feat(providers): add unit tests, ESLint suppressions, and fix hardcoded userid for TinyCMS Web

- Add unit tests for WASM init, UUID validation, challenge flow (15 tests)
- Add WASM source comment explaining binary origin
- Replace hardcoded userid with dynamic provider-specific data
- Add ESLint suppressions for no-explicit-any in WASM bridge code
- Add explanatory comments for DOM shim (runtime WASM-bindgen, not test mocks)

Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>

.cbmignore

@@ -1,201 +0,0 @@
-# codebase-memory-mcp ignore list
-#
-# Padrão gitignore-style. Linhas começando com `#` são comentários.
-# Barra final (`/`) = só diretório. Sem barra = casa arquivo OU diretório.
-#
-# O CBM também lê `.gitignore` automaticamente — esta lista deixa explícito o que
-# os hooks do CBM vão pular. Se uma regra entrar em conflito entre os dois arquivos,
-# vale a união. Editar este arquivo é mais barato do que confiar na herança implícita.
-#
-# Última reconciliação: 2026-07-31, status `ready` (513k nodes / 689k edges),
-# `auto_index_limit=50000`, total indexável medido ≈11.546 arquivos (folga 4,3×).
-#
-# Fontes cruzadas:
-#   - `codebase-memory-mcp cli index_status --project home-diegosouzapw-dev-proxys-OmniRoute`
-#     → `not_indexed.dirs` (27) + `not_indexed.files` (336), todos `BY DESIGN`.
-#   - `.gitignore` deste repo (5.691 B) — fonte canônica secundária.
-#
-# Como auditar mudanças: depois de editar este arquivo, rodar `index_repository`
-# (ou esperar `auto_watch` re-indexar) e re-checar `cli index_status` → comparar
-# contagens em `not_indexed.dirs_count` e `not_indexed.files_count`.
-
-# ─────────────────────────────────────────────────────────────────────────────
-# 1. Diretorios de runtime / pacote — nao sao codigo-fonte
-# ─────────────────────────────────────────────────────────────────────────────
-node_modules/
-node_modules
-
-# Builds e artefatos reproduziveis (Layer 1 Next.js / Electron)
-.build/
-dist/
-.next/
-out/
-
-# Electron especifico
-electron/dist-electron/
-electron/node_modules/
-icon.iconset/
-
-# Workspaces internos que tem proprio node_modules
-@omniroute/opencode-plugin/dist/
-@omniroute/opencode-plugin/node_modules/
-@omniroute/opencode-provider/dist/
-@omniroute/opencode-provider/node_modules/
-
-# Recursos nativos compilados (C/JNI/wasm)
-src/mitm/tproxy/native/build/
-
-# Artefatos locais do Stryker / Playwright / coverage
-.stryker-tmp/
-reports/mutation/
-stryker-output-*.json
-.playwright-mcp/
-test-results/
-playwright-report/
-blob-report/
-
-# Analise / linters / caches
-.analysis/
-.sisyphus/
-.plans/
-.gitnexus
-.worktrees
-.codegraph/
-
-# Quality artifacts (gerados por npm run lint --cache etc)
-.eslintcache
-.eslintcache-complexity
-
-# Claude Code local state
-.claude/scheduled_tasks.lock
-.claude/scheduled_tasks/
-.claude/sessions/
-.claude/state.json
-.claude/settings.local.json
-
-# Serena / Antigravity / outras tools locais
-.serena/
-.antigravitycli/
-.gemini/
-.config/
-
-# ─────────────────────────────────────────────────────────────────────────────
-# 2. Diretorios com prefixo `_` — locais / privados (regra global do .gitignore)
-# ─────────────────────────────────────────────────────────────────────────────
-_*/
-_artifacts/
-_cache/
-_mono_repo/
-_references/
-_tasks/
-
-# ─────────────────────────────────────────────────────────────────────────────
-# 3. Diretorios de tooling IA (state local, nao codigo)
-# ─────────────────────────────────────────────────────────────────────────────
-.agents/
-.claude/
-.vscode/
-.idea/
-.junie/
-.omc/
-.data/
-.data-dev/
-.local-data/
-.logs/
-.artifacts/
-.source/
-.superpowers/
-.claude-flow/
-.omnivscodeagent/
-omnirouteCloud/
-omnirouteSite/
-.omniroute/
-.stent/
-
-# Subpaths especificos do Claude Code que nao estao em .claude/ (criados sob repo)
-.claude/worktrees/
-
-# ─────────────────────────────────────────────────────────────────────────────
-# 4. Diretorios de dados / runtime locais (storage, env, secrets, scratch)
-# ─────────────────────────────────────────────────────────────────────────────
-data/
-# NOTA: src/lib/env/, src/app/api/{cloud,sync/cloud,system/env,agent-skills/coverage}/
-# foram removidos daqui (2026-08-05). Os nomes sugerem dados/segredos locais, mas os
-# 8 arquivos sao route handlers e modulos rastreados no git — escondia-los do grafo
-# criava pontos cegos em buscas e em analise de impacto.
-tests/golden-set/data/
-
-# Logs e saida de teste
-logs/*
-test_output.log
-home-diegosouzapw-dev-automacoes-*.txt
-
-# ─────────────────────────────────────────────────────────────────────────────
-# 5. Diretorios do monorepo por subprojeto (nao fazem parte do app principal)
-# ─────────────────────────────────────────────────────────────────────────────
-security-analysis/
-vscode-extension/
-obsidian-plugin/node_modules/
-
-# ─────────────────────────────────────────────────────────────────────────────
-# 6. Diretorios de documentacao interna / workflow
-# ─────────────────────────────────────────────────────────────────────────────
-docs/superpowers/
-# Docs traduzidas: 1.215 arquivos / 94 MB (inclui 20+ copias do CHANGELOG).
-# Sao traducoes do tree em ingles, ja indexado — no grafo so geram ruido em
-# search_code e consomem o auto_index_limit.
-docs/i18n/
-
-# ─────────────────────────────────────────────────────────────────────────────
-# 7. Arquivos especificos (nao diretorios inteiros)
-# ─────────────────────────────────────────────────────────────────────────────
-
-# Segredos e env — NUNCA indexar
-.env
-.env.*
-!.env.example
-!.env.homolog.example
-
-# TypeScript build info e next env declaration
-*.tsbuildinfo
-next-env.d.ts
-typescript
-
-# SQLite transient files (WAL/SHM/journal)
-*.sqlite-shm
-*.sqlite-wal
-*.sqlite-journal
-
-# Mapas e source maps
-*.map
-
-# Bun / npm lockfiles ruidosos
-bun.lock
-
-# `cheaper-inference-gateway.svg` e arquivos de midia na raiz/asset ja cobertos
-# pelos `ignored-suffix` do indexador (svg/png/jpg/ico/etc >50kB ou >500linhas);
-# manter a regra explicita aqui ajuda a auditar:
-cheaper-inference-gateway.svg
-cheaper-inference-gateway-*.svg
-
-# Husky internals
-.husky/_/
-
-# CI / quality metric artifacts
-config/quality/quality-metrics.json
-config/quality/test-impact-map.json
-audit-report.json
-.gh-discussions.json
-
-# i18n audit (gerado por npm run scripts)
-scripts/i18n/_audit.json
-scripts/i18n/_pending-keys.json
-
-# NOTA: bin/omniroute.mjs foi removido daqui (2026-08-05). Estava marcado como
-# "scratch", mas e o entrypoint real do CLI publicado (package.json -> bin.omniroute)
-# e consta em PACK_ARTIFACT_REQUIRED_PATHS. Precisa estar no grafo.
-
-# Deploy / docker backups
-deploy.sh
-docker-compose.yml.bak
-docker-compose.minimal.yml

.dockerignore

@@ -7,13 +7,7 @@
 **/.vscode
 
 # Dependencies and build output
-# `node_modules` alone matches the ROOT only — Docker's matcher does not cross
-# `/` like .gitignore does. Without the `**/` form, nested installs ship in the
-# build context (e.g. @omniroute/opencode-provider/node_modules, ~79 MB of
-# devDependencies). Both forms are kept: the bare one is the documented root
-# rule, the `**/` one covers every nested package.
 node_modules
-**/node_modules
 .next
 .build
 out
@@ -43,17 +37,6 @@ tests
 test-results
 playwright-report
 blob-report
-output
-.playwright-cli
-.playwright-mcp
-.stryker-tmp
-reports/mutation
-
-# Local caches and quality-gate artifacts (all gitignored). `_*` does not match
-# dot-prefixed names, so these need explicit entries.
-.artifacts
-.eslintcache
-.eslintcache-complexity
 
 # Documentation
 # Issue #2348: The Dashboard Docs viewer reads markdown from `/app/docs` at
@@ -66,10 +49,6 @@ reports/mutation
 # (English) sources at runtime, so translations are not required in the
 # container image.
 docs/i18n/**
-# Internal planning artifacts (gitignored). `*.md` above only matches the root,
-# so without this rule these land in /app/docs and become readable through the
-# dashboard's Docs viewer at runtime.
-docs/superpowers/**
 docs/diagrams/**/*.png
 docs/diagrams/**/*.jpg
 docs/diagrams/**/*.jpeg

.env.example

@@ -84,12 +84,7 @@ PORT=20128
 # endpoint display (useDisplayBaseUrl) shows https://host/omniroute/v1 instead of
 # https://host/v1. Rebuild after changing this value (Next basePath is build-time).
 # Default: "" (served at the domain root). Example: /omniroute to serve under https://host/omniroute
-# Docker: baked at image build time via build-arg; root-path images can also apply this at
-# container start (see docs/guides/DOCKER_GUIDE.md).
 # OMNIROUTE_BASE_PATH=
-# Client fetch/EventSource under this path are rewritten via installBasePathFetch
-# (src/shared/utils/basePathFetch.ts) so absolute `/api/*` and `/v1/*` hits work
-# without a reverse-proxy rewrite. Rebuild after changing (Next basePath is build-time).
 #
 # Browser-visible mirror of OMNIROUTE_BASE_PATH, inlined at build time so the
 # dashboard endpoint display can read it client-side. Set it to the same value
@@ -240,9 +235,8 @@ NODE_ENV=production
 
 # Container runtime — controls startup script behavior (permissions, advice).
 # Values: docker | podman  |  Default: docker
-# Set to "podman" for any Podman topology. The entrypoint cannot determine
-# whether the engine is local or reached through Podman Machine, so it prints
-# topology-neutral guidance and links contrib/podman/README.md.
+# Set to "podman" when running under rootless Podman so the entrypoint
+# gives the correct fix instructions (podman unshare chown vs sudo chown).
 CONTAINER_HOST=docker
 
 # Container runtime override for skill sandboxing.
@@ -811,15 +805,6 @@ PROVIDER_LIMITS_SYNC_SPACING_MS=1500
 # Default: <repo>/src/lib/db/migrations.
 #OMNIROUTE_MIGRATIONS_DIR=
 
-# Additional migration directories, as `namespace=dir` entries separated by the
-# platform path delimiter (`:` on POSIX, `;` on Windows). Files found there are
-# recorded as `<namespace>-<number>` (e.g. `ee-134`), a version space that cannot
-# collide with the upstream numeric slots — so a distribution shipping its own
-# migrations never silently loses one to a number the upstream set also claimed.
-# A malformed entry, an invalid namespace or a missing directory aborts startup
-# rather than skipping the schema. Unset = no extra directories (the default).
-#OMNIROUTE_EXTRA_MIGRATIONS_DIRS=ee=/opt/app/enterprise/db/migrations
-
 # Mass-pending-migrations safety threshold (#3416). If more than this many
 # migrations are pending on an existing DB, startup aborts (a wiped tracking
 # table could cause data loss). Raise it to restore an older backup; set to 0
@@ -1414,6 +1399,10 @@ APP_LOG_TO_FILE=true
 # Default: ~/.omniroute/plugins/  Override in dev/CI to point at a local plugin tree.
 # OMNIROUTE_PLUGIN_PATH=
 
+# Allow plugins to request the 'exec' permission (spawn child processes from the
+# plugin worker sandbox). Disabled by default; set to 1 to enable (local operator only).
+# OMNIROUTE_PLUGINS_ALLOW_EXEC=0
+
 # ── Prompt cache (system prompt deduplication) ──
 # Used by: open-sse/services — caches identical system prompts across requests.
 # PROMPT_CACHE_MAX_SIZE=50              # Max cached entries (default: 50)
@@ -1826,10 +1815,6 @@ APP_LOG_TO_FILE=true
 # Tokens reserved for completion output when computing prompt budgets.
 # Used by: open-sse/services/contextManager.ts. Default: 1024.
 # CONTEXT_RESERVE_TOKENS=1024
-# How many of the newest inline images to keep when pruning older ones to fit
-# the context window (#8560). Used by: open-sse/services/contextManager.ts.
-# Default: 2.
-# CONTEXT_KEEP_LATEST_IMAGES=2
 
 # ── Model alias rewriting (legacy compatibility) ──
 # Toggle the legacy model-alias compatibility layer used by older clients.

.github/dependabot.yml

@@ -39,17 +39,6 @@ updates:
       # the duplication gate — migrate the gate intentionally, not via dependabot.
       - dependency-name: "jscpd"
         update-types: ["version-update:semver-major"]
-      # ioredis is a SOFT/optional dependency loaded through a dynamic import
-      # (src/lib/quota/redisQuotaStore.ts — "Redis driver requires ioredis package"),
-      # so a breaking major never fails at build or typecheck time: the only consumers
-      # are the distributed quota store (redisQuotaStore.ts, storeFactory.ts) and the
-      # `import type Redis` in src/shared/utils/rateLimiter.ts. Nothing in the unit or
-      # vitest suites exercises a live Redis connection, so a v5→v6 API break would ship
-      # green and only surface at runtime for operators running distributed quota — the
-      # exact users least able to absorb it. #9310 grouped that major with 9 harmless
-      # bumps; majors here need their own PR and a deliberate migration review.
-      - dependency-name: "ioredis"
-        update-types: ["version-update:semver-major"]
       # @huggingface/transformers is HARD-PINNED at 3.5.2 (exact, no caret) — FROZEN.
       # It is load-bearing for the LLMLingua ONNX compression engine (open-sse/services/
       # compression/engines/llmlingua/ — worker.ts pins @huggingface/transformers@3.5.2)

.github/pull_request_template.md

@@ -9,14 +9,11 @@
 
 ## Validation
 
-Choose the change type and focused loop from the
-[Contribution Golden Path](../docs/dev/CONTRIBUTION_GOLDEN_PATH.md). The full unit suite,
-Vitest, the 60% coverage gate, and the production build all run in CI on this PR (#8329):
+Run only the focused loop for what you changed — the full unit suite, Vitest, the
+60% coverage gate, and the production build all run in CI on this PR (#8329):
 
-- [ ] Change type: provider / routing / UI / i18n / CLI / DB / build-deploy / other
-- [ ] Focused tests and category gates from the golden path
+- [ ] Focused tests for the change: `node --import tsx/esm --test tests/unit/<file>.test.ts`
 - [ ] `npm run lint`
-- [ ] Reconciled with the current active release base; focused checks rerun afterward
 - [ ] Production-code changes include a new or updated automated test in this PR
 - [ ] SonarQube PR analysis is green or any remaining issues are explicitly documented below
 
@@ -32,4 +29,4 @@ Vitest, the 60% coverage gate, and the production build all run in CI on this PR
 
 ## Reviewer Notes
 
-- Call out any risky areas, migrations, feature flags, or manual validation that reviewers should know about.
+- Call out any risky areas, migrations, feature flags, or manual validation that reviewers should know about.

.github/workflows/build-fork.yml

@@ -0,0 +1,71 @@
+name: Publish Fork Image to GHCR
+
+on:
+  push:
+    branches: [main]
+    tags:
+      - "v*"
+  workflow_dispatch:
+
+# Least-privilege default: read-only at the top level; the build job that pushes to
+# GHCR grants packages: write itself (Scorecard TokenPermissions).
+permissions:
+  contents: read
+
+env:
+  IMAGE_NAME: ghcr.io/kang-heewon/omniroute
+
+jobs:
+  build:
+    name: Build and Push Fork Image
+    if: github.repository == 'kang-heewon/OmniRoute'
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      packages: write
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v7
+        with:
+          persist-credentials: false
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v4
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v4
+
+      - name: Login to GitHub Container Registry
+        uses: docker/login-action@v4
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract Docker metadata
+        id: meta
+        uses: docker/metadata-action@v6
+        with:
+          images: ${{ env.IMAGE_NAME }}
+          tags: |
+            type=raw,value=latest,enable={{is_default_branch}}
+            type=sha,prefix=sha-
+            type=ref,event=tag
+          labels: |
+            org.opencontainers.image.title=omniroute
+            org.opencontainers.image.description=Unified AI proxy/router — fork image
+            org.opencontainers.image.url=https://github.com/kang-heewon/OmniRoute
+            org.opencontainers.image.source=https://github.com/kang-heewon/OmniRoute
+            org.opencontainers.image.licenses=MIT
+
+      - name: Build and push
+        uses: docker/build-push-action@v7
+        with:
+          context: .
+          target: runner-base
+          platforms: linux/amd64,linux/arm64
+          push: true
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          cache-from: type=gha
+          cache-to: type=gha,mode=max

.github/workflows/ci.yml

@@ -35,25 +35,13 @@ jobs:
       workflow: ${{ steps.classify.outputs.workflow }}
       testsOnly: ${{ steps.classify.outputs.testsOnly }}
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0
         with:
           persist-credentials: false
           fetch-depth: 0
       - uses: actions/setup-node@v7
         with:
           node-version: ${{ env.CI_NODE_VERSION }}
-      # Refuse a PR that targets its own head branch before spending anything on it. #8912 has
-      # head == base == release/v3.8.50: no diff, can never merge, and it sits in the queue with
-      # a full check board attached on every push to that branch. One field comparison.
-      - name: Reject a PR that targets its own branch
-        if: github.event_name == 'pull_request'
-        env:
-          HEAD_REF: ${{ github.head_ref }}
-          BASE_REF: ${{ github.base_ref }}
-          HEAD_SHA: ${{ github.event.pull_request.head.sha }}
-          BASE_SHA: ${{ github.event.pull_request.base.sha }}
-        run: node scripts/check/check-pr-self-target.mjs
-
       - id: classify
         env:
           EVENT_NAME: ${{ github.event_name }}
@@ -137,17 +125,6 @@ jobs:
       - run: npm run check:route-guard-membership
       - run: npm run check:test-discovery
       - run: npm run check:tracked-artifacts
-      # (gap 30) Also lives in quality.yml's PR-only "Merge integrity" job — because the
-      # CHANGELOG half of that job needs a base to diff against. This half does NOT: the
-      # generator either reproduces the committed SKILL.md files or it does not.
-      #
-      # Keeping it PR-only left a real hole. This cycle's merge trains landed in batches with
-      # `--admin`, which bypasses required checks, so three SKILL.md files drifted from the route
-      # catalog, rode the release squash into `main`, and the next cycle's sync-back turned them
-      # into a base-red that blocked EVERY PR into release/v3.8.50 until #8954. Running it here
-      # means a push to `main` catches the drift at the source instead of the next cycle
-      # inheriting it.
-      - run: npm run check:agent-skills-sync
       # WS1.7 (v3.8.49 plan): Dockerfile lint (hadolint, pinned by digest).
       # failure-threshold=error keeps the 5 pre-existing warnings (DL3008/DL3003/
       # DL3016 version pinning / WORKDIR) visible without blocking; any ERROR fails.
@@ -353,16 +330,8 @@ jobs:
           install -m 0755 /tmp/osv/*linux_amd64 "$HOME/.local/bin/osv-scanner"
           # actionlint — official download script
           bash <(curl -fsSL https://raw.githubusercontent.com/rhysd/actionlint/main/scripts/download-actionlint.bash) latest "$HOME/.local/bin"
-          # zizmor — PyPI (pipx preferred, pip --user fallback); lands in ~/.local/bin.
-          # PINNED on purpose. Unpinned, the runner installed whatever PyPI served that day and
-          # measured 1 finding MORE than the devbox on the identical commit (190 vs 189) during
-          # the v3.8.49 cycle — which cost a second rebaseline push per release, chasing a
-          # number that was never the code's. The ratchet compares counts across machines, so
-          # the auditor version has to be the same on both. Bump this deliberately, and
-          # rebaseline in the same commit: check-workflows.mjs now prints `zizmorVersion=` next
-          # to the count so the new number is traceable to the tool that produced it.
-          ZIZMOR_VERSION=1.25.2
-          pipx install "zizmor==$ZIZMOR_VERSION" || pip install --user "zizmor==$ZIZMOR_VERSION"
+          # zizmor — PyPI (pipx preferred, pip --user fallback); lands in ~/.local/bin
+          pipx install zizmor || pip install --user zizmor
           # oasdiff — download latest linux amd64 tarball via gh (authed), extract binary
           rm -rf /tmp/oasd && mkdir -p /tmp/oasd
           gh release download --repo oasdiff/oasdiff --pattern '*linux_amd64.tar.gz' --dir /tmp/oasd
@@ -480,26 +449,15 @@ jobs:
     # UI keys move with dashboard code OR message catalogs.
     if: ${{ github.event_name != 'pull_request' || (github.event.pull_request.draft == false && (needs.changes.outputs.i18n == 'true' || needs.changes.outputs.code == 'true')) }}
     steps:
-      # fetch-depth: 0 — the value-drift gate diffs en.json against the merge base to
-      # find rewritten English strings. On a shallow clone the base ref is missing and
-      # the gate self-skips (base-unresolved), so it would never actually run.
       - uses: actions/checkout@v7
         with:
           persist-credentials: false
-          fetch-depth: 0
       - uses: actions/setup-node@v7
         with:
           node-version: ${{ env.CI_NODE_VERSION }}
           cache: npm
       - uses: ./.github/actions/npm-ci-retry
       - run: node scripts/i18n/check-ui-keys-coverage.mjs --threshold=65
-      # #8463: a rewritten English value used to leave its 39 translations behind
-      # silently (googleOAuthWarning shipped wrong copy in 39 locales for months).
-      # Key parity above cannot see it — a stale translation counts as covered.
-      - name: i18n UI value drift (stale translations)
-        env:
-          BASE_REF: ${{ github.base_ref && format('origin/{0}', github.base_ref) || '' }}
-        run: node scripts/i18n/check-ui-value-drift.mjs
 
   # #8038: cheap single-locale glossary/protected-terms consistency gate —
   # complements i18n-ui-coverage (key parity) and the ICU `i18n` job below
@@ -751,13 +709,7 @@ jobs:
   test-unit:
     name: Unit Tests (${{ matrix.shard }}/8)
     # Same dynamic-runner rule as Build (own-origin only; fallback ubuntu-latest).
-    # PINNED to hosted, deliberately not on the USE_VPS_RUNNER switch (gap 19). One variable
-    # governed the build and the test jobs, which want OPPOSITE machines: the build needs the
-    # .113's RAM, the tests need the hosted runner's link. Measured on 2026-07-29 —
-    # actions/setup-node took 20m06s on .113 with 4 concurrent runners versus 16s hosted (npm
-    # cache restore saturating the link), while the tests themselves tied, 2m54 vs 2m31. So
-    # self-hosted is strictly worse here and there is nothing to configure.
-    runs-on: ubuntu-latest
+    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && (github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository)) && fromJSON('["self-hosted","omni-release"]') || 'ubuntu-latest' }}
     timeout-minutes: 25
     # needs: changes (not build) — this job never downloads the next-build artifact;
     # gating it on Build only serialized ~20min of wall-clock for nothing. Jobs that
@@ -829,13 +781,7 @@ jobs:
   test-vitest:
     name: Vitest (MCP / autoCombo / UI components)
     # Same dynamic-runner rule as Build (own-origin only; fallback ubuntu-latest).
-    # PINNED to hosted, deliberately not on the USE_VPS_RUNNER switch (gap 19). One variable
-    # governed the build and the test jobs, which want OPPOSITE machines: the build needs the
-    # .113's RAM, the tests need the hosted runner's link. Measured on 2026-07-29 —
-    # actions/setup-node took 20m06s on .113 with 4 concurrent runners versus 16s hosted (npm
-    # cache restore saturating the link), while the tests themselves tied, 2m54 vs 2m31. So
-    # self-hosted is strictly worse here and there is nothing to configure.
-    runs-on: ubuntu-latest
+    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && (github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository)) && fromJSON('["self-hosted","omni-release"]') || 'ubuntu-latest' }}
     timeout-minutes: 15
     # needs: changes (not build) — no artifact consumed; see test-unit note.
     needs: changes
@@ -1169,8 +1115,7 @@ jobs:
           if FILES=$(node scripts/quality/balance-e2e-shards.mjs "$SHARD" 9); then
             if [ -z "$FILES" ]; then echo "[e2e-balance] shard $SHARD has no files"; exit 0; fi
             echo "[e2e-balance] shard $SHARD runs:"; echo "$FILES"
-            # FILES is our own newline-separated path list, so word-splitting is intended
-            # shellcheck disable=SC2086,SC2046
+            # shellcheck disable=SC2086 — FILES is our own newline-separated path list
             npx playwright test $(echo "$FILES" | tr '\n' ' ') --reporter=line,junit
           else
             echo "[e2e-balance] balancer unavailable — plain --shard fallback"
@@ -1272,8 +1217,6 @@ jobs:
       - name: Generate dashboard
         env:
           EVENT_NAME: ${{ github.event_name }}
-          # Workflow-controlled data (job results), not user input — safe to read here.
-          NEEDS_JSON: ${{ toJSON(needs) }}
         run: |
           status() {
             case "$1" in
@@ -1288,29 +1231,6 @@ jobs:
           echo "# 🚀 CI Dashboard" >> "$GITHUB_STEP_SUMMARY"
           echo "" >> "$GITHUB_STEP_SUMMARY"
 
