts. Accept any end tag that starts with `</script`/`</style` followed by a tag-name boundary, matching what a browser does. CodeQL js/bad-tag-filter, alert #1007. |