-          # (gap 12) A cancelled job never reported a verdict, and in a long table that reads the
-          # same as a green one. `cancel-in-progress` plus incremental fixing cancels jobs on every
-          # push, and this cycle the Vitest job was cancelled in rounds 1, 2 and 3 — it only ran to
-          # completion in round 4, where it revealed a suite that had been broken the whole cycle
-          # plus two production bugs. A gate that never finishes is indistinguishable from one that
-          # passes, so name them at the TOP instead of leaving them to be spotted mid-table.
-          CANCELLED_JOBS=$(printf '%s' "$NEEDS_JSON" \
-            | jq -r 'to_entries | map(select(.value.result == "cancelled")) | .[].key' 2>/dev/null \
-            | sort | paste -sd", " -) || CANCELLED_JOBS=""
-          if [ -n "$CANCELLED_JOBS" ]; then
-            {
-              echo "> ### ⚫ Cancelled — no verdict was reported"
-              echo ">"
-              echo "> \`$CANCELLED_JOBS\`"
-              echo ">"
-              echo "> These did not fail; they never finished, so nothing was checked. Treat this"
-              echo "> run as INCOMPLETE for those gates. If the cancellation came from"
-              echo "> \`cancel-in-progress\` on a newer push, the newer run covers it — otherwise"
-              echo "> re-run them before reading this dashboard as green."
-              echo ""
-            } >> "$GITHUB_STEP_SUMMARY"
-          fi
-
           echo "## 🧱 Core Checks" >> "$GITHUB_STEP_SUMMARY"
           echo "| Job | Status |" >> "$GITHUB_STEP_SUMMARY"
           echo "|-----|--------|" >> "$GITHUB_STEP_SUMMARY"

.github/workflows/codeql.yml

@@ -19,7 +19,7 @@ jobs:
       actions: read
       contents: read
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
         with:
           persist-credentials: false
       - uses: github/codeql-action/init@e4fba868fa4b1b91e1fdab776edc8cfbe6e9fb81 # v4.37.3

.github/workflows/dast-smoke.yml

@@ -30,7 +30,7 @@ jobs:
       JWT_SECRET: ci-build-secret-with-sufficient-length-for-validation
       API_KEY_SECRET: ci-api-key-secret-with-sufficient-length-aaaa
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
         with:
           persist-credentials: false
       - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
@@ -59,14 +59,8 @@ jobs:
       - run: pip install schemathesis
       - name: Schemathesis smoke (high-risk endpoints, blocking)
         run: |
-          # /api/auth/oidc/* is a BROWSER redirect flow (302 to the IdP, 302 back to
-          # /login?oidc_error=... on every failure), not a REST endpoint: Schemathesis reads
-          # those 302s as "the API accepted a schema-violating request" and the configured-off
-          # 400 as "rejected a schema-compliant request". Documenting the flow in the spec is
-          # still right (operators need it); fuzzing it is not what this smoke is for.
           schemathesis run docs/openapi.yaml --url http://localhost:20128 \
             --include-path-regex '^/v1/(chat/completions|models)$|^/api/(auth|keys)' \
-            --exclude-path-regex '^/api/auth/oidc/' \
             --max-examples 8 --workers 4 --checks all --max-response-time 30 \
             --request-timeout 20 --suppress-health-check all --no-color
       - name: promptfoo injection-guard (blocking)

.github/workflows/docker-publish.yml

@@ -155,13 +155,13 @@ jobs:
         uses: docker/setup-buildx-action@v4
 
       - name: Login to Docker Hub
-        uses: docker/login-action@v4.5.2
+        uses: docker/login-action@v4
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
 
       - name: Login to GitHub Container Registry
-        uses: docker/login-action@v4.5.2
+        uses: docker/login-action@v4
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
@@ -255,13 +255,13 @@ jobs:
         uses: docker/setup-buildx-action@v4
 
       - name: Login to Docker Hub
-        uses: docker/login-action@v4.5.2
+        uses: docker/login-action@v4
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
 
       - name: Login to GitHub Container Registry
-        uses: docker/login-action@v4.5.2
+        uses: docker/login-action@v4
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
@@ -390,7 +390,7 @@ jobs:
       - name: Upload Trivy SARIF to Security tab
         if: needs.prepare.outputs.version != 'main'
         continue-on-error: true
-        uses: github/codeql-action/upload-sarif@v4.37.3
+        uses: github/codeql-action/upload-sarif@v4
         with:
           sarif_file: trivy-results.sarif
           category: trivy-image

.github/workflows/electron-release.yml

@@ -120,18 +120,6 @@ jobs:
         env:
           JWT_SECRET: ci-build-secret-with-sufficient-length-for-validation
           NODE_OPTIONS: "--max_old_space_size=6144"
-          # Linux builds with webpack, not Turbopack. Turbopack's production build
-          # allocates natively (Rust, off the V8 heap), so --max_old_space_size does
-          # not bound it, and on this module graph it peaks above what the hosted
-          # runner can give — the VM is reclaimed mid-compile with "The runner has
-          # received a shutdown signal", no exit code. That is what silently took the
-          # whole desktop channel out of v3.8.49: the linux leg died, `release` was
-          # skipped, and the release shipped with ZERO assets. Measured on a 32 GB
-          # box the same build passes and peaks past 14 GB. The webpack fallback is
-          # the project's documented escape hatch for RAM-constrained machines
-          # (docs/reference/ENVIRONMENT.md, #6409) and is the same remedy already
-          # applied to nightly-compat's Node 26 build (#8090).
-          OMNIROUTE_USE_TURBOPACK: ${{ matrix.platform == 'linux' && '0' || '1' }}
         run: npm run build
 
       - name: Sync version in electron/package.json
@@ -229,16 +217,6 @@ jobs:
   release:
     name: Create Release
     needs: [validate, build]
-    # Fail-partial, not fail-closed. `build` is a 4-leg matrix with `fail-fast: false`,
-    # so the legs that succeed still upload their artifacts — but a default `needs:`
-    # gate skips this job the moment ANY leg fails, discarding all of them. That is
-    # exactly what happened to v3.8.49: the linux leg died and the release shipped with
-    # ZERO assets, throwing away 1.7 GB of good Windows/macOS installers **and** the
-    # source archives + SBOM, which do not depend on a build at all. The result was
-    # indistinguishable from "this version has no desktop channel".
-    # Now: attach everything that did build, then fail the job loudly (see the last
-    # step) so an incomplete channel is visible instead of silent.
-    if: ${{ !cancelled() && needs.validate.result == 'success' }}
     runs-on: ubuntu-latest
     permissions:
       contents: write # softprops/action-gh-release creates the GitHub Release
@@ -249,33 +227,11 @@ jobs:
           persist-credentials: false
           fetch-depth: 0
 
-      # `merge-multiple` is deliberately OFF. It resolves same-name collisions by ARRIVAL
-      # ORDER, and the two macOS jobs each emit their own `latest-mac.yml` listing only their
-      # own dmg (measured: 338 and 350 bytes, different content, identical name). One silently
-      # overwrote the other — arm64 won in the published v3.8.48, and since the Intel dmg
-      # carries no arch suffix in its name, electron-updater's
-      # `files.find(url includes process.arch) ?? files.shift()` sends every Intel Mac to the
-      # ARM dmg. Downloading into per-artifact subdirectories keeps both, so they can be
-      # merged on purpose instead of by luck.
       - name: Download all artifacts
         uses: actions/download-artifact@v8
         with:
-          path: artifacts
-
-      # Writes release-assets/latest-mac.yml with BOTH dmgs, un-suffixed entry first (that is
-      # the one electron-updater can only reach through its fallback). Refuses to write when the
-      # inputs disagree on version — a manifest stitched from two builds is worse than none.
-      - name: Merge the per-arch macOS updater manifests
-        run: node scripts/release/merge-mac-update-manifest.mjs artifacts release-assets
-
-      # Everything else moves across as-is. The partial latest-mac.yml files are excluded so
-      # they cannot clobber the merged one; -n is a second belt on the same braces.
-      - name: Collect the remaining artifacts
-        run: |
-          mkdir -p release-assets
-          find artifacts -type f ! -name latest-mac.yml -exec cp -n {} release-assets/ \;
-          echo "release-assets:"
-          ls -la release-assets/
+          path: release-assets
+          merge-multiple: true
 
       - name: Create source archives
         env:
@@ -319,47 +275,6 @@ jobs:
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
 
-  verify-desktop-assets:
-    name: Verify desktop assets landed
-    needs: [validate, release]
-    # Deliberately a SEPARATE job, not a final step of `release`: failing inside
-    # `release` would cascade into `publish-npm` (which gates on `needs: release`) and
-    # block the npm channel over a desktop-only gap. Here the assets are attached, npm
-    # still publishes, and an incomplete desktop channel shows up as a red job instead
-    # of passing unnoticed — the v3.8.49 release had ZERO assets and every gate was
-    # green, because nothing ever asserted the release HAS binaries.
-    if: ${{ !cancelled() && needs.release.result == 'success' }}
-    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-    steps:
-      - name: Assert every platform is present on the release
-        env:
-          # Regex-validated (^v[0-9]+\.[0-9]+\.[0-9]+$) in the `validate` job, and
-          # passed via env rather than interpolated into the script body.
-          VERSION: ${{ needs.validate.outputs.version }}
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          names=$(gh release view "$VERSION" --repo "$GITHUB_REPOSITORY" \
-                    --json assets --jq '.assets[].name')
-          echo "Assets on $VERSION:"
-          echo "$names" | sed 's/^/  /'
-
-          missing=""
-          # `[ ... ] && missing=...` as the last command in a branch returns 1 and
-          # would abort the whole script under Actions' default `set -e`. Use if/fi.
-          for want in '\.exe$' '\.dmg$' '\.AppImage$' '\.deb$' '^latest.*\.yml$' '\.source\.tar\.gz$'; do
-            if ! echo "$names" | grep -qE "$want"; then
-              missing="$missing $want"
-            fi
-          done
-
-          if [ -n "$missing" ]; then
-            echo "::error::Desktop channel incomplete on $VERSION — no asset matching:$missing"
-            exit 1
-          fi
-          echo "✓ every platform present on $VERSION"
-
   publish-npm:
     name: Publish to npm
     needs: [validate, release]

.github/workflows/nightly-compat.yml

@@ -71,20 +71,7 @@ jobs:
           node-version: "26"
           cache: npm
       - uses: ./.github/actions/npm-ci-retry
-      # #8090 — this is the ONLY Node 26 build in the whole CI matrix (ci.yml pins
-      # CI_NODE_VERSION=24). It failed every nightly with the runner-reclaimed
-      # signature ("The runner has received a shutdown signal" / "The operation was
-      # canceled", no exit code) always at the same Turbopack compile phase — a
-      # classic OOM kill on the memory-constrained ubuntu-latest runner. Turbopack's
-      # native (Rust, off-V8-heap) allocation is NOT bounded by --max-old-space-size
-      # and peaks far higher than webpack on this large module graph (#6409), and is
-      # heavier still under Node 26. Use the documented webpack fallback here: it still
-      # validates that the app *builds* on Node 26 (the point of this compat job) at a
-      # much lower memory peak. Turbopack-on-Node-24 stays covered by ci.yml's build
-      # job. See docs/reference/ENVIRONMENT.md (OMNIROUTE_USE_TURBOPACK) and #6409.
       - run: npm run build
-        env:
-          OMNIROUTE_USE_TURBOPACK: "0"
 
   compat-tests:
     name: Node ${{ matrix.node }} Compat Tests (${{ matrix.shard }}/4)

.github/workflows/nightly-mutation.yml

@@ -148,7 +148,7 @@ jobs:
     runs-on: ubuntu-latest
     timeout-minutes: 15
     steps:
-      - uses: actions/checkout@v7
+      - uses: actions/checkout@v6
         with:
           persist-credentials: false
       - uses: actions/setup-node@v7

.github/workflows/nightly-release-green.yml

@@ -143,8 +143,7 @@ jobs:
             MODE="--with-build --full-ci"
           fi
           echo "[release-green] mode: $MODE (event: $EVENT_NAME)"
-          # MODE is an intentional flag list, so word-splitting is wanted here
-          # shellcheck disable=SC2086
+          # shellcheck disable=SC2086 — MODE is an intentional flag list
           node scripts/quality/validate-release-green.mjs --json --hermetic $MODE \
             1> release-green.json 2> release-green.log
           echo "exit=$?" >> "$GITHUB_OUTPUT"
@@ -249,8 +248,7 @@ jobs:
             MODE="--with-build --full-ci"
           fi
           echo "[main-green] mode: $MODE (event: $EVENT_NAME)"
-          # MODE is an intentional flag list, so word-splitting is wanted here
-          # shellcheck disable=SC2086
+          # shellcheck disable=SC2086 — MODE is an intentional flag list
           node scripts/quality/validate-release-green.mjs --json --hermetic $MODE \
             1> main-green.json 2> main-green.log
           echo "exit=$?" >> "$GITHUB_OUTPUT"
@@ -303,145 +301,3 @@ jobs:
             main-green.json
             main-green.log
           if-no-files-found: ignore
-
-  # ── Banking lane (#8584) ──────────────────────────────────────────────────
-  # The ratchet is asymmetric: RAISING a cap is a ten-second manual JSON edit made
-  # under merge pressure, LOWERING one requires someone to run `--update` and commit
-  # — which no workflow does. Grep `.github/workflows/` for `--update`: only
-  # wiki-sync.yml (unrelated) and ci.yml's check-quality-ratchet.mjs --require-tighten
-  # (a different script, a different metric). So a cap outlives the code that earned
-  # it and every completed decomposition silently becomes a growth allowance for
-  # whoever touches the file next. Measured on 2026-07-25: 18 frozen files already at
-  # or under the 800-line new-file cap, up to 132x (schemas.ts, 19 lines / 2523 cap),
-  # and 31 unfulfilled "tighten via --update next cycle" notes honoured exactly once
-  # (-1 unit) in six weeks.
-  #
-  # This job closes that loop by making the DOWNWARD direction as automatic as the
-  # upward one is easy. It measures the active release branch, runs the shrink-only
-  # `--update` paths, and opens ONE always-current PR with the result. It never
-  # pushes to release/* — a human still merges, so a bad measurement cannot land
-  # unreviewed. verify-ratchet-bank.mjs is the hard guarantee that the automation can
-  # only ever write in the shrink direction; if anything was raised, added, or a
-  # rebaseline note was touched, the job aborts and opens nothing.
-  #
-  # Schedule/dispatch only, deliberately NOT on push: banking has no latency
-  # requirement (a shrink banked within 8h is fine) and a per-merge run would rebuild
-  # the PR branch repeatedly during merge campaigns while paying for a full ESLint
-  # walk each time. Detection stays on push (release-green above); banking is batched.
-  bank-ratchet-shrinks:
-    name: Bank ratchet shrinks
-    if: ${{ github.event_name != 'push' }}
-    runs-on: ubuntu-latest
-    permissions:
-      contents: write
-      pull-requests: write
-    steps:
-      - uses: actions/checkout@v7
-        with:
-          fetch-depth: 0
-
-      - name: Resolve active release branch
-        id: branch
-        env:
-          INPUT_BRANCH: ${{ github.event.inputs.branch }}
-        run: |
-          set -euo pipefail
-          if [ -n "${INPUT_BRANCH:-}" ]; then
-            TARGET="$INPUT_BRANCH"
-          else
-            TARGET=$(git for-each-ref --format='%(refname:short)' 'refs/remotes/origin/release/v*' \
-              | sed 's#origin/##' \
-              | sort -t/ -k2 -V \
-              | tail -1)
-          fi
-          if [ -z "$TARGET" ]; then echo "No release/v* branch found"; exit 1; fi
-          # Same strict guard as the validation job — blocks ref/command injection
-          # through the workflow_dispatch input.
-          if ! printf '%s' "$TARGET" | grep -qE '^release/v[0-9]+\.[0-9]+\.[0-9]+$'; then
-            echo "Refusing non-canonical branch name: $TARGET"; exit 1
-          fi
-          echo "target=$TARGET" >> "$GITHUB_OUTPUT"
-          echo "bank_branch=chore/bank-ratchet-${TARGET#release/}" >> "$GITHUB_OUTPUT"
-
-      - name: Checkout the release branch
-        env:
-          TARGET: ${{ steps.branch.outputs.target }}
-        run: |
-          set -euo pipefail
-          git checkout "$TARGET"
-          git log -1 --oneline
-
-      - uses: actions/setup-node@v7
-        with:
-          node-version: "24"
-          cache: npm
-
-      - uses: ./.github/actions/npm-ci-retry
-
-      - name: Ratchet the baselines down
-        run: |
-          # Both --update paths are shrink-only by construction (file-size writes only
-          # on `improvements`, complexity-ratchets only when `.improved`), and both exit
-          # non-zero while the branch is over baseline — which is exactly when there is
-          # nothing to bank. Their exit code is not the signal; the verifier below is.
-          set +e
-          node scripts/check/check-file-size.mjs --update
-          node scripts/check/check-complexity-ratchets.mjs --update
-          exit 0
-
-      - name: Verify the write only went downward
-        id: verify
-        run: |
-          set -euo pipefail
-          # Exits 1 if ANYTHING was raised/added or a rebaseline note was touched.
-          # `set -e` then aborts the job before a commit exists — no PR is opened.
-          node scripts/quality/verify-ratchet-bank.mjs > bank-summary.md
-          if [ -n "$(git status --porcelain config/quality/)" ]; then
-            echo "changed=true" >> "$GITHUB_OUTPUT"
-          else
-            echo "changed=false" >> "$GITHUB_OUTPUT"
-            echo "Nothing to bank — baselines already match the code."
-          fi
-
-      - name: Open / update the banking PR
-        if: steps.verify.outputs.changed == 'true'
-        env:
-          GH_TOKEN: ${{ github.token }}
-          TARGET: ${{ steps.branch.outputs.target }}
-          BANK_BRANCH: ${{ steps.branch.outputs.bank_branch }}
-          RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-        run: |
-          set -euo pipefail
-          git config user.name "github-actions[bot]"
-          git config user.email "41898282+github-actions[bot]@users.noreply.github.com"
-          git checkout -B "$BANK_BRANCH"
-          git add config/quality/
-          git commit -m "chore(quality): bank ratchet shrinks measured on ${TARGET}"
-          git push --force origin "$BANK_BRANCH"
-
-          {
-            echo "Automated banking of quality-ratchet **shrinks** already present in"
-            echo "\`${TARGET}\` — the downward half of the ratchet, which nothing else runs (#8584)."
-            echo ""
-            echo "Produced by \`check:file-size --update\` + \`check:complexity-ratchets --update\`,"
-            echo "then verified by \`scripts/quality/verify-ratchet-bank.mjs\`: **nothing was raised,"
-            echo "nothing was added, no rebaseline note was touched** — the job aborts without"
-            echo "opening a PR if any of those is violated."
-            echo ""
-            echo "No product code changes. Merging retires growth allowances that the code no"
-            echo "longer needs; not merging leaves them available to whoever edits those files next."
-            echo ""
-            cat bank-summary.md
-            echo ""
-            echo "**Run:** ${RUN_URL}"
-          } > pr-body.md
-
-          EXISTING=$(gh pr list --repo "$GITHUB_REPOSITORY" --head "$BANK_BRANCH" \
-            --state open --json number --jq '.[0].number' 2>/dev/null || echo "")
-          if [ -n "$EXISTING" ]; then
-            gh pr edit "$EXISTING" --repo "$GITHUB_REPOSITORY" --body-file pr-body.md
-            echo "Updated existing PR #$EXISTING"
-          else
-            gh pr create --repo "$GITHUB_REPOSITORY" --base "$TARGET" --head "$BANK_BRANCH" \
-              --title "chore(quality): bank ratchet shrinks (${TARGET})" --body-file pr-body.md
-          fi

.github/workflows/npm-publish.yml

@@ -56,15 +56,8 @@ env:
 
 jobs:
   publish:
-    # Same dynamic-runner rule as ci.yml's `build`/`test-unit`: `build:cli` falls back to a
-    # full `next build`, whose working set outgrew the 16 GB hosted runner during the
-    # v3.8.49 cycle — the publish died with "The runner has received a shutdown signal"
-    # mid-"Creating an optimized production build" while v3.8.48 had still fit in 16min.
-    # This job never runs on `pull_request`, so the fork-safety clause is always true here;
-    # it is kept verbatim so the expression stays greppable against ci.yml.
-    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && (github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository)) && fromJSON('["self-hosted","omni-release"]') || 'ubuntu-latest' }}
+    runs-on: ubuntu-latest
     permissions:
-      actions: read # find + download the CI run's next-build artifact for this SHA
       contents: write # gh release upload (attach SBOM to the GitHub Release)
       id-token: write # npm provenance
       packages: write # publish to npm.pkg.github.com
@@ -152,74 +145,6 @@ jobs:
         run: |
           npm version "$VERSION" --no-git-tag-version --allow-same-version
 
-      # Fast path: CI already built the standalone tree for THIS commit and uploaded it as
-      # `next-build`. `build:cli` (scripts/build/prepublish.ts) only shells out to a full
-      # `next build` when `.build/next/standalone/server.js` is missing — restoring the
-      # artifact turns the heaviest step of the publish into a download. Matching on
-      # `head_sha` is the tree-equality guarantee: same commit, same tree.
-      # Best-effort by design (retention is 1 day): every miss falls through to the build
-      # step below, which is why the dynamic runner above matters as the backstop.
-      #
-      # The `head_repository.full_name == env.REPO` clause is a supply-chain guard, not a
-      # filter refinement. This artifact becomes the published npm tarball. `pull_request`
-      # runs from forks execute in THIS repository's context and upload their own
-      # `next-build` built from fork-controlled source, and the runs API returns them for a
-      # matching `head_sha` — 57 such runs exist in this repo today. Without the clause,
-      # anything that made a fork's head commit coincide with the publish commit could put
-      # attacker-built bytes on npm. Requiring the run to originate from this repository
-      # excludes every fork run while keeping the fast path intact (verified: the same
-      # single run is selected either way for the current tip).
-      # CodeQL: actions/artifact-poisoning/critical.
-      - name: Reuse CI's next-build artifact (skips the heavy rebuild)
-        if: steps.resolve.outputs.skip != 'true'
-        continue-on-error: true
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          HEAD_SHA: ${{ github.sha }}
-          REPO: ${{ github.repository }}
-        run: |
-          set -uo pipefail
-          # The question is "which run HAS the artifact", not "which run passed" (gap 16).
-          # Requiring `conclusion == "success"` on the whole run discarded a perfectly good tree
-          # whenever any unrelated shard went red — one flaky test then pushed the publish into
-          # the 40-minute build this step exists to avoid. The artifact is only uploaded if the
-          # Build job itself succeeded, so its PRESENCE is the accurate signal; the run's overall
-          # conclusion is noise from jobs that have nothing to do with the tree.
-          #
-          # `head_repository.full_name == env.REPO` stays, and it is not a filter refinement:
-          # this tree becomes the published npm tarball, and fork `pull_request` runs execute in
-          # THIS repository's context uploading their own next-build. That clause is the
-          # supply-chain guard (CodeQL actions/artifact-poisoning).
-          CANDIDATES=$(gh api "repos/$REPO/actions/runs?head_sha=$HEAD_SHA&per_page=100" \
-            --jq '[.workflow_runs[]
-                   | select(.name == "CI"
-                            and .head_repository.full_name == env.REPO)]
-                  | sort_by(.run_started_at) | reverse | .[0:5] | .[].id') || CANDIDATES=""
-          if [ -z "$CANDIDATES" ]; then
-            echo "::notice::no CI run from this repository for $HEAD_SHA — falling back to a full build"
-            exit 0
-          fi
-          RUN=""
-          for candidate in $CANDIDATES; do
-            if gh run download "$candidate" --repo "$REPO" --name next-build --dir /tmp/next-build 2>/dev/null; then
-              RUN="$candidate"
-              break
-            fi
-            echo "  run $candidate carries no usable next-build — trying the next"
-          done
-          if [ -z "$RUN" ]; then
-            echo "::notice::none of the candidate runs still carries next-build (1-day retention) — falling back to a full build"
-            exit 0
-          fi
-          tar -xzf /tmp/next-build/e2e-build.tar.gz -C .
-          rm -rf /tmp/next-build
-          if [ -f .build/next/standalone/server.js ]; then
-            echo "✅ standalone tree restored from CI run $RUN — build:cli will skip next build"
-          else
-            echo "::notice::extract did not yield .build/next/standalone — falling back to a full build"
-            rm -rf .build
-          fi
-
       - name: Build CLI bundle (standalone app)
         if: steps.resolve.outputs.skip != 'true'
         env:
@@ -256,18 +181,6 @@ jobs:
         if: steps.resolve.outputs.skip != 'true'
         run: npm run check:pack-boot
 
-      # The boot-smoke above proves a CLEAN install boots. It does not prove the path that
-      # actually broke us: installing over an existing version, where ~110 SQLite migrations
-      # run against a populated database. v3.8.48 shipped as a hotfix because the published
-      # 3.8.47 crashed on boot, and the v3.8.49 upgrade path was first exercised end-to-end
-      # by hand on a real 3.8.48 box (VPS .16) — after publishing, which is exactly backwards.
-      # Runs BEFORE `npm stage publish` so a broken upgrade never reaches the registry at all;
-      # a staged package that is never approved simply expires, with no `npm deprecate` needed.
-      - name: Prove clean-install AND upgrade-over-previous both boot
-        if: steps.resolve.outputs.skip != 'true'
-        timeout-minutes: 30
-        run: npm run check:install-upgrade
-
       # WS1.3 (D2, v3.8.49 plan): STAGED publishing by default — `npm stage publish`
       # parks the exact bytes on the registry WITHOUT making them installable; the
       # owner then verifies and approves with 2FA (`npm stage approve`), moving the

.github/workflows/quality.yml

@@ -57,31 +57,6 @@ jobs:
           git diff --name-only "$BASE_SHA" "$HEAD_SHA" > changed-files.txt
           node scripts/quality/classify-pr-changes.mjs changed-files.txt >> "$GITHUB_OUTPUT"
 
-  build:
-    name: Build (advisory)
-    needs: changes
-    if: ${{ github.event_name != 'pull_request' || ((github.event.pull_request.draft == false || startsWith(github.head_ref, 'mergify/merge-queue/')) && needs.changes.outputs.code == 'true') }}
-    # Dynamic runner — same fork-safe rule as ci.yml / fast-gates.
-    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && (github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository)) && fromJSON('["self-hosted","omni-release"]') || 'ubuntu-latest' }}
-    # #7307: advisory for the first week of release-PR runs; remove
-    # continue-on-error after the production-build signal is stable.
-    continue-on-error: true
-    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
-        with:
-          persist-credentials: false
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
-        with:
-          node-version: ${{ env.CI_NODE_VERSION }}
-          cache: npm
-      - uses: ./.github/actions/npm-ci-retry
-      - run: npm run check:node-runtime
-      - run: npm run build
-        env:
-          OMNIROUTE_USE_TURBOPACK: "1"
-      # No artifact upload here: the PR-to-release quality workflow has no
-      # downstream package/e2e jobs that consume the Next.js build output.
-
   # Docs/OpenAPI contract gates only — existence reason is doc accuracy + route refs.
   # Split out of fast-gates so pure-docs PRs skip typecheck/unit while still validating docs.
   docs-gates:
@@ -112,20 +87,7 @@ jobs:
     # release captain has USE_VPS_RUNNER=true AND this is not a fork PR (own-origin
     # branches only — a fork PR must never execute on the LAN runner). Var unset/false
     # or a fork PR falls back to ubuntu-latest, so this is inert until the flag flips.
-    # PINNED to hosted (gap 19). This job carried the USE_VPS_RUNNER expression, and that
-    # expression was DEAD CONFIGURATION: across 160 quality.yml runs the job never once landed on
-    # a self-hosted runner — every non-skipped sample is `GitHub Actions NNNN`. The classifier is
-    # not at fault: in the same window ci.yml's Build demonstrably ran on omniroute-113-7 and
-    # omniroute-113-6, so self-hosted runs are visible when they happen.
-    #
-    # And if it ever HAD fired it would have inherited the measured penalty, because this job's
-    # first two steps are exactly the bottleneck: actions/setup-node + npm ci took 20m06s on .113
-    # with 4 concurrent runners versus 16s hosted (npm cache restore saturating the link). Median
-    # here is 5.6 min hosted across 72 successful runs.
-    #
-    # With this pinned, USE_VPS_RUNNER governs ONLY build-like jobs — one variable, one coherent
-    # purpose. That is what gap 19 asked for; a second variable turned out to be unnecessary.
-    runs-on: ubuntu-latest
+    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && (github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository)) && fromJSON('["self-hosted","omni-release"]') || 'ubuntu-latest' }}
     # tsx gates (known-symbols, route-guard-membership) import modules that open
     # SQLite on load; provide DB env so a fresh CI DB initializes cleanly.
     env:
@@ -155,18 +117,7 @@ jobs:
       - run: npm run check:fetch-targets
       # docs-all / openapi-routes / docs-symbols live in docs-gates (path-filtered).
       - run: npm run check:deps
-      # #8522: --base-ref mode for PR events — compare against max(frozen, base) so
-      # inherited drift (base already over frozen cap) doesn't red an innocent PR.
-      # workflow_dispatch (no PR base) falls back to absolute comparison.
-      - name: File-size ratchet (base-relative on PR)
-        env:
-          PR_BASE_SHA: ${{ github.event.pull_request.base.sha }}
-        run: |
-          if [ -n "$PR_BASE_SHA" ]; then
-            npm run check:file-size -- --base-ref "$PR_BASE_SHA"
-          else
-            npm run check:file-size
-          fi
+      - run: npm run check:file-size
       - run: npm run check:error-helper
       - run: npm run check:migration-numbering
       - run: npm run check:public-creds
@@ -190,81 +141,6 @@ jobs:
       # Complexity + cognitive-complexity: ONE ESLint walk (both baselines still
       # enforced separately by ruleId). Avoids two cold tree walks on fast-path.
       - run: npm run check:complexity-ratchets
-      # ── G0 (trilho .50): gates do trilho A que faltavam no trilho B ──────────────
-      # The god-file refactor happens in PRs→release/**; without these, the release
-      # rail never sees a new import cycle, dead code, duplication or a security
-      # regression until the release PR to main. Deliberately NOT brought here:
-      # bundle-size (self-skips without a build — this rail's build job is advisory
-      # and uploads nothing, so it would be dead configuration) and the coverage
-      # run (fast-unit already runs the full suite; the coverage ratchet stays on
-      # the main rail via --allow-missing in lint-guard).
-      - run: npm run check:cycles
-      - run: npm run check:lockfile
-      - name: Duplication ratchet
-        run: npm run check:duplication
-      - name: Dead-code ratchet (knip)
-        run: npm run check:dead-code
-      - name: Type coverage ratchet
-        run: npm run check:type-coverage
-      - name: Compression budget ratchet
-        run: npm run check:compression-budget
-      # Security scanners — same hardened install as ci.yml quality-extended
-      # (gh release download = authenticated, 5000 req/hr; curl to api.github.com
-      # is rate-limited to 60/hr and silently no-ops when throttled). The blocking
-      # gates below SKIP (exit 0) when their binary is absent — only a measured
-      # regression vs config/quality/quality-baseline.json blocks.
-      - name: Install security scanners (gitleaks/osv/actionlint/zizmor/oasdiff)
-        continue-on-error: true
-        env:
-          GH_TOKEN: ${{ github.token }}
-        run: |
-          set +e
-          mkdir -p "$HOME/.local/bin"
-          # Ratchets compare scanner COUNTS across runs. Pin every auditor: a rule-set
-          # update must be an explicit PR that re-measures/rebaselines, never a random
-          # red (or green) caused by whatever "latest" served that morning.
-          GITLEAKS_VERSION=v8.30.1
-          OSV_SCANNER_VERSION=v2.3.8
-          ACTIONLINT_VERSION=v1.7.12
-          ZIZMOR_VERSION=1.25.2
-          OASDIFF_VERSION=v1.19.1
-          # gitleaks — pinned linux x64 tarball via gh (authed), extract binary
-          rm -rf /tmp/gl && mkdir -p /tmp/gl
-          gh release download "$GITLEAKS_VERSION" --repo gitleaks/gitleaks --pattern '*linux_x64.tar.gz' --dir /tmp/gl
-          tar -xzf /tmp/gl/*linux_x64.tar.gz -C "$HOME/.local/bin" gitleaks
-          # osv-scanner — pinned linux amd64 bare binary via gh (authed)
-          rm -rf /tmp/osv && mkdir -p /tmp/osv
-          gh release download "$OSV_SCANNER_VERSION" --repo google/osv-scanner --pattern '*linux_amd64' --dir /tmp/osv
-          install -m 0755 /tmp/osv/*linux_amd64 "$HOME/.local/bin/osv-scanner"
-          # actionlint — official installer from a pinned release tag (never main)
-          bash <(curl -fsSL "https://raw.githubusercontent.com/rhysd/actionlint/${ACTIONLINT_VERSION}/scripts/download-actionlint.bash") "$ACTIONLINT_VERSION" "$HOME/.local/bin"
-          # zizmor — pinned PyPI package (same version as ci.yml quality-extended)
-          pipx install "zizmor==$ZIZMOR_VERSION" || pip install --user "zizmor==$ZIZMOR_VERSION"
-          # oasdiff — pinned linux amd64 tarball via gh (authed), extract binary
-          rm -rf /tmp/oasd && mkdir -p /tmp/oasd
-          gh release download "$OASDIFF_VERSION" --repo Tufin/oasdiff --pattern '*linux_amd64.tar.gz' --dir /tmp/oasd
-          tar -xzf /tmp/oasd/*linux_amd64.tar.gz -C "$HOME/.local/bin" oasdiff
-          # ALWAYS export the bin dir (even if any step above failed)
-          echo "$HOME/.local/bin" >> "$GITHUB_PATH"
-          "$HOME/.local/bin/gitleaks" version || true
-          "$HOME/.local/bin/actionlint" -version || true
-          "$HOME/.local/bin/osv-scanner" --version || true
-          "$HOME/.local/bin/oasdiff" --version || true
-          zizmor --version || true
-      - name: Secret scan (gitleaks, ratchet, blocking)
-        run: npm run check:secrets -- --ratchet
-      - name: Vulnerability ratchet (osv-scanner, ratchet, blocking)
-        run: npm run check:vuln-ratchet -- --ratchet
-      - name: Workflow lint (actionlint+zizmor, ratchet, blocking)
-        run: npm run check:workflows -- --ratchet
-      # BASE_REF is read by the script from the env (never interpolated into a
-      # shell body) — workflow-injection-safe. actions/checkout fetches remote
-      # refs, not a local branch named github.base_ref, so prefix origin/ or this
-      # gate self-skips every PR with reason=base-unresolved.
-      - name: OpenAPI breaking-change (oasdiff, ratchet, blocking)
-        env:
-          BASE_REF: ${{ github.base_ref && format('origin/{0}', github.base_ref) || '' }}
-        run: npm run check:openapi-breaking -- --ratchet
       - name: Typecheck (core)
         run: npm run typecheck:core
       # #7033: dashboard-scoped typecheck gate — src/app/(dashboard) TSX is not
@@ -359,13 +235,7 @@ jobs:
     needs: changes
     if: ${{ github.event_name != 'pull_request' || ((github.event.pull_request.draft == false || startsWith(github.head_ref, 'mergify/merge-queue/')) && needs.changes.outputs.code == 'true') }}
     # Dynamic runner — see fast-gates (own-origin + flag; fork/unset → ubuntu-latest).
-    # PINNED to hosted, deliberately not on the USE_VPS_RUNNER switch (gap 19). One variable
-    # governed the build and the test jobs, which want OPPOSITE machines: the build needs the
-    # .113's RAM, the tests need the hosted runner's link. Measured on 2026-07-29 —
-    # actions/setup-node took 20m06s on .113 with 4 concurrent runners versus 16s hosted (npm
-    # cache restore saturating the link), while the tests themselves tied, 2m54 vs 2m31. So
-    # self-hosted is strictly worse here and there is nothing to configure.
-    runs-on: ubuntu-latest
+    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && (github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository)) && fromJSON('["self-hosted","omni-release"]') || 'ubuntu-latest' }}
     env:
       JWT_SECRET: ci-lint-secret-with-sufficient-length-for-validation
       API_KEY_SECRET: ci-lint-api-key-secret-long
@@ -401,13 +271,7 @@ jobs:
     # critical path again (~8.5min → ~4.5min on ubuntu-latest; ~2min on the 8-slot
     # runner box). Node's native --test-shard=N/total takes any denominator — only
     # this matrix and the TEST_SHARD env below encode the shard count.
-    # PINNED to hosted, deliberately not on the USE_VPS_RUNNER switch (gap 19). One variable
-    # governed the build and the test jobs, which want OPPOSITE machines: the build needs the
-    # .113's RAM, the tests need the hosted runner's link. Measured on 2026-07-29 —
-    # actions/setup-node took 20m06s on .113 with 4 concurrent runners versus 16s hosted (npm
-    # cache restore saturating the link), while the tests themselves tied, 2m54 vs 2m31. So
-    # self-hosted is strictly worse here and there is nothing to configure.
-    runs-on: ubuntu-latest
+    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && (github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository)) && fromJSON('["self-hosted","omni-release"]') || 'ubuntu-latest' }}
     strategy:
       fail-fast: false
       matrix:
@@ -450,12 +314,6 @@ jobs:
     if: ${{ github.event_name != 'pull_request' || ((github.event.pull_request.draft == false || startsWith(github.head_ref, 'mergify/merge-queue/')) && needs.changes.outputs.code == 'true') }}
     runs-on: ubuntu-latest
     continue-on-error: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.repo.fork == true }}
-    # G0 (trilho .50): security-events:read lets the CodeQL ratchet below read open
-    # code-scanning alerts via `gh api .../code-scanning/alerts` (same as ci.yml's
-    # quality-gate job). contents: read keeps checkout working.
-    permissions:
-      contents: read
-      security-events: read
     steps:
       - uses: actions/checkout@v7
         with:
@@ -477,29 +335,6 @@ jobs:
       - name: ESLint (baseline congelado — warning novo = vermelho)
         # lint:json writes the report; --max-warnings 0 keeps no-new-warnings policy.
         run: npm run lint:json -- --max-warnings 0
-      # ── G0 (trilho .50): motor de ratchet também no trilho B ─────────────────────
-      # This job just wrote .artifacts/eslint-results.json — collect-metrics prefers
-      # that file, so the ratchet engine lands here at ZERO extra ESLint cost (one
-      # inventory, two consumers; same reason ci.yml chains lint → quality-gate).
-      # The coverage-report artifact does not exist on this rail, so both ratchet
-      # invocations run --allow-missing: coverage.* metrics skip gracefully while
-      # the deterministic ones (eslint / openapi-coverage / i18n-ui) stay BLOCKING.
-      # Coverage authority remains on the main rail (ci.yml test-coverage → quality-gate).
-      - run: npm run quality:collect
-      - name: Ratchet check (blocking)
-        run: node scripts/quality/check-quality-ratchet.mjs --allow-missing --summary .artifacts/quality-ratchet.md
-      - name: Require-tighten (blocking)
-        run: node scripts/quality/check-quality-ratchet.mjs --allow-missing --require-tighten
-      # CodeQL alerts ratchet — same semantics as ci.yml quality-gate: exits 1 ONLY
-      # on a real regression (open alerts > baseline in quality-baseline.json);
-      # a measurement failure (gh/auth/api) self-skips with exit 0.
-      - name: CodeQL alerts ratchet (blocking)
-        run: npm run check:codeql-ratchet
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      - name: Append ratchet summary
-        if: always()
-        run: cat .artifacts/quality-ratchet.md >> "$GITHUB_STEP_SUMMARY" || true
 
   # Merge-integrity: pega no PR os dois vazamentos crônicos de merge que hoje só
   # explodem na release-PR. (1) CHANGELOG-eat — o auto-resolve do merge come
@@ -523,7 +358,7 @@ jobs:
       API_KEY_SECRET: ci-lint-api-key-secret-long
       DISABLE_SQLITE_AUTO_BACKUP: "true"
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
         with:
           fetch-depth: 0
           persist-credentials: false

.github/workflows/semgrep.yml

@@ -18,7 +18,7 @@ jobs:
     container:
       image: semgrep/semgrep
     steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
         with:
           persist-credentials: false
       - name: Run semgrep (advisory)

.gitignore

@@ -171,6 +171,7 @@ config/quality/test-impact-map.json
 # GitNexus local index
 .gitnexus
 .worktrees
+bin/omniroute.mjs
 
 # Consistent with .dockerignore / .npmignore
 .omc/
@@ -200,10 +201,7 @@ scripts/i18n/_pending-keys.json
 .codegraph/
 
 # Fumadocs generated source
-/.source/
-
-# Temporary local worktrees used to build unpublished npm tarballs
-/.deploy-build-*/
+.source/
 
 # AI agent local settings and configs
 .agents/
@@ -235,10 +233,7 @@ omniroute.md
 
 # mise configuration
 mise.toml
-# release-green artifacts (.gitignore has no inline comments — a trailing
-# `# ...` becomes part of the pattern, so it must sit on its own line).
-# Already covered by /_*/ above; kept explicit for discoverability.
-_artifacts/
+_artifacts/          # release-green artifacts
 .claude-flow/
 
 # ESLint file cache (npm run lint --cache / complexity ratchets)
@@ -255,8 +250,3 @@ tests/homolog/.auth/
 tests/homolog/ui/.auth/
 homolog-report/
 docker-compose.yml.bak
-.playwright-cli/
-# Playwright screenshot/log output. Today every artifact happens to land inside
-# output/**/.playwright-cli/ (covered above), but anything written directly to
-# output/ would otherwise show up as untracked.
-/output/

.gitleaks.toml

@@ -86,10 +86,4 @@
     regexes = [
       '''latencyP\d{2}Ms''',
       '''interleaved-thinking-2025-05-14''',
-      # v3.8.49 pre-flight (2026-07-28). Nenhum dos dois e credencial:
-      # - chave de localStorage do banner de patrocinio (#8723), so um identificador de UI;
-      # - x-api-key PUBLICO do Firefly web (documentado em open-sse/utils/publicCreds.ts:207);
-      #   as duas ocorrencias sinalizadas estao em COMENTARIOS JSDoc, o runtime le de resolvePublicCred().
-      '''omniroute-kimi-sponsor-banner-dismissed-v1''',
-      '''SunbreakWebUI1''',
     ]

.mergify.yml

@@ -17,13 +17,6 @@
 #   • Fallback path if Mergify misbehaves or the OSS plan changes: the manual
 #     merge-train runbook (docs/ops/MERGE_TRAIN.md) — remove labels, proceed by hand.
 
-# Auto-enqueue (current Mergify model, 2026): auto_merge_conditions in
-# merge_protections_settings — the rules-based queue action / autoqueue path is
-# deprecated (EOL 2026-07-16). The owner-applied `queue` label IS the approval.
-merge_protections_settings:
-  auto_merge_conditions:
-    - label = queue
-
 queue_rules:
   - name: release
     # Any current or future release branch — the reason GitHub's native queue was
@@ -41,26 +34,14 @@ queue_rules:
     # is intentionally NOT a condition here: the owner-applied `queue` label IS the
     # approval in this repo's single-maintainer model (see governance header).
     merge_conditions:
-      # "Zero failures" — EXCEPT the advisory "Build (advisory)" job (quality.yml):
-      # continue-on-error by design, and its GH-hosted Turbopack build hangs
-      # recurrently mid-"Creating an optimized production build" (100% failure rate
-      # across every sampled PR since the job was added 2026-07-27, always killed by
-      # a runner timeout/shutdown signal, never a real compile error). Any OTHER
-      # failure still blocks (anti-fail-open kept). The prior dast-smoke exception
-      # (#7225) was dropped here: dast-smoke's hang (#7226) has been dormant for
-      # weeks (0 failures in the last 30 runs; 2 all-time, none since 2026-07-13) —
-      # carrying its tolerance forward would mask problems it no longer causes.
-      - or:
-          - "#check-failure=0"
-          - and:
-              - "#check-failure=1"
-              - check-failure=Build (advisory)
+      - "#check-failure=0"
       - "#check-pending=0"
       - "#check-success>=1"
       - check-success=Merge integrity (changelog + generated skills)
-    # NO batching: 'Merge Queue Batch' requires a paid Mergify tier (live finding
-    # 2026-07-15 — the queue command fails with "Cannot use Merge Queue batch" on
-    # the free plan). Serial queue (1 PR at a time) still automates the train.
+    # Batching: validate up to 10 queued PRs together (the manual train's sweet spot);
+    # don't hold a lone PR hostage waiting for siblings.
+    batch_size: 10
+    batch_max_wait_time: 5 min
     # Squash keeps the one-commit-per-PR history the CHANGELOG reconciliation expects.
     merge_method: squash
 

.npmignore

@@ -4,14 +4,11 @@ data/
 **/db.json
 
 # VS Code extension test runtime (large binary, not needed in npm package)
+app/vscode-extension/
 **/data/
 **/db.json
 
-# Source code (pre-built dist/ is published instead)
-#
-# NOTA (2026-08-05): as entradas `app/*` foram removidas — o diretorio `app/`
-# foi renomeado para `dist/` na Layer 1 e nao existe mais. Elas sugeriam um
-# layout que ja nao e o do projeto.
+# Source code (pre-built app/ is published instead)
 #
 # NOTE (#3578 / #3821-review): package.json "files" is the source of truth for what
 # ships. It now allowlists the backend source closure the MCP server needs at runtime
@@ -52,6 +49,8 @@ scripts/
 .vscode/
 .agents/
 .env*
+app/.env
+app/.env*
 eslint.config.mjs
 prettier.config.mjs
 postcss.config.mjs
@@ -83,6 +82,8 @@ bun.lock
 *.deb
 *.rpm
 electron/
+app/electron/
+app/vscode-extension/
 
 # Subprojects
 clipr/
@@ -92,6 +93,10 @@ vscode-extension/
 
 # Root-level underscore-prefixed directories (private/draft — never publish)
 /_*/
+app/_*/
+app/coverage/
+app/logs/
+app/tests/
 
 # Consistent with .gitignore and .dockerignore
 .DS_Store

.prettierignore

@@ -1,11 +1,6 @@
 # Long reference tables are manually aligned; formatting the whole file causes noisy diffs.
 docs/reference/ENVIRONMENT.md
 
-# Generated by `npm run gen:provider-reference`; the generator aligns the tables and
-# is their formatter of record. Without this, lint-staged reformats the file whenever
-# it is staged and the next generator run reverts it — a diff ping-pong.
-docs/reference/PROVIDER_REFERENCE.md
-
 # Dense auto-generated free-tier budget rows (one object per line) — prettier multi-line expand blows past file-size cap 800.
 open-sse/config/freeModelCatalog.data.ts
 

.source/dynamic.ts

@@ -0,0 +1,8 @@
+// @ts-nocheck
+import { dynamic } from 'fumadocs-mdx/runtime/dynamic';
+import * as Config from '../source.config';
+
+const create = await dynamic<typeof Config, import("fumadocs-mdx/runtime/types").InternalTypeConfig & {
+  DocData: {
+  }
+}>(Config, {"configPath":"source.config.ts","environment":"next","outDir":".source"}, {"doc":{"passthroughs":["extractedReferences"]}});

.source/source.config.mjs

@@ -0,0 +1,22 @@
+// source.config.ts
+import { defineDocs, defineConfig } from "fumadocs-mdx/config";
+var docs = defineDocs({
+  dir: "docs",
+  docs: {
+    files: [
+      "./architecture/**/*.md",
+      "./guides/**/*.md",
+      "./reference/**/*.md",
+      "./frameworks/**/*.md",
+      "./routing/**/*.md",
+      "./security/**/*.md",
+      "./compression/**/*.md",
+      "./ops/**/*.md"
+    ]
+  }
+});
+var source_config_default = defineConfig();
+export {
+  source_config_default as default,
+  docs
+};

@omniroute/opencode-plugin/package.json

@@ -23,7 +23,7 @@
   "scripts": {
     "build": "tsup",
     "clean": "rm -rf dist",
-    "test": "node --import tsx/esm --test tests/scaffold.test.ts tests/auth.test.ts tests/options-schema.test.ts tests/multi-instance.test.ts tests/fetch-interceptor.test.ts tests/provider.test.ts tests/gemini-sanitize.test.ts tests/combos.test.ts tests/config-shim.test.ts tests/features.test.ts tests/feature-defaults.test.ts tests/usable-combo.test.ts tests/disk-snapshot-perms.test.ts tests/fork-features.test.ts tests/auto-combo-context.test.ts tests/provider-id-routing.test.ts tests/management-read-token.test.ts tests/auto-sync.test.ts",
+    "test": "node --import tsx/esm --test tests/scaffold.test.ts tests/auth.test.ts tests/options-schema.test.ts tests/multi-instance.test.ts tests/fetch-interceptor.test.ts tests/provider.test.ts tests/gemini-sanitize.test.ts tests/combos.test.ts tests/config-shim.test.ts tests/features.test.ts tests/usable-combo.test.ts tests/disk-snapshot-perms.test.ts tests/fork-features.test.ts tests/auto-combo-context.test.ts tests/provider-id-routing.test.ts tests/management-read-token.test.ts tests/auto-sync.test.ts",
     "prepublishOnly": "npm run clean && npm run build && npm test"
   },
   "keywords": [

@omniroute/opencode-plugin/src/index.ts

@@ -216,59 +216,6 @@ const optionsSchema = z
  */
 export type OmniRoutePluginOptions = z.infer<typeof optionsSchema>;
 
-/**
- * Explicit default state for every boolean `features.*` toggle.
- *
- * #7624: `featuresSchema` marks every flag `.optional()` with no default, and
- * the effective value is applied implicitly at each read site (default-ON flags
- * use the `features.X !== false` convention, default-OFF flags use
- * `features.X === true`). That implicit convention is scattered across the file,
- * so an operator who omits the `features` block cannot tell whether
- * combos / autoCombos / enrichment are enabled — they think features are
- * disabled when they are actually on. Centralising the declared defaults here
- * (mirroring the read-site conventions exactly, so runtime behaviour is
- * unchanged) makes the effective flags introspectable and self-documenting.
- */
-export const OMNIROUTE_FEATURE_DEFAULTS = {
-  // default-ON (read sites use `features.X !== false`)
-  combos: true,
-  autoCombos: true,
-  enrichment: true,
-  diskCache: true,
-  providerTag: true,
-  fetchInterceptor: true,
-  geminiSanitization: true,
-  // default-OFF (read sites use `features.X === true`)
-  compressionMetadata: false,
-  usableOnly: false,
-  mcpAutoEmit: false,
-  debugLog: false,
-  startupDebug: false,
-} as const;
-
-/** Union of the boolean feature-flag keys declared in `OMNIROUTE_FEATURE_DEFAULTS`. */
-export type OmniRouteFeatureFlag = keyof typeof OMNIROUTE_FEATURE_DEFAULTS;
-
-/**
- * Resolve the EFFECTIVE boolean state of every feature toggle, applying the
- * declared default for any flag the operator omitted. A missing `features`
- * block (or an empty one) yields the full default set — so callers and the
- * startup diagnostics can surface exactly which features are active instead of
- * relying on the implicit `!== false` / `=== true` conventions. Purely
- * derived: it does not mutate options nor change any read-site behaviour.
- */
-export function resolveEffectiveFeatureFlags(
-  features?: OmniRoutePluginOptions["features"]
-): Record<OmniRouteFeatureFlag, boolean> {
-  const f: Partial<Record<OmniRouteFeatureFlag, boolean>> = features ?? {};
-  const out = {} as Record<OmniRouteFeatureFlag, boolean>;
-  for (const key of Object.keys(OMNIROUTE_FEATURE_DEFAULTS) as OmniRouteFeatureFlag[]) {
-    const val = f[key];
-    out[key] = typeof val === "boolean" ? val : OMNIROUTE_FEATURE_DEFAULTS[key];
-  }
-  return out;
-}
-
 export const OMNIROUTE_PROVIDER_KEY = "omniroute" as const;
 
 /** Deployed plugin version (injected at build time by tsup define). */
@@ -2053,7 +2000,6 @@ async function writeStartupDiagnostics(params: {
   autoComboCount: number;
   enrichment: OmniRouteEnrichmentMap;
   autoCombos: OmniRouteRawAutoCombo[];
-  features?: OmniRoutePluginOptions["features"];
 }): Promise<void> {
   const {
     providerId,
@@ -2064,7 +2010,6 @@ async function writeStartupDiagnostics(params: {
     autoComboCount,
     enrichment,
     autoCombos,
-    features,
   } = params;
   const enriched = [...enrichment.entries()];
   const withName = enriched.filter(([, e]) => e.name);
@@ -2077,16 +2022,6 @@ async function writeStartupDiagnostics(params: {
   lines.push(
     `models=${modelCount} combos=${comboCount} enrichment=${enrichmentSize} autoCombos=${autoComboCount}`
   );
-  // #7624: surface the EFFECTIVE feature flags so an operator who omitted the
-  // `features` block can see combos/autoCombos/enrichment are on (the counts
-  // above can read 0 for reasons unrelated to the flags — e.g. missing auth).
-  const effectiveFlags = resolveEffectiveFeatureFlags(features);
-  lines.push(
-    `features(effective): ` +
-      (Object.keys(effectiveFlags) as OmniRouteFeatureFlag[])
-        .map((k) => `${k}=${effectiveFlags[k] ? "on" : "off"}`)
-        .join(" ")
-  );
   lines.push(
     `enrichment: ${withName.length} with name, ${withPricing.length} with pricing, ${withFree.length} free`
   );
@@ -3195,7 +3130,6 @@ export function createOmniRouteProviderHook(
             autoComboCount: rawAutoCombos.length,
             enrichment: rawEnrichment,
             autoCombos: rawAutoCombos,
-            features: resolved.features,
           });
         }
       }
@@ -5274,7 +5208,6 @@ export function createOmniRouteConfigHook(
           autoComboCount: rawAutoCombos.length,
           enrichment: rawEnrichment,
           autoCombos: rawAutoCombos,
-          features: resolved.features,
         });
       }
 

@omniroute/opencode-plugin/tests/feature-defaults.test.ts

@@ -1,95 +0,0 @@
-/**
- * #7624 — explicit feature-flag defaults.
- *
- * The `features` block in opencode.json marks every toggle `.optional()` with
- * no default. The effective value was previously only knowable by tracing the
- * implicit `features.X !== false` (default-ON) / `features.X === true`
- * (default-OFF) convention scattered across each read site, which left
- * operators unsure whether combos / autoCombos / enrichment were enabled when
- * they omitted the block.
- *
- * `OMNIROUTE_FEATURE_DEFAULTS` declares those defaults explicitly and
- * `resolveEffectiveFeatureFlags(features)` derives the effective state for any
- * (possibly-undefined) features object, mirroring the read-site conventions
- * exactly. These are purely derived — runtime routing behaviour is unchanged.
- */
-
-import test from "node:test";
-import assert from "node:assert/strict";
-
-import {
-  OMNIROUTE_FEATURE_DEFAULTS,
-  resolveEffectiveFeatureFlags,
-} from "../src/index.js";
-
-const DEFAULT_ON = [
-  "combos",
-  "autoCombos",
-  "enrichment",
-  "diskCache",
-  "providerTag",
-  "fetchInterceptor",
-  "geminiSanitization",
-] as const;
-
-const DEFAULT_OFF = [
-  "compressionMetadata",
-  "usableOnly",
-  "mcpAutoEmit",
-  "debugLog",
-  "startupDebug",
-] as const;
-
-test("OMNIROUTE_FEATURE_DEFAULTS: declares each flag with its documented default", () => {
-  for (const key of DEFAULT_ON) {
-    assert.equal(OMNIROUTE_FEATURE_DEFAULTS[key], true, `${key} defaults ON`);
-  }
-  for (const key of DEFAULT_OFF) {
-    assert.equal(OMNIROUTE_FEATURE_DEFAULTS[key], false, `${key} defaults OFF`);
-  }
-});
-
-test("resolveEffectiveFeatureFlags: undefined features → full declared default set", () => {
-  const flags = resolveEffectiveFeatureFlags(undefined);
-  for (const key of DEFAULT_ON) {
-    assert.equal(flags[key], true, `${key} effective ON when features omitted`);
-  }
-  for (const key of DEFAULT_OFF) {
-    assert.equal(flags[key], false, `${key} effective OFF when features omitted`);
-  }
-});
-
-test("resolveEffectiveFeatureFlags: empty features object → same as omitted", () => {
-  assert.deepEqual(
-    resolveEffectiveFeatureFlags({}),
-    resolveEffectiveFeatureFlags(undefined)
-  );
-});
-
-test("resolveEffectiveFeatureFlags: explicit false disables a default-ON flag", () => {
-  const flags = resolveEffectiveFeatureFlags({ autoCombos: false });
-  assert.equal(flags.autoCombos, false, "explicit autoCombos:false honoured");
-  // Untouched flags keep their declared defaults.
-  assert.equal(flags.combos, true);
-  assert.equal(flags.enrichment, true);
-});
-
-test("resolveEffectiveFeatureFlags: explicit true enables a default-OFF flag", () => {
-  const flags = resolveEffectiveFeatureFlags({ compressionMetadata: true });
-  assert.equal(flags.compressionMetadata, true, "explicit compressionMetadata:true honoured");
-  assert.equal(flags.usableOnly, false, "other opt-in flags stay OFF");
-});
-
-test("resolveEffectiveFeatureFlags: non-boolean sibling keys do not leak into flags", () => {
-  // features may also carry mcpToken/logLevel/apiFormat — the resolver must
-  // only ever return the boolean toggle keys.
-  const flags = resolveEffectiveFeatureFlags({
-    mcpAutoEmit: true,
-    mcpToken: "sk-mcp-token-abc",
-    logLevel: "debug",
-  });
-  assert.equal(flags.mcpAutoEmit, true);
-  assert.equal(Object.keys(flags).length, Object.keys(OMNIROUTE_FEATURE_DEFAULTS).length);
-  assert.equal("mcpToken" in flags, false);
-  assert.equal("logLevel" in flags, false);
-});

AGENTS.md

@@ -1,117 +1,600 @@
-# OmniRoute agent guide
+# omniroute — Agent Guidelines
 
 ## Project
 
-OmniRoute is a unified AI proxy/router. The repository contains the Next.js application
-(`src/`), streaming engine workspace (`open-sse/`), Electron desktop app (`electron/`),
-CLI (`bin/`), and tests (`tests/`).
+Unified AI proxy/router — route any LLM through one endpoint. Multi-provider support
+with **290 provider entries** (OpenAI, Anthropic, Gemini, DeepSeek, Groq, xAI, Mistral, Fireworks,
+Cohere, NVIDIA, Cerebras, Pollinations, Puter, Cloudflare AI, HuggingFace, DeepInfra,
+SambaNova, Meta Llama API, Moonshot AI, AI21 Labs, Databricks, Snowflake, and many more)
+with **MCP Server** (104 tools), **A2A v0.3 Protocol**, and **Electron desktop app**.
 
-## Setup and focused checks
+> **Live counts (v3.8.49)**: providers 290 · MCP tools 104 · MCP scopes 30 · A2A skills 6 ·
+> open-sse services 134 · routing strategies 17 · auto-combo scoring factors 12 ·
+> DB modules 95 · DB migrations 110 · base tables 17 · search providers 11 ·
+> i18n locales 42. **Refresh with `npm run check:docs-all`.**
 
-- Runtime: Node.js `>=22.22.3 <23` or `>=24.0.0 <27`; npm 10+.
-- Install dependencies: `npm install`.
-- Start development: `npm run dev`.
-- Build: `npm run build`; release build: `npm run build:release`.
-- Lint: `npm run lint`.
-- Core type check: `npm run typecheck:core`.
-- Run the most focused test for changed code first:
-  `node --import tsx/esm --test tests/unit/<file>.test.ts`.
-- Other suites: `npm run test:vitest`, `npm run test:e2e`,
-  `npm run test:protocols:e2e`, and `npm run test:ecosystem`.
-- Run `npm run check:docs-all` after changing documentation.
+## Doc Accuracy Discipline (read before writing any doc)
 
-For the complete test matrix, coverage requirements, and pull-request gates, read
-[`CONTRIBUTING.md`](CONTRIBUTING.md#running-tests).
+> **If `grep -rn "name" src/ open-sse/ bin/` returns nothing, the name does not exist. Do not document it.**
 
-## Documentation accuracy
+The recurring failure mode in AI-generated docs is _plausible-but-unverified specifics_.
+Every claim in a `.md` file under `docs/` should be verifiable against the source.
 
-Documentation must describe verified behavior, not plausible behavior.
+**Rules (enforced by `npm run check:fabricated-docs`):**
 
-1. Before documenting an API name, endpoint, path, CLI command, or environment variable,
-   search for it: `rg -n "name" src/ open-sse/ bin/`. If it has no source match, do not
-   document it.
-2. Measure mutable counts instead of writing them from memory: use `wc -l <file>` or a
-   directory-specific count command.
-3. Copy code examples from working usage or run them. Prefer a source link such as
-   `path/to/file.ts:line` to an invented signature.
-4. Run `npm run check:docs-all` for edits under `docs/`; it includes the fabricated-docs
-   validation.
+1. **Never state an API name, endpoint, path, CLI command, or env var without grepping for it first.**
+   ```bash
+   grep -rn "theName" src/ open-sse/ bin/
+   # 0 hits → do not document
+   ```
+2. **Never write a line count, file size, migration count, provider count, or strategy count from memory.**
+   ```bash
+   wc -l <file>           # exact line count
+   ls <dir>/*.ts | wc -l  # file count
+   ```
+3. **Every code example should be copy-pasted from real usage or actually run** — not synthesized.
+   Link to a real call site (`path:line`) instead of inventing a signature.
+4. **Prefer citing real source (`file.ts:line`) over paraphrasing behavior** — verifiable and self-correcting.
+5. **A shorter doc that is 100% accurate beats a comprehensive one with fabrications.**
+   Wrong docs cost more than missing docs, because people trust and act on them.
 
-## Code conventions
+The script `scripts/check/check-fabricated-docs.mjs` extracts every route path, env var, hook
+name, function name, and file reference from `docs/**/*.md` and verifies each one against the
+codebase. Run it locally before pushing docs; it runs in CI via `npm run check:docs-all`.
 
-- Format with Prettier: two spaces, semicolons, double quotes, 100-character line width,
-  and ES5 trailing commas. Run Prettier on changed files.
-- TypeScript target is ES2022 with bundler module resolution. Prefer explicit types.
-- Import order: external, internal (`@/` and `@omniroute/open-sse`), then relative.
-- Do not add logic to `src/lib/localDb.ts`; import from the owning `src/lib/db/` module.
-- Use specific errors and contextual logging. Do not silently swallow SSE-stream failures;
-  use abort signals for cleanup and return appropriate HTTP status codes.
+## Stack
 
-## Security requirements
+- **Runtime**: Next.js 16 (App Router), Node.js `>=22.0.0 <23 || >=24.0.0 <27`, ES Modules (`"type": "module"`)
+- **Language**: TypeScript 6.0 (`src/`) + JavaScript (`open-sse/`, `electron/`)
+- **Database**: better-sqlite3 (SQLite) — `DATA_DIR` configurable, default `~/.omniroute/`
+- **Streaming**: SSE via `open-sse` internal workspace package
+- **Styling**: Tailwind CSS v4
+- **i18n**: next-intl with 42 locales (`src/i18n/messages/`) — refresh with `ls src/i18n/messages/*.json | wc -l`
+- **Desktop**: Electron (cross-platform: Windows, macOS, Linux)
+- **Schemas**: Zod v4 for all API / MCP input validation
 
-- Never commit credentials or log SQLite encryption keys.
-- Validate API inputs with Zod and use the route's required authentication path.
-- Sanitize user HTML with DOMPurify.
-- Use `resolvePublicCred()` for public upstream OAuth identifiers; never add them as string
-  literals. See [`docs/security/PUBLIC_CREDS.md`](docs/security/PUBLIC_CREDS.md).
-- Use `buildErrorBody()` or `sanitizeErrorMessage()` for HTTP, SSE, executor, and MCP errors;
-  do not return raw `err.stack` or `err.message`. See
-  [`docs/security/ERROR_SANITIZATION.md`](docs/security/ERROR_SANITIZATION.md).
-- Pass runtime values to `exec()` or `spawn()` through `env`, not interpolation into a script.
+---
 
-## Repository map
+## Build, Lint, and Test Commands
 
-Read the nearest `AGENTS.md` and the linked deep-dive before making a non-trivial change.
+| Command                             | Description                                                        |
+| ----------------------------------- | ------------------------------------------------------------------ |
+| `npm run dev`                       | Start Next.js dev server                                           |
+| `npm run build`                     | Production build: `next build` → `.build/next/` + assemble `dist/` |
+| `npm run build:release`             | Clean rebuild + HEAD sentinel (`dist/BUILD_SHA`) — use for deploy  |
+| `npm run start`                     | Run production build                                               |
+| `npm run build:cli`                 | Build CLI package                                                  |
+| `npm run lint`                      | ESLint on all source files                                         |
+| `npm run typecheck:core`            | TypeScript core type checking                                      |
+| `npm run typecheck:noimplicit:core` | Strict checking (no implicit any)                                  |
+| `npm run check`                     | Run lint + test                                                    |
+| `npm run check:cycles`              | Check for circular dependencies                                    |
+| `npm run electron:dev`              | Run Electron app in dev mode                                       |
+| `npm run electron:build`            | Build Electron app for current OS                                  |
 
-| Area                               | Location                                                | Start here                                                                                                                                       |
-| ---------------------------------- | ------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------ |
-| API routes                         | `src/app/api/v1/`                                       | [`docs/architecture/ARCHITECTURE.md`](docs/architecture/ARCHITECTURE.md)                                                                         |
-| Streaming request handling         | `open-sse/handlers/`                                    | [`docs/architecture/ARCHITECTURE.md`](docs/architecture/ARCHITECTURE.md)                                                                         |
-| Provider execution and translation | `open-sse/executors/`, `open-sse/translator/`           | [`docs/architecture/CODEBASE_DOCUMENTATION.md`](docs/architecture/CODEBASE_DOCUMENTATION.md)                                                     |
-| Routing and resilience             | `open-sse/services/`                                    | [`open-sse/services/AGENTS.md`](open-sse/services/AGENTS.md), [`docs/routing/AUTO-COMBO.md`](docs/routing/AUTO-COMBO.md)                         |
-| Database and migrations            | `src/lib/db/`, `db/migrations/`                         | [`src/lib/db/AGENTS.md`](src/lib/db/AGENTS.md)                                                                                                   |
-| Domain policy                      | `src/domain/`                                           | [`docs/architecture/ARCHITECTURE.md`](docs/architecture/ARCHITECTURE.md)                                                                         |
-| MCP and A2A                        | `open-sse/mcp-server/`, `src/lib/a2a/`                  | [`docs/frameworks/MCP-SERVER.md`](docs/frameworks/MCP-SERVER.md), [`docs/frameworks/A2A-SERVER.md`](docs/frameworks/A2A-SERVER.md)               |
-| Agent features                     | `src/lib/{acp,memory,skills,cloudAgent}/`               | [`docs/frameworks/AGENT_PROTOCOLS_GUIDE.md`](docs/frameworks/AGENT_PROTOCOLS_GUIDE.md), [`docs/frameworks/SKILLS.md`](docs/frameworks/SKILLS.md) |
-| Safety and governance              | `src/lib/{guardrails,compliance}/`, `src/server/authz/` | [`docs/security/GUARDRAILS.md`](docs/security/GUARDRAILS.md), [`docs/architecture/AUTHZ_GUIDE.md`](docs/architecture/AUTHZ_GUIDE.md)             |
-| Operations                         | `src/mitm/`, tunnel modules, `electron/`                | [`docs/ops/TUNNELS_GUIDE.md`](docs/ops/TUNNELS_GUIDE.md), [`docs/guides/ELECTRON_GUIDE.md`](docs/guides/ELECTRON_GUIDE.md)                       |
+**Build output layout:**
 
-## Review focus
+| Directory | Purpose                                            | Gitignored |
+| --------- | -------------------------------------------------- | ---------- |
+| `src/`    | Application source (TypeScript / TSX)              | No         |
+| `.build/` | Build intermediates (`distDir = .build/next`)      | Yes        |
+| `dist/`   | Shippable bundle assembled by `assembleStandalone` | Yes        |
 
-- Keep database operations in `src/lib/db/`; do not issue raw SQL from routes.
-- Send provider requests through `open-sse/handlers/`.
-- Keep MCP and A2A pages as tabs inside `/dashboard/endpoint`.
-- Preserve SSE cleanup, rate-limit header parsing, Zod validation, and provider-schema
-  validation.
-- Treat Memory and Skills as cross-cutting changes that can affect MCP tools, the request
-  pipeline, and A2A skills.
-- Do not close a contributor pull request after using its code; merge it through GitHub so
-  the contributor receives credit.
+The pipeline is a single `next build` pass — intermediates land in `.build/next/`, the
+assembled bundle in `dist/`. VPS deploys rsync `dist/` into the remote
+`/usr/lib/node_modules/omniroute/app/` directory (VPS image path is unchanged).
 
-## Upstream contributions
+### Running Tests
 
-This checkout is a fork of `diegosouzapw/OmniRoute`. Keep fork-only deployment and personal
-automation changes out of upstream PRs.
+```bash
+# All tests (unit + vitest + ecosystem + e2e)
+npm run test:all
 
-Start upstream work from the active upstream default branch, not `main`:
+# Single test file (Node.js native test runner — most tests use this)
+node --import tsx/esm --test tests/unit/your-file.test.ts
+node --import tsx/esm --test tests/unit/plan3-p0.test.ts
+node --import tsx/esm --test tests/unit/fixes-p1.test.ts
+node --import tsx/esm --test tests/unit/security-fase01.test.ts
+
+# Integration tests
+node --import tsx/esm --test tests/integration/*.test.ts
+
+# Vitest (MCP server, autoCombo)
+npm run test:vitest
+
+# E2E with Playwright
+npm run test:e2e
+
+# Protocol clients E2E (MCP transports, A2A)
+npm run test:protocols:e2e
+
+# Ecosystem compatibility tests
+npm run test:ecosystem
+
+# Coverage (see CONTRIBUTING.md)
+npm run test:coverage
+```
+
+**For authoritative coverage requirements, test execution, and PR gates, see [`CONTRIBUTING.md`](CONTRIBUTING.md#running-tests).**
+
+---
+
+## Code Style Guidelines
+
+### Formatting (Prettier — enforced via lint-staged)
+
+2 spaces · semicolons required · double quotes (`"`) · 100 char width · es5 trailing commas.
+Always run `prettier --write` on changed files.
+
+### TypeScript
+
+- **Target**: ES2022 · **Module**: `esnext` · **Resolution**: `bundler`
+- `strict: false` — prefer explicit types, don't rely on inference
+- Path aliases: `@/*` → `src/`, `@omniroute/open-sse` → `open-sse/`, `@omniroute/open-sse/*` → `open-sse/*`
+
+### ESLint Rules
+
+- **Security (error, everywhere)**: `no-eval`, `no-implied-eval`, `no-new-func`
+- **Relaxed in `open-sse/` and `tests/`**: `@typescript-eslint/no-explicit-any` = warn
+- React hooks rules and `@next/next/no-assign-module-variable` disabled in `open-sse/` and `tests/`
+
+### Naming
+
+| Element             | Convention                       | Example                              |
+| ------------------- | -------------------------------- | ------------------------------------ |
+| Files               | camelCase / kebab-case           | `chatCore.ts`, `tokenHealthCheck.ts` |
+| React components    | PascalCase                       | `Dashboard.tsx`, `ProviderCard.tsx`  |
+| Functions/variables | camelCase                        | `getHealth()`, `switchCombo()`       |
+| Constants           | UPPER_SNAKE                      | `MAX_RETRIES`, `DEFAULT_TIMEOUT`     |
+| Interfaces          | PascalCase (`I` prefix optional) | `ProviderConfig`                     |
+| Enums               | PascalCase (members too)         | `LogLevel.Error`                     |
+
+### Imports
+
+- **Order**: external → internal (`@/`, `@omniroute/open-sse`) → relative (`./`, `../`)
+- **No barrel imports** from `localDb.ts` — import from the specific `db/` module instead
+
+### Error Handling
+
+- try/catch with specific error types; always log with context (pino logger)
+- Never silently swallow errors in SSE streams — use abort signals for cleanup
+- Return proper HTTP status codes (4xx client, 5xx server)
+
+### Security
+
+- **NEVER** commit API keys, secrets, or credentials
+- Validate all user inputs with Zod schemas
+- Auth middleware required on all API routes
+- Never log SQLite encryption keys
+- Sanitize user content (dompurify for HTML)
+- **Public upstream OAuth identifiers** (Gemini / Antigravity / Windsurf-style client_id/secret + Firebase Web keys extracted from public CLIs): use `resolvePublicCred()` from `open-sse/utils/publicCreds.ts`, **never** as string literals. Full pattern in `docs/security/PUBLIC_CREDS.md`.
+- **Error responses** (HTTP / SSE / executor / MCP): use `buildErrorBody()` or `sanitizeErrorMessage()` from `open-sse/utils/error.ts`, **never** put raw `err.stack` / `err.message` in a Response body. Full pattern in `docs/security/ERROR_SANITIZATION.md`.
+- **`exec()` / `spawn()` with runtime values**: pass via the `env` option, **never** string-interpolate paths/values into the script body. Reference: `src/mitm/cert/install.ts::updateNssDatabases`.
+- Prefer secure-by-default libraries when available — see [tldrsec/awesome-secure-defaults](https://github.com/tldrsec/awesome-secure-defaults) for the curated list (Helmet.js, DOMPurify, ssrf-req-filter, safe-regex, Google Tink, etc.).
+
+---
+
+## Architecture
+
+### Data Layer (`src/lib/db/`)
+
+All persistence uses SQLite through **95 domain-specific modules** in `src/lib/db/`. Top modules:
+
+- Core: `core.ts`, `migrationRunner.ts`, `encryption.ts`, `stateReset.ts`
+- Providers / catalog: `providers.ts`, `models.ts`, `providerLimits.ts`, `compressionAnalytics.ts`
+- Routing: `combos.ts`, `modelComboMappings.ts`, `domainState.ts`, `commandCodeAuth.ts`
+- Auth: `apiKeys.ts`, `secrets.ts`, `registeredKeys.ts`, `sessionAccountAffinity.ts`
+- Usage / billing: `quotaSnapshots.ts`, `creditBalance.ts`, `usage*.ts`, `compressionCacheStats.ts`
+- Storage: `backup.ts`, `cleanup.ts`, `jsonMigration.ts`, `healthCheck.ts`, `databaseSettings.ts`
+- Extension modules: `evals.ts`, `webhooks.ts`, `reasoningCache.ts`, `readCache.ts`, `tierConfig.ts`, `compressionCombos.ts`, `compressionScheduler.ts`, `batches.ts`, `files.ts`, `syncTokens.ts`, `proxies.ts`, `oneproxy.ts`, `upstreamProxy.ts`, `versionManager.ts`, `cliToolState.ts`, `prompts.ts`, `detailedLogs.ts`, `contextHandoffs.ts`, `compression.ts`, `stats.ts`
+
+Live count: `ls src/lib/db/*.ts | wc -l` (currently 95). Drift detection: `npm run check:docs-counts`.
+Schema migrations live in `db/migrations/` (**110 files** as of v3.8.43) and run via `migrationRunner.ts`.
+`src/lib/localDb.ts` is a **re-export layer only** — never add logic there.
+
+#### DB Internals
+
+- **`core.ts`**: `getDbInstance()` returns a singleton `better-sqlite3` instance with WAL
+  journaling. `SCHEMA_SQL` defines **17 base tables** (verify with `grep -c "CREATE TABLE" src/lib/db/core.ts` minus 1 for the bookkeeping `_omniroute_migrations` table). Helpers: `rowToCamel`, `encryptConnectionFields`.
+- **`migrationRunner.ts`**: Applies versioned SQL files from `db/migrations/` inside transactions.
+  Tracks applied migrations in `_omniroute_migrations` table.
+- **Migrations**: 110 files (`001_initial_schema.sql` → `110_*.sql`).
+  Each migration is idempotent and runs in a transaction. Live count: `ls src/lib/db/migrations/*.sql | wc -l`.
+- **Domain modules** import `getDbInstance()` from `core.ts` for all CRUD operations.
+  Each module owns a specific table/set of tables (e.g., `providers.ts` → `provider_connections`,
+  `combos.ts` → `combos`). Encryption helpers protect sensitive fields at rest.
+- **`localDb.ts`** re-exports all domain modules — consumers import from here for convenience.
+
+### API Route Layer (`src/app/api/v1/`)
+
+Next.js App Router routes — each follows a consistent pattern:
+
+```
+Route → CORS preflight → Body validation (Zod) → Optional auth (extractApiKey/isValidApiKey)
+  → API key policy enforcement (enforceApiKeyPolicy) → Handler delegation (open-sse)
+```
+
+| Route                           | Handler                   | Notes                                                         |
+| ------------------------------- | ------------------------- | ------------------------------------------------------------- |
+| `chat/completions/route.ts`     | `handleChat()`            | + prompt injection guard (clones request)                     |
+| `responses/route.ts`            | `handleChat()` (unified)  | Responses API format                                          |
+| `embeddings/route.ts`           | `handleEmbedding()`       | Model listing + creation                                      |
+| `images/generations/route.ts`   | `handleImageGeneration()` | Model listing + creation                                      |
+| `audio/transcriptions/route.ts` | audio handler             | Multipart form data                                           |
+| `audio/speech/route.ts`         | TTS handler               | Binary audio response                                         |
+| `videos/generations/route.ts`   | video handler             | ComfyUI/SD WebUI                                              |
+| `music/generations/route.ts`    | music handler             | ComfyUI workflows                                             |
+| `moderations/route.ts`          | moderation handler        | Content safety                                                |
+| `rerank/route.ts`               | rerank handler            | Document relevance                                            |
+| `search/route.ts`               | search handler            | Web search (12 providers per `open-sse/handlers/search.ts:6`) |
+
+**No global Next.js middleware file** — interception is route-specific. Auth is optional
+(controlled by `REQUIRE_API_KEY` env). Prompt injection guard is unique to chat completions.
+
+### Request Pipeline (`open-sse/`)
+
+The `open-sse/` workspace is the core streaming engine. Full request flow:
+
+```
+Client Request
+  → src/app/api/v1/.../route.ts (Next.js route)
+    → open-sse/handlers/chatCore.ts::handleChatCore()
+      → Semantic/signature cache check
+      → Rate limit check (rateLimitManager)
+      → Combo routing? → open-sse/services/combo.ts::handleComboChat()
+        → resolveComboTargets() → ordered ResolvedComboTarget[]
+        → For each target: handleSingleModel() (wraps chatCore)
+      → translateRequest() (open-sse/translator/)
+        → Convert source format (e.g., OpenAI) → target format (e.g., Claude)
+      → getExecutor() → provider-specific executor instance
+        → executor.execute() (BaseExecutor → DefaultExecutor or provider-specific)
+          → buildUrl() + buildHeaders() + transformRequest()
+          → fetch() to upstream provider
+          → Retry logic with exponential backoff
+      → Response translation back to client format
+      → If Responses API: responsesTransformer.ts TransformStream
+  → SSE stream or JSON response to client
+```
+
+**Handlers** (`open-sse/handlers/`): `chatCore.ts`, `responsesHandler.ts`, `embeddings.ts`,
+`imageGeneration.ts`, `videoGeneration.ts`, `musicGeneration.ts`, `audioSpeech.ts`,
+`audioTranscription.ts`, `moderations.ts`, `rerank.ts`, `search.ts`.
+
+**Upstream headers**: merged after default auth; same header name replaces executor value.
+**T5 intra-family fallback** recomputes headers using only the fallback model id.
+Forbidden header names: `src/shared/constants/upstreamHeaders.ts` — keep sanitize,
+Zod schemas, and unit tests aligned when editing.
+
+### Provider Categories
+
+- **Free** (2): Qoder AI, Kiro AI
+- **OAuth** (13): Claude Code, Antigravity, Codex, GitHub Copilot, Cursor, Kimi Coding, Kilo Code, Cline, Kiro, Qoder, Gemini, Windsurf (v3.8), GitLab Duo (v3.8)
+- **API Key** (120+): OpenAI, Anthropic, Gemini, DeepSeek, Groq, xAI, Mistral, Perplexity,
+  Together, Fireworks, Cerebras, Cohere, NVIDIA, Nebius, SiliconFlow, Hyperbolic,
+  HuggingFace, OpenRouter, Vertex AI, Cloudflare AI, Scaleway, AI/ML API, Pollinations,
+  Puter, Longcat, Alibaba, Kimi, Minimax, Blackbox, Synthetic, Kilo Gateway,
+  Z.AI, GLM, Deepgram, AssemblyAI, ElevenLabs, Cartesia, PlayHT, Inworld,
+  NanoBanana, SD WebUI, ComfyUI, Ollama Cloud, Perplexity Search, Serper, Brave, Exa,
+  Tavily, OpenCode Zen/Go, Bailian Coding Plan, DeepInfra, Vercel AI Gateway,
+  Lambda AI, SambaNova, nScale, OVHcloud AI, Baseten, PublicAI, Moonshot AI,
+  Meta Llama API, v0 (Vercel), Morph, Featherless AI, FriendliAI, LlamaGate,
+  Galadriel, Weights & Biases Inference, Volcengine, AI21 Labs, Venice.ai,
+  Codestral, Upstage, Maritalk, Xiaomi MiMo, Inference.net, NanoGPT, Predibase,
+  Bytez, Heroku AI, Databricks, Snowflake Cortex, GigaChat (Sber), CrofAI,
+  AgentRouter, ChatGPT Web, Baidu Qianfan, AWS Polly, RunwayML, GitLab Duo,
+  Amazon Q, Empower, Poe, and many more.
+- **Self-Hosted** (8+): LM Studio, vLLM, Lemonade, Llamafile, Triton, Docker Model Runner, Xinference, Oobabooga
+- **Custom**: OpenAI-compatible (`openai-compatible-*`) and Anthropic-compatible (`anthropic-compatible-*`) prefixes
+
+Providers are registered in `src/shared/constants/providers.ts` with Zod validation at module load.
+
+### Executors (`open-sse/executors/`)
+
+Provider-specific request executors: `base.ts`, `default.ts`, `cursor.ts`, `codex.ts`,
+`antigravity.ts`, `github.ts`, `kiro.ts`, `qoder.ts`, `vertex.ts`,
+`cloudflare-ai.ts`, `opencode.ts`, `pollinations.ts`, `puter.ts`.
+
+#### Executor Internals
+
+- **`base.ts`** (`BaseExecutor`): Abstract base with `buildUrl()`, `buildHeaders()`,
+  `transformRequest()`, retry logic (exponential backoff), and `execute()`. Subclasses
+  override URL/header/transform methods for provider-specific behavior.
+- **`default.ts`** (`DefaultExecutor extends BaseExecutor`): Handles most OpenAI-compatible
+  providers. Reads provider config from `providerRegistry.ts` to resolve base URL, auth
+  header format, and request transformations.
+- **`getExecutor()`** (`executors/index.ts`): Factory that returns the correct executor
+  instance based on provider ID. Provider-specific executors (Cursor, Codex, Vertex, etc.)
+  override only what differs from the default.
+
+### Translator (`open-sse/translator/`)
+
+Translates between API formats (OpenAI-format ↔ Anthropic, Gemini, etc.).
+Includes request/response translators with helpers for image handling.
+
+#### Translator Internals
+
+- **`translator/index.ts`**: Exports `translateRequest()` and format constants. Called by
+  `chatCore.ts` before executor dispatch.
+- **Flow**: `translateRequest(body, sourceFormat, targetFormat)` → detects source format
+  (OpenAI, Anthropic, Gemini) → applies the matching translator module → returns
+  transformed body ready for the target provider.
+- **Response translation** runs in reverse after upstream response, converting back to
+  the client's expected format.
+
+### Transformer (`open-sse/transformer/`)
+
+`responsesTransformer.ts` — transforms Responses API format to/from Chat Completions format.
+
+#### Transformer Internals
+
+- **`createResponsesApiTransformStream()`**: Returns a `TransformStream` that converts
+  Chat Completions SSE chunks (`data: {"choices":[...]}`) into Responses API SSE events
+  (`response.output_item.added`, `response.output_text.delta`, etc.).
+- Used when the client sends a Responses API request: the request is internally converted
+  to Chat Completions format, dispatched normally, and the response is piped through this
+  transform stream before reaching the client.
+
+### Services (`open-sse/services/`)
+
+134 service modules in `open-sse/services/` (top-level only; more including sub-dirs like `autoCombo/` and `compression/`). Refresh: `ls open-sse/services/*.ts | wc -l`. Key modules:
+`combo.ts` (routing engine), `usage.ts`, `tokenRefresh.ts`,
+`rateLimitManager.ts`, `accountFallback.ts`, `sessionManager.ts`, `wildcardRouter.ts`,
+`autoCombo/`, `intentClassifier.ts`, `taskAwareRouter.ts`, `thinkingBudget.ts`,
+`contextManager.ts`, `modelDeprecation.ts`, `modelFamilyFallback.ts`,
+`emergencyFallback.ts`, `workflowFSM.ts`, `backgroundTaskDetector.ts`, `ipFilter.ts`,
+`signatureCache.ts`, `volumeDetector.ts`, `contextHandoff.ts`, `compression/` (prompt
+compression pipeline), and more.
+
+#### Prompt Compression Pipeline (`compression/`)
+
+Modular prompt compression that runs proactively before the existing reactive context manager.
+
+- **`strategySelector.ts`**: Selects compression mode based on config, compression combo assignments,
+  combo overrides, auto-trigger thresholds, and defaults. Priority: assigned compression combo >
+  combo override > auto-trigger > default mode > off.
+- **`lite.ts`**: 5 lite-mode techniques: `collapseWhitespace`, `dedupSystemPrompt`,
+  `compressToolResults`, `removeRedundantContent`, `replaceImageUrls`. Target: 10-15% savings at
+  <1ms latency.
+- **`caveman.ts` / `cavemanRules.ts`**: Caveman-style semantic condensation backed by built-in
+  rules plus file-loaded language packs under `compression/rules/`.
+- **`engines/rtk/`**: Rule-based terminal/tool-output compression inspired by RTK patterns. Detects
+  command output classes, applies JSON filter packs, deduplicates repeated lines, strips ANSI/code
+  noise, and preserves errors/actionable context. The RTK JSON DSL supports replace,
+  match-output short-circuit, strip/keep, per-line truncation, head/tail/max-line truncation,
+  inline tests, trust-gated project/global custom filters, and optional redacted raw-output
+  retention for authenticated recovery.
+- **`engines/registry.ts`**: Registers engines (`caveman`, `rtk`) and powers stacked pipelines.
+- **`stats.ts`**: Per-request compression stats tracking (original tokens, compressed tokens,
+  savings %, techniques used, engine breakdown, compression combo id).
+- **`types.ts`**: `CompressionMode` (off/lite/standard/aggressive/ultra/rtk/stacked),
+  `CompressionConfig`, `CompressionStats`, `CompressionResult`.
+- DB settings in `src/lib/db/compression.ts`, compression combos in
+  `src/lib/db/compressionCombos.ts`, API routes under `src/app/api/settings/compression/`,
+  `src/app/api/context/*`, and preview/language-pack routes under `src/app/api/compression/*`.
+
+#### Combo Routing Engine (`combo.ts`)
+
+- **`handleComboChat()`**: Entry point for combo-routed requests. Receives the combo config
+  and iterates through targets in order until one succeeds or all fail.
+- **`resolveComboTargets()`**: Expands a combo configuration into an ordered array of
+  `ResolvedComboTarget[]`, each specifying provider + model + account + credentials.
+- **Strategies** (17): priority, weighted, fill-first, round-robin, P2C, random, least-used, reset-aware (v3.8),
+  reset-window, cost-optimized, strict-random, auto, lkgp, context-optimized, context-relay, headroom, fusion. Source: `ROUTING_STRATEGY_VALUES` in `src/shared/constants/routingStrategies.ts`.
+- Each target calls **`handleSingleModel()`** which wraps `handleChatCore()` with
+  per-target error handling and circuit breaker checks.
+
+### Domain Layer (`src/domain/`)
+
+Policy engine modules: `policyEngine.ts`, `comboResolver.ts`, `costRules.ts`,
+`degradation.ts`, `fallbackPolicy.ts`, `lockoutPolicy.ts`, `modelAvailability.ts`,
+`providerExpiration.ts`, `quotaCache.ts`, `responses.ts`, `configAudit.ts`.
+
+### MCP Server (`open-sse/mcp-server/`)
+
+**104 tools** total (`TOTAL_MCP_TOOL_COUNT`, `open-sse/mcp-server/server.ts`): a 42-entry base registry (`MCP_TOOLS` in `schemas/tools.ts`, bundling the core / cache / compression / 1proxy / advanced tools) **plus** standalone module sets — memory (3), skill (4), agentSkill (3), pool (6), gamification (8), plugin (8), notion (6), obsidian (22). 3 transports (stdio / SSE / Streamable HTTP). Scoped auth (31 scopes — see `OMNIROUTE_MCP_SCOPES`), Zod schemas. See [`docs/frameworks/MCP-SERVER.md`](docs/frameworks/MCP-SERVER.md).
+
+**Core tools** (20): get_health, list_combos, get_combo_metrics, switch_combo, check_quota,
+route_request, cost_report, list_models_catalog, web_search, simulate_route, set_budget_guard,
+set_routing_strategy, set_resilience_profile, test_combo, get_provider_metrics,
+best_combo_for_task, explain_route, get_session_snapshot, db_health_check, sync_pricing.
+
+**Cache tools** (2): cache_stats, cache_flush.
+
+**Compression tools** (5): compression_status, compression_configure, set_compression_engine,
+list_compression_combos, compression_combo_stats.
+
+**1proxy tools** (3): oneproxy_fetch, oneproxy_rotate, oneproxy_stats.
+
+**Memory tools** (3): memory_search, memory_add, memory_clear.
+
+**Skill tools** (4): skills_list, skills_enable, skills_execute, skills_executions.
+
+**Agent-skill tools** (3): A2A skill discovery / invocation bridges.
+
+**Gamification tools** (8): levels, badges, leaderboard, and community-federation queries.
+
+**Plugin tools** (8): plugin marketplace listing, install/enable/disable, and runtime inspection.
+
+**Notion tools** (6) + **Obsidian tools** (22): knowledge-base read/write integrations (the largest tool family — vault search, note CRUD, WebDAV-backed file ops).
+
+#### MCP Internals
+
+- **Tool registration**: Each tool is an object with `{ name, description, inputSchema: ZodSchema,
+handler: async (args) => {...} }`. Zod validates inputs before the handler fires.
+- **`createMcpServer()`** and **`startMcpStdio()`** exported from `mcp-server/index.ts`.
+  `createMcpServer()` wires all tool sets; `startMcpStdio()` launches the stdio transport.
+- **Transports**: stdio (CLI `omniroute --mcp`), SSE (`/api/mcp/sse`), Streamable HTTP
+  (`/api/mcp/stream`). All share the same tool/scope engine.
+- **Scopes** (30): Control which tool categories an API key can access. Enforcement happens
+  before handler dispatch.
+- **Audit**: Every tool invocation is logged to SQLite (`mcp_audit` table) with tool name,
+  args, success/failure, API key attribution, and timestamp.
+
+### A2A Server (`src/lib/a2a/`)
+
+JSON-RPC 2.0, SSE streaming, Task Manager with TTL cleanup.
+Agent Card at `/.well-known/agent.json`.
+Skills (6): `smartRouting.ts`, `quotaManagement.ts`, `providerDiscovery.ts`, `costAnalysis.ts`, `healthReport.ts`, `listCapabilities.ts`.
+
+#### A2A Internals
+
+- **`taskManager.ts`**: State machine lifecycle for tasks: `submitted → working →
+completed | failed | canceled`. Tasks have TTL and are cleaned up automatically.
+- **JSON-RPC methods**: `message/send` (sync), `message/stream` (SSE), `tasks/get`,
+  `tasks/cancel`. Dispatched via `POST /a2a`.
+- **Skills**: Registered in a DB-backed registry. Each skill receives task context
+  (messages, metadata) and returns structured results. `quotaManagement.ts` summarizes
+  quota; `smartRouting.ts` recommends routing decisions.
+- **Agent Card**: `/.well-known/agent.json` exposes capabilities, skills, and metadata
+  for client auto-discovery.
+
+### ACP Module (`src/lib/acp/`)
+
+Agent Communication Protocol registry and manager.
+
+### Memory System (`src/lib/memory/`)
+
+Extraction, injection, retrieval, summarization, and store modules for persistent
+conversational memory across sessions.
+
+### Skills System (`src/lib/skills/`)
+
+Extensible skill framework: registry, executor, sandbox, built-in skills,
+custom skill support, interception, and injection.
+
+#### Skills Internals
+
+- **`registry.ts`**: DB-backed skill registration and discovery. Skills have metadata
+  (name, description, version, enabled status) stored in SQLite.
+- **`executor.ts`**: Execution engine with configurable timeout and retry logic.
+  Receives skill name + input, looks up the skill, runs it in the sandbox.
+- **`sandbox.ts`**: Isolation layer for custom (user-provided) skills. Limits resource
+  access and execution time.
+- **Built-in skills**: Ship with OmniRoute (e.g., quota management, routing). Located
+  alongside the registry.
+- **Interception/Injection**: Skills can intercept requests in the pipeline (pre/post
+  processing) or inject context into prompts.
+
+### Compliance (`src/lib/compliance/`)
+
+Policy index for compliance enforcement.
+
+### MITM Proxy (`src/mitm/`)
+
+MITM proxy capability with certificate management, DNS handling, and target routing.
+
+### Middleware (`src/middleware/`)
+
+Request middleware including `promptInjectionGuard.ts`.
+
+### Guardrails (`src/lib/guardrails/`)
+
+Hot-reloadable guardrails framework (3 built-in: pii-masker, prompt-injection, vision-bridge). Fail-open. The `pii-masker` guardrail is registered and runs on every request, but its data-mutating logic is **opt-in** and OFF by default — it only redacts when `PII_REDACTION_ENABLED` (request) / `PII_RESPONSE_SANITIZATION` (response + streaming) are enabled (both `defaultValue: "false"`); with them off, payloads pass through untouched. A request can additionally opt OUT of any guardrail via header (`x-omniroute-disabled-guardrails`). Never make PII default-on (Hard Rule #20). See [`docs/security/GUARDRAILS.md`](docs/security/GUARDRAILS.md).
+
+### Cloud Agents (`src/lib/cloudAgent/`)
+
+`CloudAgentBase` abstract class + 3 agents (codex-cloud, devin, jules). Tasks persisted in `cloud_agent_tasks`; management auth required. See [`docs/frameworks/CLOUD_AGENT.md`](docs/frameworks/CLOUD_AGENT.md).
+
+### Evals (`src/lib/evals/`)
+
+Generic eval framework: `evalRunner.ts`, `runtime.ts`. Targets: combo / model / suite-default. See [`docs/frameworks/EVALS.md`](docs/frameworks/EVALS.md).
+
+### Webhooks (`src/lib/webhookDispatcher.ts`)
+
+HMAC-signed delivery, exponential backoff, auto-disable after 10 failures. 7 event types. See [`docs/frameworks/WEBHOOKS.md`](docs/frameworks/WEBHOOKS.md).
+
+### Authorization Pipeline (`src/server/authz/`)
+
+`classify → policies → enforce`. 3 route classes (PUBLIC / CLIENT_API / MANAGEMENT). See [`docs/architecture/AUTHZ_GUIDE.md`](docs/architecture/AUTHZ_GUIDE.md).
+
+### Reasoning Replay (`src/lib/db/reasoningCache.ts` + `open-sse/services/reasoningCache.ts`)
+
+Hybrid in-memory + SQLite cache for `reasoning_content`. Re-injects on multi-turn for strict providers (DeepSeek V4, Kimi K2, Qwen-Thinking, GLM, xiaomi-mimo). See [`docs/routing/REASONING_REPLAY.md`](docs/routing/REASONING_REPLAY.md).
+
+### Tunnels (`src/lib/{cloudflaredTunnel,ngrokTunnel}.ts` + `src/app/api/tunnels/`)
+
+Cloudflare Quick/Named, ngrok, Tailscale Funnel. See [`docs/ops/TUNNELS_GUIDE.md`](docs/ops/TUNNELS_GUIDE.md).
+
+### Adding a New Provider
+
+1. Register in `src/shared/constants/providers.ts`
+2. Add executor in `open-sse/executors/` (if custom logic needed)
+3. Add translator in `open-sse/translator/` (if non-OpenAI format)
+4. Add OAuth config in `src/lib/oauth/constants/oauth.ts` (if OAuth-based)
+5. Add models in `open-sse/config/providerRegistry.ts`
+
+---
+
+## Subdirectory AGENTS.md Files
+
+- **[`src/lib/db/AGENTS.md`](src/lib/db/AGENTS.md)** — SQLite persistence, domain modules, migrations
+- **[`open-sse/services/AGENTS.md`](open-sse/services/AGENTS.md)** — Routing engine, combo resolution, strategy selection
+
+## Reference Documentation (docs/)
+
+For any non-trivial change, read the matching deep-dive first:
+
+| Area                                       | Doc                                                                                                             |
+| ------------------------------------------ | --------------------------------------------------------------------------------------------------------------- |
+| Repo navigation                            | [`docs/architecture/REPOSITORY_MAP.md`](docs/architecture/REPOSITORY_MAP.md)                                    |
+| Architecture                               | [`docs/architecture/ARCHITECTURE.md`](docs/architecture/ARCHITECTURE.md)                                        |
+| Engineering reference                      | [`docs/architecture/CODEBASE_DOCUMENTATION.md`](docs/architecture/CODEBASE_DOCUMENTATION.md)                    |
+| Auto-Combo (12-factor, 18 strategies)      | [`docs/routing/AUTO-COMBO.md`](docs/routing/AUTO-COMBO.md)                                                      |
+| Resilience (3 layers)                      | [`docs/architecture/RESILIENCE_GUIDE.md`](docs/architecture/RESILIENCE_GUIDE.md)                                |
+| Skills                                     | [`docs/frameworks/SKILLS.md`](docs/frameworks/SKILLS.md)                                                        |
+| Memory                                     | [`docs/frameworks/MEMORY.md`](docs/frameworks/MEMORY.md)                                                        |
+| Cloud agents                               | [`docs/frameworks/CLOUD_AGENT.md`](docs/frameworks/CLOUD_AGENT.md)                                              |
+| Guardrails                                 | [`docs/security/GUARDRAILS.md`](docs/security/GUARDRAILS.md)                                                    |
+| Evals                                      | [`docs/frameworks/EVALS.md`](docs/frameworks/EVALS.md)                                                          |
+| Compliance                                 | [`docs/security/COMPLIANCE.md`](docs/security/COMPLIANCE.md)                                                    |
+| Webhooks                                   | [`docs/frameworks/WEBHOOKS.md`](docs/frameworks/WEBHOOKS.md)                                                    |
+| Authz                                      | [`docs/architecture/AUTHZ_GUIDE.md`](docs/architecture/AUTHZ_GUIDE.md)                                          |
+| Stealth                                    | [`docs/security/STEALTH_GUIDE.md`](docs/security/STEALTH_GUIDE.md)                                              |
+| Reasoning replay                           | [`docs/routing/REASONING_REPLAY.md`](docs/routing/REASONING_REPLAY.md)                                          |
+| Agent protocols (A2A / ACP / Cloud)        | [`docs/frameworks/AGENT_PROTOCOLS_GUIDE.md`](docs/frameworks/AGENT_PROTOCOLS_GUIDE.md)                          |
+| MCP server                                 | [`docs/frameworks/MCP-SERVER.md`](docs/frameworks/MCP-SERVER.md)                                                |
+| A2A server                                 | [`docs/frameworks/A2A-SERVER.md`](docs/frameworks/A2A-SERVER.md)                                                |
+| API reference                              | [`docs/reference/API_REFERENCE.md`](docs/reference/API_REFERENCE.md) + [`docs/openapi.yaml`](docs/openapi.yaml) |
+| Provider catalog (auto-generated)          | [`docs/reference/PROVIDER_REFERENCE.md`](docs/reference/PROVIDER_REFERENCE.md)                                  |
+| Tunnels                                    | [`docs/ops/TUNNELS_GUIDE.md`](docs/ops/TUNNELS_GUIDE.md)                                                        |
+| Electron desktop                           | [`docs/guides/ELECTRON_GUIDE.md`](docs/guides/ELECTRON_GUIDE.md)                                                |
+| Release flow                               | [`docs/ops/RELEASE_CHECKLIST.md`](docs/ops/RELEASE_CHECKLIST.md)                                                |
+| Quality gates (35 gates, allowlist policy) | [`docs/architecture/QUALITY_GATES.md`](docs/architecture/QUALITY_GATES.md)                                      |
+| Cluster opt-in profiles (memory, bifrost)  | [`docs/architecture/cluster-decisions.md`](docs/architecture/cluster-decisions.md)                              |
+
+---
+
+## Fork / Upstream Workflow
+
+This repository is a fork of `diegosouzapw/OmniRoute`. Keep fork-only operational
+changes (for example GHCR image publishing, personal deployment workflows, or local
+automation) out of upstream contribution PRs.
+
+When preparing a PR for upstream, always start the work branch from the upstream
+**default branch** — the active `release/vX.Y.Z` line (today `release/v3.8.49`).
+Never branch from `main`: `main` only receives release squash-merges, so a branch
+cut there is weeks behind and produces conflict-heavy PRs
+(see `CONTRIBUTING.md` and `docs/ops/BRANCHING_MODEL.md`):
 
 ```bash
 git fetch upstream
-git switch -c <branch-name> upstream/<default-branch>
+# the default branch is the active release line, e.g. release/v3.8.49
+git switch -c <branch-name> upstream/release/vX.Y.Z
 ```
 
-Target that same release branch in the pull request. Stage only the intended files, run the
-focused checks, and use a Conventional Commit message (for example, `docs: slim AGENTS.md`).
+Only cherry-pick or reapply the changes intended for the upstream PR.
 
-## Reference documentation
+---
 
-Use the source of truth for the area you are changing:
+## Review Focus
 
-| Area                                   | Reference                                                                                                                                                                                                                                              |
-| -------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
-| Repository navigation and architecture | [`docs/architecture/REPOSITORY_MAP.md`](docs/architecture/REPOSITORY_MAP.md), [`docs/architecture/ARCHITECTURE.md`](docs/architecture/ARCHITECTURE.md)                                                                                                 |
-| API and providers                      | [`docs/reference/API_REFERENCE.md`](docs/reference/API_REFERENCE.md), [`docs/reference/PROVIDER_REFERENCE.md`](docs/reference/PROVIDER_REFERENCE.md), [`docs/openapi.yaml`](docs/openapi.yaml)                                                         |
-| Routing, resilience, and reasoning     | [`docs/routing/AUTO-COMBO.md`](docs/routing/AUTO-COMBO.md), [`docs/architecture/RESILIENCE_GUIDE.md`](docs/architecture/RESILIENCE_GUIDE.md), [`docs/routing/REASONING_REPLAY.md`](docs/routing/REASONING_REPLAY.md)                                   |
-| Security                               | [`docs/security/GUARDRAILS.md`](docs/security/GUARDRAILS.md), [`docs/security/COMPLIANCE.md`](docs/security/COMPLIANCE.md), [`docs/security/STEALTH_GUIDE.md`](docs/security/STEALTH_GUIDE.md)                                                         |
-| Platform features                      | [`docs/frameworks/MCP-SERVER.md`](docs/frameworks/MCP-SERVER.md), [`docs/frameworks/A2A-SERVER.md`](docs/frameworks/A2A-SERVER.md), [`docs/frameworks/SKILLS.md`](docs/frameworks/SKILLS.md), [`docs/frameworks/MEMORY.md`](docs/frameworks/MEMORY.md) |
-| Releases and quality                   | [`docs/ops/RELEASE_CHECKLIST.md`](docs/ops/RELEASE_CHECKLIST.md), [`docs/architecture/QUALITY_GATES.md`](docs/architecture/QUALITY_GATES.md)                                                                                                           |
+- **DB ops** go through `src/lib/db/` modules, never raw SQL in routes
+- **Provider requests** flow through `open-sse/handlers/`
+- **MCP/A2A pages** are tabs inside `/dashboard/endpoint`, not standalone routes
+- **No memory leaks** in SSE streams (abort signals, cleanup)
+- **Rate limit headers** must be parsed correctly
+- All API inputs validated with **Zod schemas**
+- **Provider constants** validated at module load via Zod (`src/shared/validation/providerSchema.ts`)
+- **Pricing data** syncs from LiteLLM via `src/lib/pricingSync.ts`
+- **Memory/Skills** are cross-cutting: affect MCP tools, request pipeline, and A2A skills
+- **⛔ NEVER close a contributor's PR** after using their code — always merge via GitHub so they get credit. See `.agents/workflows/review-prs.md` for full policy.

CLAUDE.md

@@ -45,7 +45,7 @@ For full test matrix, see `CONTRIBUTING.md` → "Running Tests". For deep archit
 | Translators   | `open-sse/translator/`  | Format conversion (OpenAI↔Claude↔Gemini)                                                                                                                |
 | Transformer   | `open-sse/transformer/` | Responses API ↔ Chat Completions                                                                                                                        |
 | Services      | `open-sse/services/`    | Combo routing, rate limits, caching, etc                                                                                                                |
-| Database      | `src/lib/db/`           | SQLite domain modules (130 migrations)                                                                                                                  |
+| Database      | `src/lib/db/`           | SQLite domain modules (95 files, 110 migrations)                                                                                                        |
 | Domain/Policy | `src/domain/`           | Policy engine, cost rules, fallback logic                                                                                                               |
 | MCP Server    | `open-sse/mcp-server/`  | 104 tools (42 base + memory/skill/agentSkill/pool/notion/obsidian/gamification/plugin modules), 3 transports (stdio / SSE / Streamable HTTP), 31 scopes |
 | A2A Server    | `src/lib/a2a/`          | JSON-RPC 2.0 agent protocol                                                                                                                             |
@@ -72,7 +72,7 @@ Client → /v1/chat/completions (Next.js route)
 
 API routes follow a consistent pattern: `Route → CORS preflight → Zod body validation → Optional auth (extractApiKey/isValidApiKey) → API key policy enforcement → Handler delegation (open-sse)`. No global Next.js middleware — interception is route-specific.
 
-**Combo routing** (`open-sse/services/combo.ts`): 19 public strategies (priority, weighted, fill-first, round-robin, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, cache-optimized, context-relay, fusion, pipeline). Each target calls `handleSingleModel()` which wraps `handleChatCore()` with per-target error handling and circuit breaker checks. The `fusion` strategy is the exception: it fans out to a panel of models in parallel, then a judge model synthesizes one final answer (`open-sse/services/fusion.ts`). See `docs/routing/AUTO-COMBO.md` for the 13-factor Auto-Combo scoring + the full strategy table and `docs/architecture/RESILIENCE_GUIDE.md` for the 3 resilience layers.
+**Combo routing** (`open-sse/services/combo.ts`): 18 strategies (priority, weighted, fill-first, round-robin, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline). Each target calls `handleSingleModel()` which wraps `handleChatCore()` with per-target error handling and circuit breaker checks. The `fusion` strategy is the exception: it fans out to a panel of models in parallel, then a judge model synthesizes one final answer (`open-sse/services/fusion.ts`). See `docs/routing/AUTO-COMBO.md` for the 12-factor Auto-Combo scoring + the full strategy table and `docs/architecture/RESILIENCE_GUIDE.md` for the 3 resilience layers.
 
 ---
 
@@ -332,7 +332,7 @@ For any non-trivial change, read the matching deep-dive first:
 | Repo navigation                               | `docs/architecture/REPOSITORY_MAP.md`                   |
 | Architecture                                  | `docs/architecture/ARCHITECTURE.md`                     |
 | Engineering reference                         | `docs/architecture/CODEBASE_DOCUMENTATION.md`           |
-| Auto-Combo (13-factor scoring, 19 strategies) | `docs/routing/AUTO-COMBO.md`                            |
+| Auto-Combo (12-factor scoring, 18 strategies) | `docs/routing/AUTO-COMBO.md`                            |
 | Resilience (3 mechanisms)                     | `docs/architecture/RESILIENCE_GUIDE.md`                 |
 | Reasoning replay                              | `docs/routing/REASONING_REPLAY.md`                      |
 | Skills framework                              | `docs/frameworks/SKILLS.md`                             |
@@ -461,18 +461,10 @@ own dedicated branch, and you MUST confirm the base branch with the operator bef
    git fetch origin "$BASE_BRANCH"
    git worktree add ".claude/worktrees/${TASK##*/}" -b "$TASK" "origin/$BASE_BRANCH"
    cd ".claude/worktrees/${TASK##*/}"
-   # Reuse the main checkout's node_modules to skip a per-worktree npm install.
-   # HARD LINKS (`cp -al`), never a symlink: ~5s for the whole tree and near-zero extra
-   # disk (the inodes are shared), and unlike a symlink it does not break the dev server.
-   cp -al "$(git -C <main_checkout> rev-parse --show-toplevel)/node_modules" node_modules
+   # symlink node_modules from the main checkout to skip a per-worktree npm install:
+   ln -s "$(git -C <main_checkout> rev-parse --show-toplevel)/node_modules" node_modules
    ```
 
-   **Never `ln -s` node_modules.** Turbopack rejects a symlink that resolves outside the
-   project root, so `npm run dev` dies with a FATAL panic (`Symlink [project]/node_modules
-   is invalid, it points out of the filesystem root`) while typecheck, lint and the test
-   runners all keep passing — the error names "filesystem root", not the worktree, so it
-   reads like a Next/build bug and costs real time to trace (incident 2026-07-31, #9043).
-
    In Claude Code prefer the native `EnterWorktree` tool (it already creates worktrees under
    `.claude/worktrees/`): create the worktree with the command above, then call `EnterWorktree`
    with its `path`.

CONTRIBUTING.md

@@ -2,11 +2,6 @@
 
 Thank you for your interest in contributing! This guide covers everything you need to get started.
 
-For the official per-change workflow, start with the
-[Contribution Golden Path](docs/dev/CONTRIBUTION_GOLDEN_PATH.md). It maps provider, routing,
-UI/UX, i18n, CLI, database, and build/deploy changes to their contracts, focused tests, CI
-coverage, and reconciliation steps.
-
 ---
 
 ## Development Setup
@@ -203,11 +198,10 @@ Coverage notes:
 
 ### Pull Request Requirements
 
-Before opening a PR, use the
-[Contribution Golden Path](docs/dev/CONTRIBUTION_GOLDEN_PATH.md) to run the focused loop for
-what you changed. The full unit suite (4 CI shards), Vitest, the **60%+** coverage gate, and
-the production build are CI's responsibility — running them locally adds no signal the PR
-checks will not already give you, and on smaller machines it can saturate the host (#8084):
+Before opening a PR, run the focused loop for what you changed. The full unit suite
+(4 CI shards), Vitest, the **60%+** coverage gate, and the production build are CI's
+responsibility — running them locally adds no signal the PR checks will not already
+give you, and on smaller machines it can saturate the host (#8084):
 
 - Run the test files that cover your change: `node --import tsx/esm --test tests/unit/<file>.test.ts`
 - Run `npm run lint`
@@ -277,7 +271,7 @@ src/                        # TypeScript (.ts / .tsx)
 │   ├── a2a/                # Agent-to-Agent v0.3 protocol server
 │   ├── acp/                # Agent Communication Protocol registry
 │   ├── compliance/         # Compliance policy engine
-│   ├── db/                 # SQLite domain modules + 130 migrations
+│   ├── db/                 # SQLite database layer (21 modules + 16 migrations)
 │   ├── memory/             # Persistent conversational memory
 │   ├── oauth/              # OAuth providers, services, and utilities
 │   ├── skills/             # Extensible skill framework
@@ -287,7 +281,7 @@ src/                        # TypeScript (.ts / .tsx)
 ├── mitm/                   # MITM proxy (cert, DNS, target routing)
 ├── shared/
 │   ├── components/         # React components (.tsx)
-│   ├── constants/          # Provider definitions (290), MCP scopes, 19 routing strategies
+│   ├── constants/          # Provider definitions (177), MCP scopes, 14 routing strategies
 │   ├── utils/              # Circuit breaker, sanitizer, auth helpers
 │   └── validation/         # Zod v4 schemas
 └── sse/                    # SSE proxy pipeline
@@ -295,7 +289,7 @@ src/                        # TypeScript (.ts / .tsx)
 open-sse/                   # @omniroute/open-sse workspace
 ├── executors/              # 14 provider-specific request executors
 ├── handlers/               # 11 request handlers (chat, responses, embeddings, images, etc.)
-├── mcp-server/             # MCP server (104 tools, 3 transports, 31 scopes)
+├── mcp-server/             # MCP server (25 tools, 3 transports, 10 scopes)
 ├── services/               # 36+ services (combo, autoCombo, rateLimitManager, etc.)
 ├── translator/             # Format translators (OpenAI ↔ Claude ↔ Gemini ↔ Responses ↔ Ollama)
 ├── transformer/            # Responses API transformer

Dockerfile

@@ -1,5 +1,5 @@
 # ── Common base with runtime deps ──────────────────────────────────────────
-FROM node:26-trixie-slim AS base
+FROM node:24-trixie-slim AS base
 WORKDIR /app
 
 # `apt-get upgrade` pulls the security-patched versions of the Debian (trixie)
@@ -8,8 +8,8 @@ WORKDIR /app
 # that already have a fix published in trixie. CVEs without an upstream fix yet
 # (local-only TOCTOU, etc.) remain until the distro patches them and the image
 # is rebuilt; none are reachable from the proxy's request surface at runtime.
-RUN --mount=type=cache,id=apt-cache,target=/var/cache/apt,sharing=locked \
-  --mount=type=cache,id=apt-lists,target=/var/lib/apt/lists,sharing=locked \
+RUN --mount=type=cache,id=apt-cache,target=/var/cache/apt,sharing=shared \
+  --mount=type=cache,id=apt-lists,target=/var/lib/apt/lists,sharing=shared \
   apt-get update \
   && apt-get upgrade -y \
   && apt-get install -y --no-install-recommends libsecret-1-0 ca-certificates \
@@ -29,8 +29,8 @@ FROM base AS builder
 
 # Build tools for native module compilation
 # apt-get update needed here because base's rm -rf clears the shared cache
-RUN --mount=type=cache,id=apt-cache,target=/var/cache/apt,sharing=locked \
-  --mount=type=cache,id=apt-lists,target=/var/lib/apt/lists,sharing=locked \
+RUN --mount=type=cache,id=apt-cache,target=/var/cache/apt,sharing=shared \
+  --mount=type=cache,id=apt-lists,target=/var/lib/apt/lists,sharing=shared \
   apt-get update \
   && apt-get install -y --no-install-recommends python3 make g++ \
   && rm -rf /var/lib/apt/lists/*
@@ -95,11 +95,6 @@ RUN --mount=type=cache,id=npm-cache,target=/root/.npm \
 # See docs/ops/QUALITY_GATE_PLAYBOOK.md Parte 6.
 ENV OMNIROUTE_USE_TURBOPACK=1
 
-# Next.js basePath is fixed at build time; pass OMNIROUTE_BASE_PATH here when the
-# image should serve under a reverse-proxy subpath without a runtime patch.
-ARG OMNIROUTE_BASE_PATH=""
-ENV OMNIROUTE_BASE_PATH=$OMNIROUTE_BASE_PATH
-
 # Docker containers cannot run the MITM/Agent-Bridge stack (no host DNS/cert
 # access), so keep @/mitm/manager on the graceful stub (#3390). This flag is
 # Docker-only: npm/Electron/VPS builds must bundle the REAL manager (#6344).

README.md

@@ -67,46 +67,32 @@
 
 <table>
   <tr>
-    <td align="right"><b>🚀 Start</b></td>
-    <td align="center"><a href="#-quick-start">🚀 Quick Start</a></td>
-    <td align="center"><a href="#-more-install-methods--docker-source-pnpm-arch">📦 Install</a></td>
-    <td align="center"><a href="#-works-the-second-you-install-it--no-keys-no-config">🆓 Zero-config</a></td>
+    <td align="center"><a href="#-quick-start"><b>🚀 Quick Start</b></a></td>
+    <td align="center"><a href="#-combos--the-flagship"><b>🎯 Combos</b></a></td>
+    <td align="center"><a href="#-290-ai-providers--90-free"><b>🌐 Providers</b></a></td>
   </tr>
   <tr>
-    <td align="right"><b>💡 Learn</b></td>
+    <td align="center"><a href="#-full-cli--a2a--mcp"><b>🔌 CLI &amp; MCP</b></a></td>
+    <td align="center"><a href="#%EF%B8%8F-save-1595-tokens--automatically"><b>🗜️ Compression</b></a></td>
+    <td align="center"><a href="https://omniroute.online"><b>🌍 Website</b></a></td>
+  </tr>
+</table>
+
+<table>
+  <tr>
     <td align="center"><a href="#-the-promise">💥 The Promise</a></td>
-    <td align="center"><a href="#-why-omniroute">🤔 Why OmniRoute</a></td>
+    <td align="center"><a href="#-why-omniroute">🤔 Why</a></td>
     <td align="center"><a href="#-what-sets-omniroute-apart">🏆 What Sets Apart</a></td>
   </tr>
   <tr>
-    <td align="right"><b>⚙️ Features</b></td>
-    <td align="center"><a href="#-combos--the-flagship">🎯 Combos</a></td>
-    <td align="center"><a href="#-290-ai-providers--90-free">🌐 Providers</a></td>
-    <td align="center"><a href="#-full-cli--a2a--mcp">🔌 CLI &amp; MCP</a></td>
-  </tr>
-  <tr>
-    <td align="right"></td>
-    <td align="center"><a href="#%EF%B8%8F-save-1595-tokens--automatically">🗜️ Compression</a></td>
+    <td align="center"><a href="#-compatible-clis--coding-agents">🤖 Compatible CLIs</a></td>
     <td align="center"><a href="#%EF%B8%8F-where-omniroute-runs--anywhere">🖥️ Where It Runs</a></td>
     <td align="center"><a href="#-private--local-first">🔒 Private</a></td>
   </tr>
   <tr>
-    <td align="right"><b>👀 See it</b></td>
     <td align="center"><a href="#-omniroute-in-action">🎬 In Action</a></td>
-    <td align="center"><a href="#-whats-new">✨ What's New</a></td>
-    <td align="center"><a href="#-compatible-clis--coding-agents">🤖 Compatible CLIs</a></td>
-  </tr>
-  <tr>
-    <td align="right"><b>💚 Support</b></td>
-    <td align="center"><a href="#-support-omniroute">💚 Support / Donate</a></td>
-    <td align="center"><a href="#-community--help">💬 Community</a></td>
-    <td align="center"><a href="#-sponsors">💖 Sponsors</a></td>
-  </tr>
-  <tr>
-    <td align="right"><b>📦 Project</b></td>
-    <td align="center"><a href="#%EF%B8%8F-tech-stack">🛠️ Tech Stack</a></td>
-    <td align="center"><a href="#-documentation">📖 Docs</a></td>
-    <td align="center"><a href="#-500-contributors">👥 Contributors</a></td>
+    <td align="center"><a href="#-dashboard-screenshots">📸 Screenshots</a></td>
+    <td align="center"><a href="#-support--community">📧 Support</a></td>
   </tr>
 </table>
 
@@ -241,53 +227,12 @@ curl http://localhost:20128/v1/chat/completions \
       <b>What Kimi's support powers:</b> Kimi's API credits power OmniRoute's AI-validated release pipeline — the <i>merge validation powered by Kimi K3</i> stage that reviews every pull request before it ships — plus day-to-day feature development. First-class Kimi support ships on both rails: the direct <a href="https://platform.kimi.ai?aff=omniroute">Kimi API</a> (<code>kimi-k3</code>) and the <a href="https://www.kimi.com/code?aff=omniroute">Kimi Code coding plan</a> (OAuth and API key). OmniRoute is also the first Brazilian open-source project in Kimi's support program. <a href="https://platform.kimi.ai?aff=omniroute"><b>Get a Kimi API key →</b></a>
     </td>
   </tr>
-  <tr>
-    <td align="center" width="150">
-      <a href="https://cheaperinference.com/?utm_source=omniroute">
-        <img src="public/providers/cheaperinference.svg" width="64" alt="Cheaper Inference"/>
-      </a>
-      <br/><b>Cheaper Inference</b><br/><sub>cheaperinference.com</sub><br/><br/>
-      <img src="https://img.shields.io/badge/Open_Source_Friend-31f889?style=flat-square&labelColor=04170d" alt="Open Source Friend"/>
-    </td>
-    <td>
-      Thanks to <b>Cheaper Inference</b>, an OmniRoute Open Source Friend, for backing this project! Cheaper Inference is a cost-ranked gateway that resells 42 frontier models — Claude, GPT-5.x, Gemini, Kimi K3, GLM, DeepSeek, Grok and MiniMax — behind one OpenAI-compatible endpoint, routing each request to the cheapest eligible provider without ever charging above the model maker's list price.
-      <br/><br/>
-      <b>First-class support in OmniRoute:</b> Chat Completions, the native <code>/v1/responses</code> endpoint, vision, tool calling and 3 image models (<code>grok-imagine</code>, <code>nano-banana-pro</code>, <code>nano-banana-2</code>, reachable as <code>cheaperinference/&lt;model&gt;</code>). <a href="https://cheaperinference.com/?utm_source=omniroute"><b>Get an API key →</b></a>
-    </td>
-  </tr>
 </table>
 
 <sub>Links tagged <code>aff=omniroute</code> are partner links. They fund the project at no extra cost to you.</sub>
 
 <br/>
 
-<details open>
-<summary><sub><b>🎟️ Affiliates Promo</b> — free signup coupons from providers we don't sponsor (click to expand)</sub></summary>
-
-<sub><i>This section is for referral/coupon codes only. Sponsored partnerships live in <b>🤝 Supported by our Open Source Friends</b> above. OmniRoute has no sponsorship or partnership with the providers listed here — these are public coupons anyone can use.</i></sub>
-
-<table>
-  <tr>
-    <td align="center" width="120">
-      <a href="https://agentrouter.org/register?aff=70LM">
-        <img src="public/providers/agentrouter.png" width="32" alt="AgentRouter"/>
-      </a>
-      <br/><sub><b>AgentRouter</b></sub><br/><sub>agentrouter.org</sub>
-    </td>
-    <td>
-      <sub><b><a href="https://agentrouter.org/register?aff=70LM">AgentRouter</a></b> — affiliate signup · <b>$100 free credits</b> on signup (free server, expect higher latency — best for testing, not production). First-class support in OmniRoute since <b>v3.8.50</b>: Chat Completions, the Anthropic-compatible wire format and the OpenAI-compatible path. Available models include <code>claude-opus-4-8</code>, <code>claude-opus-5</code>, <code>gpt-5.6-sol</code> and more. <b><a href="https://agentrouter.org/register?aff=70LM">Grab your $100 →</a></b></sub>
-      <br/><br/>
-      <sub>⚠️ <i>Affiliate link — OmniRoute has no sponsorship or partnership with this provider.</i></sub>
-    </td>
-  </tr>
-</table>
-
-<sub>Know another provider with a generous free signup coupon that benefits OmniRoute users? Open an issue and we'll add it here.</sub>
-
-</details>
-
-<br/>
-
 <div align="center">
 
 ## 🎯 Combos — The Flagship
@@ -445,50 +390,13 @@ All **19** strategies — mix & match per combo step:
 
 <br/>
 
-## 💚 Support OmniRoute
+## ❤️ Support
 
-OmniRoute is MIT-licensed and maintained in the open. If it saves you time or money, here's how to keep it independent — pick whatever fits you. Sponsorship never affects routing priority; it buys visibility, not ranking.
+OmniRoute is free and open source, built and maintained in the open. If it saves you time or money, consider supporting development:
 
-<table>
-  <tr><td nowrap>⭐ <b>Star the repo</b></td><td>Free — genuinely helps visibility</td><td><a href="https://github.com/diegosouzapw/OmniRoute">Star OmniRoute</a></td></tr>
-  <tr><td nowrap>🐙 <b>GitHub Sponsors</b></td><td>One-off or monthly · zero platform fee</td><td><a href="https://github.com/sponsors/diegosouzapw">github.com/sponsors/diegosouzapw</a></td></tr>
-  <tr><td nowrap>🏢 <b>Open Collective</b></td><td><b>Companies</b> — issues an invoice/receipt · transparent books</td><td><a href="https://opencollective.com/omniroute">opencollective.com/omniroute</a></td></tr>
-  <tr><td nowrap>☕ <b>Ko-fi</b></td><td>Quick one-off tip, no signup for the donor</td><td><a href="https://ko-fi.com/diegosouzapw">ko-fi.com/diegosouzapw</a></td></tr>
-  <tr><td nowrap>🧋 <b>Buy Me a Coffee</b></td><td>Small, informal gesture</td><td><a href="https://www.buymeacoffee.com/diegosouzapw">buymeacoffee.com/diegosouzapw</a></td></tr>
-  <tr><td nowrap>🖐 <b>Liberapay</b></td><td>Recurring · non-profit · open source</td><td><a href="https://liberapay.com/diegosouzapw">liberapay.com/diegosouzapw</a></td></tr>
-  <tr><td nowrap>🇧🇷 <b>PIX</b> (Brazil)</td><td>Instant, no fees</td><td>key &amp; QR below</td></tr>
-  <tr><td nowrap>₿ <b>Crypto</b></td><td>BTC · ETH · USDT-TRC20 · USDC-Solana</td><td>addresses below</td></tr>
-</table>
-
-**🇧🇷 PIX** — instant, no fees (Brazil)
-
-<img src="docs/assets/pix-qr.png" width="140" align="right" alt="OmniRoute PIX QR code"/>
-
-Key (random): `5d865059-bc44-483a-962d-43ceb80126eb`
-
-Pix copia-e-cola:
-
-```
-00020101021126580014br.gov.bcb.pix01365d865059-bc44-483a-962d-43ceb80126eb5204000053039865802BR5922OMNIROUTE CONTRIBUICAO6006BRASIL62070503***630475DD
-```
-
-<br clear="right"/>
-
-<details>
-<summary><b>₿ Crypto</b> — BTC · ETH · USDT-TRC20 · USDC-Solana (click to expand)</summary>
-
-<table>
-  <tr><td nowrap><b>₿ BTC</b></td><td nowrap>Bitcoin (SegWit)</td><td><code>bc1qh00smz004sy85wyl28v77tenkt3ckl6eaep7fd</code></td></tr>
-  <tr><td nowrap><b>Ξ ETH</b></td><td nowrap>Ethereum (ERC20)</td><td><code>0x64Cf6B68A6Ff34288e89172950a2d00102337a84</code></td></tr>
-  <tr><td nowrap><b>₮ USDT</b></td><td nowrap>Tron (TRC20)</td><td><code>TKAF41JpuQrHbKTnsQa9svJE2T192Hvsc2</code></td></tr>
-  <tr><td nowrap><b>$ USDC</b></td><td nowrap>Solana</td><td><code>2emNNZzVVWQc3FQ2wk9M6qXUQmW8AKdjjL174fXR28Tu</code></td></tr>
-</table>
-
-<sub>⚠️ Send each coin only on the network shown — sending on the wrong network can lose the funds.</sub>
-
-</details>
-
-🐛 Found a bug or have feedback? Open a [Discussion](https://github.com/diegosouzapw/OmniRoute/discussions).
+- ⭐ **Star the repo** — it genuinely helps visibility
+- 💖 **[GitHub Sponsors](https://github.com/sponsors/diegosouzapw)** — fund ongoing maintenance and new providers
+- 🐛 **Report bugs and share feedback** in [Discussions](https://github.com/diegosouzapw/OmniRoute/discussions)
 
 <br/>
 
@@ -510,12 +418,11 @@ Pix copia-e-cola:
 - **🕵️ Transparent MITM decrypt (TPROXY)** — capture CLIs that ignore proxy env vars, with a per-SNI CA + trust-store installer. → [MITM/TPROXY](docs/security/MITM-TPROXY-DECRYPT.md)
 - **💸 Cost telemetry everywhere** — `X-OmniRoute-*` cost/usage headers on every endpoint, cache-HIT savings header, per-key USD spend quotas. → [API Reference](docs/reference/API_REFERENCE.md)
 - **🧠 Memory you control** — off by default, opt-in int8 vector quantization + typed decay, per-request `x-omniroute-no-memory`. → [Memory](docs/frameworks/MEMORY.md)
-- **🛡️ Security** — prompt-injection guard on every LLM route (red-team suite), opt-in credential-masking guardrail (redacts leaked API keys/secrets in both directions), free DuckDuckGo last-resort web search, and an optional OIDC login gate for the dashboard (password login always stays available). → [Guardrails](docs/security/GUARDRAILS.md)
+- **🛡️ Security** — prompt-injection guard on every LLM route (red-team suite) + free DuckDuckGo last-resort web search. → [Guardrails](docs/security/GUARDRAILS.md)
 - **🖼️ New endpoints** — `/v1/ocr` (Mistral OCR) and `/v1/audio/translations` (Whisper-style) round out the media surface. → [API Reference](docs/reference/API_REFERENCE.md)
 - **🎨 Image / video / audio generation** — one API for media: xAI Grok Imagine & Novita AI video, ComfyUI, Freepik, Adobe Firefly, Microsoft Designer, Google Imagen, Segmind, EdgeTTS. → [API Reference](docs/reference/API_REFERENCE.md)
 - **🌍 Deployment & ops** — reverse-proxy `basePath`, browser-language auto-detect, per-key device tracking, root-less MITM trust, zh-TW localization. → [Environment](docs/reference/ENVIRONMENT.md)
-- **🤝 More providers & agents** — Cursor Cloud Agent, Grok Build (xAI) with browser + OAuth login, Ollama first-class card, Claude Opus 5 & Sonnet 5, Kimi official partnership (Code/Web/Moonshot), Zed, Requesty, SenseNova, Yuanbao, Agnes AI… and a refreshed **290-provider catalog**. → [Providers](docs/reference/PROVIDER_REFERENCE.md)
-- **📡 Routing transparency** — every response carries an `X-OmniRoute-Decision` header naming the strategy/provider/latency that served it, a new `cache-optimized` combo strategy + Auto-Combo `cacheAffinity` factor route repeat requests back to the connection holding the cached prefix, and a read-only `/v1/auto-combo/{channel}/candidates` endpoint exposes an `auto/*` channel's live candidate pool. → [Auto-Combo](docs/routing/AUTO-COMBO.md)
+- **🤝 More providers & agents** — Cursor Cloud Agent, Grok Build (xAI), Ollama first-class card, Claude Sonnet 5, Zed, Requesty, SenseNova, Yuanbao, Agnes AI… and a refreshed **290-provider catalog**. → [Providers](docs/reference/PROVIDER_REFERENCE.md)
 - **⚡ Local performance & infra** — one-click local Redis, Cloudflare Workers / Deno Deploy relay deployers, Bifrost & Mux as supervised embedded services. → [Embedded Services](docs/frameworks/EMBEDDED-SERVICES.md)
 
 <br/>
@@ -926,22 +833,18 @@ devbox run npm run dev
 **🦭 Podman**
 
 ```bash
-# 1. Prepare the bind-mounted data directory
-mkdir -p data
+# 1. Build the image
+podman build --target runner-base -t omniroute:base .
 
-# 2. Linux + local rootless Podman only (never a remote Podman Machine client):
-podman unshare chown 1000:1000 ./data
+# 2. Fix data directory permissions for rootless Podman
+mkdir -p data && podman unshare chown 1000:1000 ./data
 
-# 3. Set the runtime hint, build the local Compose image, and start
+# 3. Set runtime in .env, then run (see contrib/podman/ for Quadlet)
 echo "CONTAINER_HOST=podman" >> .env
-podman compose --profile base up -d --build
+podman compose --profile base up -d
 ```
 
-On macOS or Windows, Podman uses a remote Podman Machine: skip `podman unshare` and
-follow the [topology-specific data directory guidance](contrib/podman/README.md#data-directory-permissions-by-topology).
-
-📖 [Podman Guide](contrib/podman/README.md) — Compose builds, Podman Machine, and
-Linux/systemd Quadlet setup.
+📖 [Podman Guide](contrib/podman/README.md) — Quadlet setup, podman-compose, Quadlet.
 
 **⚡ Faster / leaner install (skip the native build)**
 
@@ -997,7 +900,7 @@ same process on one port, so there is no separate CLI-only package today.
 
 <div align="center">
 
-# 📧 Community & Help
+# 📧 Support & Community
 
 > Everything in one place — follow the maintainer, chat with the community, or open an issue.
 
@@ -1013,7 +916,7 @@ same process on one port, so there is no separate CLI-only package today.
 | 📦 **Source code**                         | [github.com/diegosouzapw/OmniRoute](https://github.com/diegosouzapw/OmniRoute)                                            |
 | 🐛 **Report a bug**                        | [open an issue](https://github.com/diegosouzapw/OmniRoute/issues) — attach `npm run system-info` output                   |
 | 🤝 **Contribute**                          | [CONTRIBUTING.md](CONTRIBUTING.md) · [Branching & Release Model](docs/ops/BRANCHING_MODEL.md) · pick a `good first issue` |
-| 💚 **Support the project**                 | [Ways to support ↑](#-support-omniroute) · [GitHub Sponsors](https://github.com/sponsors/diegosouzapw)                    |
+| ⭐ **Support the project**                 | [Star the repo](https://github.com/diegosouzapw/OmniRoute) · [GitHub Sponsors](https://github.com/sponsors/diegosouzapw)  |
 
 </div>
 

bin/cli/commands/launch-codex.mjs

@@ -1,7 +1,6 @@
 import { spawn } from "node:child_process";
 import { t } from "../i18n.mjs";
 import { resolveActiveContext } from "../contexts.mjs";
-import { quoteShellArgs } from "../utils/winShellArgs.mjs";
 
 /** OpenAI/Codex env keys stripped from the child so a stale OpenAI key/base-url
  *  in the shell can't shadow the omniroute provider (defense-in-depth). Mirrors
@@ -30,23 +29,6 @@ export function resolveCodexSpawn(platform) {
   return { command: "codex", shell: undefined };
 }
 
-/**
- * `shell: true` makes Node join argv with plain spaces and no escaping (the
- * DEP0190 warning). That mangles every launch-codex invocation on Windows, not
- * just the ones with a multi-word user argument: the injected `-c` provider
- * flags carry TOML values whose quotes cmd.exe strips
- * (`model_providers.omniroute.name="OmniRoute"` arrives unquoted and no longer
- * parses as TOML). Quote the args ourselves on that path; off Windows there is
- * no shell, so argv is passed through untouched. Same fix as `launch` (#8837).
- *
- * @param {string[]} args
- * @param {NodeJS.Platform|string} platform
- * @returns {string[]}
- */
-export function quoteCodexArgs(args, platform) {
-  return quoteShellArgs(args, platform);
-}
-
 function stripTrailingSlash(value) {
   let s = String(value);
   let end = s.length;
@@ -171,7 +153,7 @@ export async function runLaunchCodexCommand(opts = {}, codexArgs = []) {
 
   return await new Promise((resolve) => {
     const { command: codexLaunch, shell: shellValue } = resolveCodexSpawn(process.platform);
-    const child = spawn(codexLaunch, quoteCodexArgs(extraArgs, process.platform), {
+    const child = spawn(codexLaunch, extraArgs, {
       env,
       stdio: "inherit",
       shell: shellValue,
@@ -213,10 +195,7 @@ export function registerLaunchCodex(program) {
     .argument("[codexArgs...]", "arguments passed through to the codex binary")
     .action(async (codexArgs, opts) => {
       const merged = { ...opts, profile: opts.profile ?? opts.p };
-      // process.exit() here aborted the process with a libuv assertion on
-      // Windows (`!(handle->flags & UV_HANDLE_CLOSING)`, async.c:94): it tears
-      // the loop down while the inherited stdio handles of the just-exited
-      // child are still closing. Setting exitCode lets the loop drain first.
-      process.exitCode = await runLaunchCodexCommand(merged, codexArgs ?? []);
+      const exitCode = await runLaunchCodexCommand(merged, codexArgs ?? []);
+      if (exitCode !== 0) process.exit(exitCode);
     });
 }

bin/cli/commands/launch.mjs

@@ -3,7 +3,6 @@ import { join } from "node:path";
 import os from "node:os";
 import { t } from "../i18n.mjs";
 import { resolveActiveContext } from "../contexts.mjs";
-import { quoteShellArgs } from "../utils/winShellArgs.mjs";
 
 function stripTrailingSlash(value) {
   let s = String(value);
@@ -91,34 +90,6 @@ export function resolveLaunchTarget(opts = {}) {
   return { baseUrl, authToken };
 }
 
-/**
- * #8246: on Windows, npm installs claude as a `.cmd` shim — spawn() without a
- * shell cannot resolve PATHEXT shims (and Node refuses to exec `.cmd` directly
- * since CVE-2024-27980), so the Windows path must go through cmd.exe.
- *
- * @param {NodeJS.Platform|string} platform
- * @returns {{ command: string, shell: true|undefined }}
- */
-export function resolveClaudeSpawn(platform) {
-  return platform === "win32"
-    ? { command: "claude.cmd", shell: true }
-    : { command: "claude", shell: undefined };
-}
-
-/**
- * `shell: true` makes Node join argv with plain spaces and no escaping (the
- * DEP0190 warning), so `-p "two words"` used to reach claude as `-p two` plus
- * three stray positional arguments. Quote the args ourselves on that path.
- * Off Windows there is no shell, so argv is passed through untouched.
- *
- * @param {string[]} args
- * @param {NodeJS.Platform|string} platform
- * @returns {string[]}
- */
-export function quoteClaudeArgs(args, platform) {
-  return quoteShellArgs(args, platform);
-}
-
 /**
  * @param {{port?:string, remote?:string, token?:string, apiKey?:string, profile?:string, claudeHome?:string}} opts
  * @param {string[]} claudeArgs  pass-through args for the claude binary
@@ -149,12 +120,13 @@ export async function runLaunchCommand(opts = {}, claudeArgs = []) {
   const env = buildClaudeEnv(process.env, baseUrl, authToken, { configDir });
 
   return await new Promise((resolve) => {
-    const { command, shell } = resolveClaudeSpawn(process.platform);
-    const child = spawn(command, quoteClaudeArgs(claudeArgs, process.platform), {
+    // #8246: on Windows, npm installs claude as a .cmd shim — spawn() without
+    // shell:true cannot resolve PATHEXT shims and fails with ENOENT.
+    const claudeCommand = process.platform === "win32" ? "claude.cmd" : "claude";
+    const child = spawn(claudeCommand, claudeArgs, {
       env,
       stdio: "inherit",
-      shell,
-      ...(process.platform === "win32" ? { windowsHide: true } : {}),
+      ...(process.platform === "win32" ? { shell: true, windowsHide: true } : {}),
     });
     child.on("error", (err) => {
       if (err && err.code === "ENOENT") {
@@ -187,10 +159,7 @@ export function registerLaunch(program) {
     .allowExcessArguments(true)
     .argument("[claudeArgs...]", "arguments passed through to the claude binary")
     .action(async (claudeArgs, opts) => {
-      // process.exit() here aborted the process with a libuv assertion on
-      // Windows (`!(handle->flags & UV_HANDLE_CLOSING)`, async.c:94): it tears
-      // the loop down while the inherited stdio handles of the just-exited
-      // child are still closing. Setting exitCode lets the loop drain first.
-      process.exitCode = await runLaunchCommand(opts, claudeArgs ?? []);
+      const exitCode = await runLaunchCommand(opts, claudeArgs ?? []);
+      if (exitCode !== 0) process.exit(exitCode);
     });
 }

bin/cli/commands/serve.mjs

@@ -7,11 +7,6 @@ import { t } from "../i18n.mjs";
 import { writePidFile, cleanupPidFile, waitForServer } from "../utils/pid.mjs";
 import { ServerSupervisor, detectMitmCrash } from "../runtime/processSupervisor.mjs";
 import { isTermux } from "../../../scripts/build/postinstallSupport.mjs";
-import {
-  ensureAndroidCacheDir,
-  isFatalInstrumentationHookFailure,
-  formatAndroidInstrumentationFailureHint,
-} from "../utils/ensureAndroidCacheDir.mjs";
 import {
   resolveMaxOldSpaceMb,
   calibrateHeapFallbackMb,
@@ -67,37 +62,9 @@ export function registerServe(program) {
     });
 }
 
-/** Once-per-process guard so the Android/Termux cache hint is not spammed. */
-let instrumentationFailureHintPrinted = false;
-
-/**
- * If child output looks like Next.js failed to load its instrumentation hook
- * on Android/Termux, print a clear operator-facing fix hint.
- * Exported for unit tests.
- *
- * @param {string} text
- * @returns {boolean} true when a hint was printed
- */
-export function maybeReportInstrumentationHookFailure(text) {
-  if (instrumentationFailureHintPrinted) return false;
-  if (!isFatalInstrumentationHookFailure(text)) return false;
-  instrumentationFailureHintPrinted = true;
-  process.stderr.write(formatAndroidInstrumentationFailureHint(process.env.XDG_CACHE_HOME));
-  return true;
-}
-
-/** Test-only reset for the once-per-process hint guard. */
-export function resetInstrumentationFailureHintForTests() {
-  instrumentationFailureHintPrinted = false;
-}
-
 export async function runServe(opts = {}) {
   const startedAt = performance.now();
 
-  // Same prep as bin/omniroute.mjs — keep it here so a direct `runServe()` call
-  // (tests / programmatic) still gets a writable Next.js cache dir before spawn.
-  ensureAndroidCacheDir({ env: process.env });
-
   const { isNativeBinaryCompatible } =
     await import("../../../scripts/build/native-binary-compat.mjs");
   const { getNodeRuntimeSupport, getNodeRuntimeWarning } =
@@ -167,11 +134,7 @@ export async function runServe(opts = {}) {
     "Release",
     "better_sqlite3.node"
   );
-  if (
-    !process.versions.bun &&
-    existsSync(sqliteBinary) &&
-    !isNativeBinaryCompatible(sqliteBinary)
-  ) {
+  if (!process.versions.bun && existsSync(sqliteBinary) && !isNativeBinaryCompatible(sqliteBinary)) {
     console.error(
       "\x1b[31m✖ better-sqlite3 native module is incompatible with this platform.\x1b[0m"
     );
@@ -267,16 +230,15 @@ export async function runServe(opts = {}) {
 function runDaemon(serverJs, env, memoryLimit, dashboardPort, apiPort) {
   // #5238: skip the explicit CLI --max-old-space-size when the user pinned the
   // heap via NODE_OPTIONS (a CLI arg would shadow/override their value).
-  const server = spawn(
-    process.versions.bun ? process.execPath : "node",
-    [...(process.versions.bun ? [] : buildNodeHeapArgs(process.env, memoryLimit)), serverJs],
-    {
-      cwd: APP_DIR,
-      env,
-      stdio: "ignore",
-      detached: true,
-    }
-  );
+  const server = spawn(process.versions.bun ? process.execPath : "node", [
+    ...(process.versions.bun ? [] : buildNodeHeapArgs(process.env, memoryLimit)),
+    serverJs,
+  ], {
+    cwd: APP_DIR,
+    env,
+    stdio: "ignore",
+    detached: true,
+  });
   writePidFile("server", server.pid);
   server.unref();
   console.log(`\x1b[32m✔ OmniRoute started in background (PID: ${server.pid})\x1b[0m`);
@@ -287,15 +249,14 @@ function runDaemon(serverJs, env, memoryLimit, dashboardPort, apiPort) {
 function runWithoutRecovery(serverJs, env, memoryLimit, dashboardPort, apiPort, noOpen, startedAt) {
   // #5238: skip the explicit CLI --max-old-space-size when the user pinned the
   // heap via NODE_OPTIONS (a CLI arg would shadow/override their value).
-  const server = spawn(
-    process.versions.bun ? process.execPath : "node",
-    [...(process.versions.bun ? [] : buildNodeHeapArgs(process.env, memoryLimit)), serverJs],
-    {
-      cwd: APP_DIR,
-      env,
-      stdio: "pipe",
-    }
-  );
+  const server = spawn(process.versions.bun ? process.execPath : "node", [
+    ...(process.versions.bun ? [] : buildNodeHeapArgs(process.env, memoryLimit)),
+    serverJs,
+  ], {
+    cwd: APP_DIR,
+    env,
+    stdio: "pipe",
+  });
 
   writePidFile("server", server.pid);
 
@@ -304,7 +265,6 @@ function runWithoutRecovery(serverJs, env, memoryLimit, dashboardPort, apiPort,
   server.stdout.on("data", (data) => {
     const text = data.toString();
     process.stdout.write(text);
-    maybeReportInstrumentationHookFailure(text);
     if (
       !started &&
       (text.includes("Ready") || text.includes("started") || text.includes("listening"))
@@ -314,11 +274,7 @@ function runWithoutRecovery(serverJs, env, memoryLimit, dashboardPort, apiPort,
     }
   });
 
-  server.stderr.on("data", (data) => {
-    const text = data.toString();
-    process.stderr.write(text);
-    maybeReportInstrumentationHookFailure(text);
-  });
+  server.stderr.on("data", (data) => process.stderr.write(data));
 
   server.on("error", (err) => {
     console.error("\x1b[31m✖ Failed to start server:\x1b[0m", err.message);
@@ -427,9 +383,6 @@ export function reportReadinessTimeout(dashboardPort, supervisor) {
     console.error("--- Recent server output ---");
     recentLog.forEach((l) => console.error(l));
     console.error("--- End recent output ---\n");
-    // If the buffered log already shows the Android instrumentation failure,
-    // print the actionable hint even when --log was off (default).
-    maybeReportInstrumentationHookFailure(recentLog.join("\n"));
   }
 }
 

bin/cli/runtime/processSupervisor.mjs

@@ -10,21 +10,11 @@ import {
 } from "./supervisorPolicy.mjs";
 import { buildNodeHeapArgs } from "../../../scripts/build/runtime-env.mjs";
 import { stopProcessGracefully } from "../../../src/shared/platform/windowsProcess.ts";
-import {
-  isFatalInstrumentationHookFailure,
-  formatAndroidInstrumentationFailureHint,
-} from "../utils/ensureAndroidCacheDir.mjs";
 
 const CRASH_LOG_LINES = 50;
 
 export class ServerSupervisor {
-  constructor({
-    serverPath,
-    env,
-    maxRestarts = DEFAULT_MAX_RESTARTS,
-    memoryLimit = 512,
-    onCrashCallback,
-  }) {
+  constructor({ serverPath, env, maxRestarts = DEFAULT_MAX_RESTARTS, memoryLimit = 512, onCrashCallback }) {
     this.serverPath = serverPath;
     this.env = env;
     this.maxRestarts = maxRestarts;
@@ -35,13 +25,11 @@ export class ServerSupervisor {
     this.crashLog = [];
     this.child = null;
     this.isShuttingDown = false;
-    this.instrumentationFailureHintPrinted = false;
   }
 
   start() {
     this.startedAt = Date.now();
     this.crashLog = [];
-    this.instrumentationFailureHintPrinted = false;
 
     const showLog = process.env.OMNIROUTE_SHOW_LOG === "1";
     // #5238: skip the explicit CLI --max-old-space-size when the user pinned the
@@ -53,35 +41,23 @@ export class ServerSupervisor {
     // silently, so a boot that never becomes ready looked like a dead hang with zero
     // output even at APP_LOG_LEVEL=debug. Pipe stdout too and buffer it alongside
     // stderr so a readiness timeout can surface what the child actually printed.
-    this.child = spawn(
-      process.versions.bun ? process.execPath : "node",
-      [...(process.versions.bun ? [] : heapArgs), this.serverPath],
-      {
-        cwd: dirname(this.serverPath),
-        env: this.env,
-        stdio: showLog ? "inherit" : ["ignore", "pipe", "pipe"],
-      }
-    );
+    this.child = spawn(process.versions.bun ? process.execPath : "node", [
+      ...(process.versions.bun ? [] : heapArgs),
+      this.serverPath,
+    ], {
+      cwd: dirname(this.serverPath),
+      env: this.env,
+      stdio: showLog ? "inherit" : ["ignore", "pipe", "pipe"],
+    });
 
     writePidFile("server", this.child.pid);
 
     const bufferOutput = (data) => {
-      const text = data.toString();
-      const lines = text.split("\n").filter(Boolean);
+      const lines = data.toString().split("\n").filter(Boolean);
       this.crashLog.push(...lines);
       if (this.crashLog.length > CRASH_LOG_LINES) {
         this.crashLog = this.crashLog.slice(-CRASH_LOG_LINES);
       }
-      // Surface Android/Termux instrumentation-hook failures even when --log is
-      // off (output is only buffered otherwise).
-      if (!this.instrumentationFailureHintPrinted && isFatalInstrumentationHookFailure(text)) {
-        this.instrumentationFailureHintPrinted = true;
-        process.stderr.write(
-          formatAndroidInstrumentationFailureHint(
-            this.env?.XDG_CACHE_HOME || process.env.XDG_CACHE_HOME
-          )
-        );
-      }
     };
 
     if (this.child.stdout) {

bin/cli/utils/ensureAndroidCacheDir.mjs

@@ -1,122 +0,0 @@
-/**
- * Next.js cache-dir prep for Android / Termux.
- *
- * Next.js `getCacheDirectory()` has no dedicated branch for
- * `process.platform === "android"`. On that path it only accepts a cache root
- * that *already* exists (`fs.existsSync` on `~/.cache` or a generic tmp dir).
- * If neither exists it prints `Unsupported platform: android` and exits — the
- * CLI can still look "running" while every request returns a bare HTTP 500
- * because the instrumentation hook never loads (and so neither does logging).
- *
- * Termux Node sometimes reports `platform === "android"` and sometimes
- * `"linux"` with Termux env signals (`TERMUX_VERSION` / `PREFIX`). Creating
- * `~/.cache` (and pointing `XDG_CACHE_HOME` at it when unset) makes the probe
- * succeed on both shapes.
- *
- * Call this *before* spawning or loading Next.js. Safe no-op on desktop
- * platforms that are not Termux.
- */
-
-import { existsSync, mkdirSync } from "node:fs";
-import { homedir } from "node:os";
-import { join } from "node:path";
-import { isTermux } from "../../../scripts/build/postinstallSupport.mjs";
-
-/**
- * @param {string} [platform]
- * @param {NodeJS.ProcessEnv} [env]
- * @returns {boolean}
- */
-export function needsAndroidCacheDirPrep(platform = process.platform, env = process.env) {
-  return platform === "android" || isTermux(env);
-}
-
-/**
- * @param {() => string} [homedirFn]
- * @param {NodeJS.ProcessEnv} [env]
- * @returns {string}
- */
-export function resolveAndroidCacheDir(homedirFn = homedir, env = process.env) {
-  if (typeof env.XDG_CACHE_HOME === "string" && env.XDG_CACHE_HOME.trim()) {
-    return env.XDG_CACHE_HOME;
-  }
-  return join(homedirFn(), ".cache");
-}
-
-/**
- * Ensure a writable cache directory exists for Next.js on Android/Termux.
- *
- * @param {object} [options]
- * @param {string} [options.platform]
- * @param {NodeJS.ProcessEnv} [options.env]
- * @param {() => string} [options.homedirFn]
- * @param {typeof mkdirSync} [options.mkdirSyncFn]
- * @param {typeof existsSync} [options.existsSyncFn]
- * @param {boolean} [options.setEnv] When true (default), set `XDG_CACHE_HOME` on `env`
- *   if unset so child processes inherit a known-writable cache root.
- * @returns {{ prepared: boolean, cacheDir: string | null, created: boolean }}
- */
-export function ensureAndroidCacheDir(options = {}) {
-  const {
-    platform = process.platform,
-    env = process.env,
-    homedirFn = homedir,
-    mkdirSyncFn = mkdirSync,
-    existsSyncFn = existsSync,
-    setEnv = true,
-  } = options;
-
-  if (!needsAndroidCacheDirPrep(platform, env)) {
-    return { prepared: false, cacheDir: null, created: false };
-  }
-
-  const cacheDir = resolveAndroidCacheDir(homedirFn, env);
-  let created = false;
-  if (!existsSyncFn(cacheDir)) {
-    mkdirSyncFn(cacheDir, { recursive: true });
-    created = true;
-  }
-
-  if (setEnv && !(typeof env.XDG_CACHE_HOME === "string" && env.XDG_CACHE_HOME.trim())) {
-    env.XDG_CACHE_HOME = cacheDir;
-  }
-
-  return { prepared: true, cacheDir, created };
-}
-
-/**
- * Detect Next.js instrumentation-hook failures that leave the server looking
- * "up" while requests get silent HTTP 500s (typical when the Android cache
- * probe failed before logging started).
- *
- * @param {string} text
- * @returns {boolean}
- */
-export function isFatalInstrumentationHookFailure(text) {
-  if (!text) return false;
-  return (
-    /Unsupported platform:\s*android/i.test(text) ||
-    /error occurred while loading instrumentation hook/i.test(text)
-  );
-}
-
-/**
- * Operator-facing hint when that instrumentation failure shows up in child
- * output — defense in depth if prep was skipped or a future Next.js probe
- * regresses.
- *
- * @param {string} [cacheDir]
- * @returns {string}
- */
-export function formatAndroidInstrumentationFailureHint(cacheDir) {
-  const dir = cacheDir || join(homedir(), ".cache");
-  return (
-    `\n\x1b[31m✖ Next.js instrumentation failed on Android/Termux (likely missing cache dir).\x1b[0m\n` +
-    `  OmniRoute tried to create a writable cache at:\n` +
-    `    \x1b[36m${dir}\x1b[0m\n` +
-    `  Manual workaround (survives reinstalls — do NOT patch dist/server.js):\n` +
-    `    \x1b[36mmkdir -p ~/.cache\x1b[0m\n` +
-    `    then restart: \x1b[36momniroute serve\x1b[0m\n` +
-    `  See: docs/guides/TERMUX_GUIDE.md → Troubleshooting → Unsupported platform: android\n`
-  );
-}

bin/cli/utils/winShellArgs.mjs

@@ -1,47 +0,0 @@
-/**
- * Argument escaping for child processes spawned with `shell: true` on Windows.
- *
- * The launchers (`omniroute launch`, `omniroute launch-codex`) must go through
- * cmd.exe on win32 because the target binaries are npm `.cmd` shims that Node
- * cannot exec directly (CVE-2024-27980). With `shell: true` Node joins argv with
- * plain spaces and no escaping at all (the DEP0190 warning), so anything with a
- * space, a quote or a cmd metacharacter reaches the child mangled.
- */
-
-/** cmd.exe metacharacters that stay live inside a quoted argument. */
-const WIN_META_CHARS = /([()\][%!^"`<>&|;, *?])/g;
-
-/**
- * Escape one argument for a cmd.exe command line built by `shell: true`.
- *
- * Two layers, in order:
- *  1. the CRT argv rules the target binary parses (double the backslashes that
- *     precede a quote, escape embedded quotes, wrap in quotes);
- *  2. cmd.exe's metacharacters, caret-escaped — applied TWICE because the
- *     target is an npm `.cmd` shim that forwards `%*` to node, so the line is
- *     parsed by cmd a second time. Single-escaping truncated any argument at
- *     the first `&` or `|`. (Same rule as cross-spawn's doubleEscapeMetaChars.)
- *
- * @param {unknown} arg
- * @returns {string}
- */
-export function escapeWindowsShellArg(arg) {
-  const s = String(arg);
-  if (s === "") return '""';
-  let out = s.replace(/(\\*)"/g, '$1$1\\"').replace(/(\\*)$/, "$1$1");
-  out = `"${out}"`;
-  return out.replace(WIN_META_CHARS, "^$1").replace(WIN_META_CHARS, "^$1");
-}
-
-/**
- * Escape a whole argv for the `shell: true` path. Off Windows there is no shell,
- * so argv is passed through untouched.
- *
- * @param {string[]} args
- * @param {NodeJS.Platform|string} platform
- * @returns {string[]}
- */
-export function quoteShellArgs(args, platform) {
-  const list = [...(args ?? [])];
-  return platform === "win32" ? list.map(escapeWindowsShellArg) : list;
-}

bin/omniroute.mjs

@@ -145,15 +145,6 @@ function loadEnvFile() {
 
 loadEnvFile();
 
-// Next.js has no android branch in getCacheDirectory(): if ~/.cache (and tmp)
-// do not already exist it aborts the instrumentation hook, and every request
-// then returns a silent HTTP 500 even though the CLI still looks "running".
-// Create the cache dir (and set XDG_CACHE_HOME when unset) before serve/Next.
-{
-  const { ensureAndroidCacheDir } = await import("./cli/utils/ensureAndroidCacheDir.mjs");
-  ensureAndroidCacheDir();
-}
-
 // Generate STORAGE_ENCRYPTION_KEY if not set (persisted to ~/.omniroute/.env)
 // This ensures the key survives across upgrades and is not regenerated on each install.
 // See: https://github.com/diegosouzapw/OmniRoute/issues/1622

changelog.d/features/6022-decision-trace-header.md

@@ -0,0 +1 @@
+- **feat(sse):** every completion response now carries an `X-OmniRoute-Decision: strategy=<name>; provider=<alias>; latency_ms=<n>` header exposing the routing decision — `<name>` is the combo strategy (`priority`, `weighted`, `fusion`, etc.) or `single` for a non-combo request — for client-side debugging/analytics without server log access (#6022 — thanks @chirag127).

changelog.d/features/6072-ws-server-inprocess-autostart.md

@@ -0,0 +1 @@
+- **feat(ws):** the live-dashboard WebSocket server now auto-starts in-process (via `instrumentation-node.ts`) across every deployment mode — dev, production, Docker, Electron — with no separate sidecar script; the default WS port moved from 20129 to **20132** to avoid colliding with `API_PORT` in split-port setups, the deprecated `OMNIROUTE_DISABLE_LIVE_WS` env was consolidated into `OMNIROUTE_ENABLE_LIVE_WS` (default enabled), and the WS path is now derived from `NEXT_PUBLIC_LIVE_WS_PUBLIC_URL`'s pathname (`/live-ws` fallback) (#6072 — thanks @ianriizky).

changelog.d/features/6354-per-model-timeout.md

@@ -0,0 +1 @@
+- feat(sse): configurable per-model upstream header-response timeout override, precedence model > provider > global; applied to codex reasoning-heavy tiers (gpt-5.5-high/xhigh, gpt-5.6-\*-high/xhigh) (#6354)

changelog.d/features/6540-hidepaid-ui-selects.md

@@ -0,0 +1 @@
+- **feat(dashboard):** Replace free-text model inputs in the Routing (web search route), Combo Defaults (handoff model), and Background Degradation tabs with a `hidePaidModels`-aware `ModelSelectField`, add a fail-open "paid-only pattern" warning to the per-model routing rule pattern field, and reject paid-only model targets at save time on `PATCH /api/settings`, `PATCH /api/settings/combo-defaults`, and `PUT /api/settings/background-degradation` when `hidePaidModels` is on ([#6540](https://github.com/diegosouzapw/OmniRoute/issues/6540))

changelog.d/features/6556-omniglyph-engine.md

@@ -0,0 +1 @@
+- **feat(compression):** new **omniglyph** engine (context-as-image) — renders system prompt, tool docs, and dense history as compact PNG pages the model reads instead of text (~10× fewer tokens on the converted block; 59–70% end-to-end measured). Works stacked with RTK/Caveman (`stackPriority: 90`) or standalone (`mode: omniglyph`); restricted to Claude Fable 5 over the direct Anthropic route, fail-closed gates with `skip:<reason>` techniques, preview (`stable: false`, off by default) (#6556). Dependency bumped to `omniglyph@^1.0.2` for upstream ReDoS fixes (#6661).

changelog.d/features/6593-ratelimit-admission-control.md

@@ -0,0 +1 @@
+- **feat(sse):** rate-limit request queue admission control — `resilienceSettings.requestQueue.maxQueueDepth` (default `0` = disabled, opt-in 0–100000) fast-rejects a request with a typed `RATE_LIMIT_QUEUE_FULL` error once the local per-provider+connection queue already holds `maxQueueDepth` requests, instead of growing the queue unboundedly; the factory default for `requestQueue.maxWaitMs` (how long a request may wait before being dropped) also fell from 120s to 15s so a saturated queue fails fast (#6593 — thanks @chirag127).

changelog.d/features/6611-native-container-runtimes.md

@@ -0,0 +1 @@
+- **feat(sandbox):** the skill sandbox gained a container-provider abstraction that auto-detects and uses the best native runtime per host — Apple Container (macOS 26+), WSL container (`wslc.exe`), OrbStack, Podman — instead of hardcoding `docker run`, removing the Docker Desktop requirement on macOS/Windows (#6611 — thanks @KooshaPari).

changelog.d/features/6636-codex-session-import.md

@@ -0,0 +1 @@
+- feat(oauth): accept the full ChatGPT session JSON (not just a bare access token) when pasting Codex credentials manually or via `POST /api/oauth/codex/import-token` (#6636)

changelog.d/features/6650-g4f-space-gateway.md

@@ -0,0 +1 @@
+- feat(sse): add 5 no-key g4f.space gateway providers — `g4f-groq`, `g4f-gemini`, `g4f-pollinations`, `g4f-ollama`, `g4f-nvidia` — a free, no-signup reverse proxy (gpt4free project) fronting Groq, Gemini, Pollinations, Ollama, and NVIDIA NIM, rate-limited to 5 req/min per IP (#6650 — thanks @chirag127).

changelog.d/features/6653-deepinfra-video-provider.md

@@ -0,0 +1 @@
+- feat(sse): add DeepInfra as a video-generation provider via its native synchronous inference endpoint (#6653)

changelog.d/features/6654-freepik-pikaso-image-provider.md

@@ -0,0 +1 @@
+- feat(providers): add Freepik (Magnific Mystic) API-key image generation provider — async submit/poll flow with realism/fluid/zen/flexible/super_real/editorial_portraits models (#6654)

changelog.d/features/6655-revai-stt-provider.md

@@ -0,0 +1 @@
+- feat(providers): add Rev AI speech-to-text provider with async job upload/poll/transcript flow (#6655)

changelog.d/features/6656-segmind-image-video-provider.md

@@ -0,0 +1 @@
+- **feat(providers):** add **Segmind** as an image + video generation provider — `x-api-key` auth against `POST https://api.segmind.com/v1/{model}`, with a curated starter model list (Flux, Stable Diffusion XL/3.5, Kandinsky for image; Wan, Hunyuan, LTX, Kling for video) (#6656).

changelog.d/features/6657-gladia-stt-provider.md

@@ -0,0 +1 @@
+- feat(providers): add Gladia as an async speech-to-text provider (#6657)

changelog.d/features/6658-novita-video-gen-provider.md

@@ -0,0 +1 @@
+- feat(video): add Novita AI as a video-generation provider (Wan/Kling async submit-poll) (#6658)

changelog.d/features/6659-speechmatics-stt-provider.md

@@ -0,0 +1 @@
+- **feat(providers):** add Speechmatics as an STT provider — async batch transcription (Enhanced operating point), 8 hours/month free tier, no credit card required. Streaming (real-time) mode is out of scope for v1. (#6659)

changelog.d/features/6660-mixedbread-embeddings-provider.md

@@ -0,0 +1 @@
+- feat(providers): add Mixedbread AI as an embeddings provider (`mxbai-embed-large-v1`, `mxbai-embed-2d-large-v1`, free tier) (#6660)

changelog.d/features/6666-felo-chat-aggregator-provider.md

@@ -0,0 +1 @@
+- **feat(providers):** add Felo (felo.ai) as a free, no-signup, no-API-key chat/search-agent aggregator provider (`felo-web`) — joins the existing `-web` family (DuckDuckGo AI Chat, Blackbox, etc). Five models (`felo-chat`, `felo-search`, `felo-scholar`, `felo-social`, `felo-document`) map to Felo's search categories; the executor opens a search thread then translates Felo's bespoke SSE stream into OpenAI-compatible chunks (#6666).

changelog.d/features/6667-gtts-audio-tts-provider.md

@@ -0,0 +1 @@
+- **feat(providers):** add gTTS (Google Translate TTS) as a free, no-signup audio-speech provider — routes through Google's current `batchexecute` RPC endpoint (the previously proposed `translate_tts` endpoint is deprecated), splitting input at the 100-char-per-request limit. (#6667)

changelog.d/features/6668-edgetts-audio-tts-provider.md

@@ -0,0 +1 @@
+- **feat(sse):** add EdgeTTS (Microsoft Edge "Read Aloud") as a free, no-API-key `audio-tts` provider — the first WebSocket-transport speech provider, with per-client-IP rate limiting. (#6668)

changelog.d/features/6670-freetheai-gateway-provider.md

@@ -0,0 +1 @@
+- feat(providers): add FreeTheAi as an OpenAI-compatible gateway provider with a free Discord-signup tier (#6670)

changelog.d/features/6672-microsoftdesigner-image-provider.md

@@ -0,0 +1 @@
+- feat(sse): add Microsoft Designer as an unofficial web-session image provider, reverse-engineered submit-then-poll DallE.ashx flow (#6672)

changelog.d/features/6673-hailuoai-web-provider.md

@@ -0,0 +1 @@
+- **feat(providers):** add Hailuo Web (`hailuo-web`) — a free, `_token`-based web-cookie chat provider for the MiniMax consumer chat product at hailuo.ai, ported from the g4f reference implementation (MD5-chain request signing, custom `event:`/`data:` SSE parsing). Distinct from the existing paid API-key `minimax`/`minimax-cn` providers. (#6673)

changelog.d/features/6683-cli-auth-export.md

@@ -0,0 +1 @@
+- **feat(cli):** new `omniroute auth export` command dumps DECRYPTED provider credentials (`apiKey`/`accessToken`/`refreshToken`/`idToken`) for one connection (`--id <id>`) or all connections, as `json` or `env` (`--format`), local-only and gated behind `--force` — no DB access happens without it, a stderr warning banner prints before any plaintext, `--out <file>` writes with `0600` permissions, and per-field decrypt failures surface as a `<field>DecryptFailed` boolean instead of aborting the export or leaking the caught error text (#6683)

changelog.d/features/6684-mitm-root-ca.md

@@ -0,0 +1 @@
+- **feat(mitm):** the AgentBridge static MITM server (`server.cjs`) can now issue a per-host TLS leaf from a persisted local root CA (`src/mitm/cert/rootCa.ts`, reusing the CA/leaf crypto already proven for TPROXY in `tproxy/dynamicCert.ts`) instead of a single static self-signed leaf scoped only to the 4 antigravity hosts — a fresh install covers the full `MITM_TOOL_HOSTS` set automatically; an install that already trusted the old leaf keeps using it until the operator opts in via `MITM_ROOT_CA_ENABLED=true` (`src/mitm/cert/migration.ts`), so no existing install is silently upgraded to the more powerful any-host-signing CA trust model (#6684).

changelog.d/features/6708-gemma4-thinkingconfig-guard.md

@@ -0,0 +1 @@
+- **fix(sse):** skip `thinkingConfig` for Gemma models on the OpenAI→Gemini path so OpenAI-shape clients no longer get a 400 from Vertex. (thanks @chy1211)

changelog.d/features/6709-xai-responses-endpoint.md

@@ -0,0 +1 @@
+- **feat(xai):** route xAI clients to Grok's native `/v1/responses` endpoint instead of the chat-completions bridge. (thanks @ryanngit)

changelog.d/features/6727-model-capability-override-ui.md

@@ -0,0 +1 @@
+- **feat(models):** add a Settings → AI "Model Overrides" UI plus `/api/model-capability-overrides` CRUD and a `model_capability_overrides` table, letting operators set a manual max-output-token override per provider/model (#6727 — thanks @xz-dev).

changelog.d/features/6737-vary-accept-encoding.md

@@ -0,0 +1 @@
+- **feat(api):** add `Vary: Accept-Encoding` to token-authenticated `/v1*`/`/v1beta*` responses so downstream caches distinguish compressed vs uncompressed variants (RFC 9110 §12.5.5). (thanks @chirag127)

changelog.d/features/6758-notion-web-provider.md

@@ -0,0 +1 @@
+- feat(sse): add Notion AI Web (Unofficial/Experimental) cookie-session provider (#6758)

changelog.d/features/6760-compression-mode-selector-context-cache.md

@@ -0,0 +1 @@
+- **feat(dashboard):** add per-routing-combo compression-mode override to the Compression Combos page under Context & Cache, alongside the existing combo-card quick override. (#6760)

changelog.d/features/6763-operator-configurable-account-rotation.md

@@ -0,0 +1 @@
+- **feat(resilience):** operator-configurable account rotation policy — a new `rotationConfig` layer lets operators tune how connections rotate on failure, wired into `accountFallback` (#6763 — thanks @artickc).

changelog.d/features/6771-fusion-preserve-tools-bypass.md

@@ -0,0 +1 @@
+- **feat(sse):** preserve `tools`/`tool_choice` for tool-bearing requests through fusion combos — bypass panel synthesis and route straight to the judge with tools intact (#6771 — thanks @chirag127).

changelog.d/features/6774-cursor-grok-4-5-effort-fast.md

@@ -0,0 +1 @@
+- **chore(cursor): add Grok 4.5 effort/fast model IDs** (#6774 — thanks @andrewmunsell).

changelog.d/features/6776-discover-live-codex-models.md

@@ -0,0 +1 @@
+- **feat(codex):** Codex provider model discovery now fetches the live catalog from `chatgpt.com/backend-api/codex/models` using Codex-shaped headers, falling back to a GitHub-hosted model manifest and then to the local static catalog when the live/GitHub sources are unavailable or return an unexpected shape — new `src/app/api/providers/[id]/models/discovery/codex.ts` (normalization, version-gating, merge/enrich against the local catalog) covered by `tests/unit/provider-models-discovery-split.test.ts` and `tests/unit/provider-models-route-codex.test.ts` (#6776 — thanks @JxnLexn).

changelog.d/features/6779-cursor-opus48-fable5-sonnet5.md

@@ -0,0 +1 @@
+- **feat(cursor):** register the Opus 4.8, Fable 5, and Sonnet 5 model families for the Cursor Agent provider so the latest Claude/Fable model ids route correctly (#6779 — thanks @andrewmunsell).

changelog.d/features/6783-changelog-fragments.md

@@ -0,0 +1 @@
+- **Changelog fragments (`changelog.d/`)**: PRs now add their changelog entry as a new fragment file (`changelog.d/{features|fixes|maintenance}/<PR>-<slug>.md`) instead of editing `CHANGELOG.md` — two PRs never touch the same file, structurally eliminating the CHANGELOG-eat merge conflicts that forced a re-sync push + full CI re-run after every sibling merge (O(N²) CI runs in a merge-storm). `scripts/release/aggregate-changelog.mjs` (npm run changelog:aggregate) folds fragments into the living section at release reconciliation, and `check:changelog-integrity` now also validates fragment well-formedness. Regression guard: `tests/unit/changelog-fragments.test.ts`.

changelog.d/features/6798-latency-optimized-proxy-rotation.md

@@ -0,0 +1 @@
+- **feat(proxy):** add a latency-optimized proxy rotation strategy that ranks pool entries by measured round-trip latency, extending the existing round-robin/random/sticky proxy-pool selection (#6798 — thanks @iamraydoan).

changelog.d/features/6801-xp-audit-log-retention.md

@@ -0,0 +1 @@
+- feat(db): include `xp_audit_log` in the automatic retention/prune cycle, with a configurable `retention.xpAuditLog` setting (#6801)

changelog.d/features/6804-fusion-judge-own-knowledge.md

@@ -0,0 +1 @@
+- **feat(fusion):** the fusion judge may now draw on its own knowledge and override the panel when every panel answer is wrong or incomplete, instead of being restricted to synthesizing only from panel output (#6804 — thanks @chirag127).

changelog.d/features/6811-playground-model-picker-search.md

@@ -0,0 +1 @@
+- **feat(dashboard):** search box on the Playground's raw model `<select>` (#4086) — the shared `ModelSelectModal` (combo builder + CLI-code cards) already had search, but Playground's `StudioConfigPane` model dropdown stayed a flat unsearchable list, unusable once a provider like OpenRouter contributed 50+ models. Typing now filters the dropdown (Turkish-safe accent/case-insensitive match via `matchesSearch`), while the currently selected model always stays pinned in the list even if it no longer matches the query, so typing never silently swaps the active selection. Reuses the existing `common.search` i18n key (already translated in all 42 locales) — no new translation key needed. Regression guard: `tests/unit/playground-model-selection-3731.test.ts` (`filterModelsByQuery`), `tests/unit/ui/playground-model-search-4086.test.tsx`.

changelog.d/features/6818-antigravity-weekly-quota.md

@@ -0,0 +1 @@
+- **feat(usage):** Antigravity/agy quota widget now surfaces the **weekly** window alongside the existing per-model 5-hour window ([#4017](https://github.com/diegosouzapw/OmniRoute/issues/4017)) — the weekly limit isn't part of the per-model `retrieveUserQuota` response the fetcher already calls; it only appears in a separate, undocumented `retrieveUserQuotaSummary` RPC that groups models into families ("Gemini Models", "Claude and GPT models") with one weekly bucket per family. A new `usage/antigravityWeeklyQuota.ts` leaf fetches that RPC (cached, best-effort — a failure or unavailable RPC never breaks the existing per-model quotas) and parses the weekly bucket per group into `gemini_weekly`/`claude_gpt_weekly` quota entries, merged into the same `quotas` map the widget already renders generically. Regression guard: `tests/unit/antigravity-weekly-quota-4017.test.ts` (bucket parsing, the alternate `quotaSummary`-nested envelope, and end-to-end merge via `getUsageForProvider`).

changelog.d/features/6820-codex-effort-model-echo.md

@@ -0,0 +1 @@
+- **feat(codex):** Codex CLI compatibility shim — the Responses API `response.created`/`response.in_progress`/`response.completed` payloads now carry a `model` field (previously absent), and for Codex-CLI-originated requests it echoes the client-requested, effort-suffixed model id (e.g. `gpt-5.5-xhigh`) instead of the bare upstream id (`gpt-5.5`), so the Codex CLI status line/model button shows the active reasoning effort ([#3697](https://github.com/diegosouzapw/OmniRoute/issues/3697)). `openaiToOpenAIResponsesResponse` (`open-sse/translator/response/openai-responses.ts`) now threads the upstream model into the Responses event objects; a new `isCodexOriginatedHeaders()` (`open-sse/config/codexIdentity.ts`, reusing PR #3481's `originator`/User-Agent detection) makes chatCore's existing opt-in `echoRequestedModelName` (#1311) model-echo pipeline fire automatically for Codex clients regardless of the setting, detected by request headers so it still applies when `codex/gpt-5.5-xhigh` is routed through a combo to a non-codex upstream; `echoModelInObject`/`echoModelInSseLine` (`open-sse/services/responseModelEcho.ts`) now also rewrite the nested `response.model` field the Responses API uses. `/v1/models` still returns `models: []` for Codex (unchanged). Regression guard: `tests/unit/codex-effort-model-echo-3697.test.ts`.

changelog.d/features/6823-zai-web-cookie-provider.md

@@ -0,0 +1 @@
+- **Z.ai Web (free web-session provider)**: new `zai-web` web-cookie provider drives the free chat.z.ai consumer chat UI via a pasted browser session cookie, distinct from the existing API-key `zai`/`glm`/`glm-cn`/`glmt` providers (`api.z.ai`) — modeled on the `doubao-web`/`venice-web` cookie executors and the pre-existing `chatglm-web` credential requirement/token-extraction entries. `ZaiWebExecutor` (`open-sse/executors/zai-web.ts`) posts to `chat.z.ai/api/chat/completions` with the cookie forwarded both as `Cookie` and as `Authorization: Bearer <token>`, and normalizes both z.ai's internal `delta_content`/`phase` SSE envelope and a pass-through OpenAI-shaped `choices[].delta` frame into standard chat-completion chunks. Registered in `WEB_COOKIE_PROVIDERS`, `WEB_SESSION_CREDENTIAL_REQUIREMENTS`, the provider registry (`zai-web` entry, GLM-4.6/4.5/4.5V models), and `tokenExtractionConfig.ts` for in-app cookie capture. Regression guard: `tests/unit/executor-zai-web.test.ts` (16 tests — token extraction, frame parsing for both SSE shapes, streaming and non-streaming aggregation, error paths). (#4056)

changelog.d/features/6836-import-providers-from-file.md

@@ -0,0 +1 @@
+- feat(dashboard): import multiple, possibly different providers from a CSV/JSON file — per-row validation, a checklist to pick which parsed rows to import, and a new `POST /api/providers/import` route with partial-failure results (#6836)

changelog.d/features/6838-headroom-gcf-v3.2-nested-flattening.md

@@ -0,0 +1 @@
+- **feat(compression):** update the vendored **GCF** codec behind the **Headroom** engine to spec **v3.2 (nested flattening)** ([#6837](https://github.com/diegosouzapw/OmniRoute/issues/6837)). Homogeneous arrays whose rows carry nested objects/arrays now tabularize via `>`-prefixed path fields instead of a low-yield per-row fallback, so nested MCP tool-result rows (`meta:{...}`, `tags:[...]`) compact like flat rows. On representative shapes the update takes deeply-nested payloads the old codec left near-uncompressed from ~3% to ~32% vs JSON (k8s pods, `cl100k_base`), with shallow-nested rows seeing a small bump and flat arrays unchanged. Re-vendored from current gcf-typescript (zero runtime deps, MIT, SPDX-marked, generic-profile only); also folds in the `[N]:` inline-array quoting fix and canonical decimal formatting. Round-trip stays lossless (order-insensitive), and the decoder is hardened against prototype pollution (a `__proto__`/`constructor` path segment never mutates `Object.prototype`, and keys shadowing built-ins like `toString` now round-trip correctly instead of misparsing). Regression guard: `tests/unit/compression/headroom-smartcrusher.test.ts` (deep-nested + prototype-pollution cases).

changelog.d/features/6842-openrouter-quota-tracking.md

@@ -0,0 +1 @@
+- **feat(sse):** OpenRouter quota tracking — a dedicated fetcher polls `/api/v1/key` + `/api/v1/credits` (per-key credit cap/remaining/reset, daily/weekly/monthly USD spend, BYOK usage) with a 45s cache and graceful degradation, a local per-account counter tracks the `:free`-model 50-or-1000-per-day + 20 RPM windows (corrected from `X-RateLimit-*` headers and `Retry-After` on 429), and OpenRouter `402` responses now lock the connection with a real cooldown instead of triggering an immediate reselection of the same credit-exhausted key (#6842).

changelog.d/features/6844-grok-cli-quota.md

@@ -0,0 +1 @@
+- **feat(sse):** live gRPC-web quota fetcher for Grok Build (`grok-cli`) — polls xAI's shared weekly credit pool (`grok.com/grok_api_v2.GrokBuildBilling/GetGrokCreditsConfig`) with the connection's existing bearer token, hand-decoding the framed/raw protobuf response (proto3 omission = 0% used, malformed/expired-token fails open) instead of relying solely on the static 864 req/day estimate, which now serves as an explicit fallback when the live fetch is unavailable (#6844).

changelog.d/features/6845-v0-vercel-quota-tracking.md

@@ -0,0 +1 @@
+- feat(sse): add dual-window quota tracking for the `v0-vercel` provider — polls the credits (`/v1/user/billing`) and daily Platform-API operation (`/v1/rate-limits`) endpoints with the existing routing API key, defensively degrading to an `unknown` billing type rather than misparsing a future v0 billing-model migration, feeding preflight + the dashboard's Provider Quota card (#6845).

changelog.d/features/6846-nvidia-nim-quota.md

@@ -0,0 +1 @@
+- feat(sse): add a static local RPM budget (default 40/min, operator-overridable), per-model 429 scoping (already covered by #6773's `passthroughModels`), and a per-connection concurrency cap (default 6, operator-overridable) for the `nvidia` (NVIDIA NIM) provider, which sends no rate-limit headers and has no usage API — Phase 1 of client-side quota tracking; adaptive AIMD learning and the dashboard quota card are deferred follow-ups (#6846).

changelog.d/features/6850-agentrouter-quota-tracking.md

@@ -0,0 +1 @@
+- feat(sse): add quota tracking for the `agentrouter` provider — polls the New-API `/api/user/self` balance endpoint with a separate System Access Token + `New-Api-User` id (configured via `providerSpecificData.consoleApiKey` / `newApiUserId`), converting raw quota units into a dollar balance and feeding preflight + the dashboard's Provider Quota card (#6850).

changelog.d/features/6862-gpt-5-6-providers.md

@@ -0,0 +1,2 @@
+- **feat(providers):** Add GPT-5.6 support across OpenAI API, Codex, and ChatGPT Web, including Codex Max/Ultra efforts, VS Code metadata, Fast-tier credit accounting, curated live discovery, the Codex 0.144.1 client identity, and correct chat routing for models that also support image generation ([#6862](https://github.com/diegosouzapw/OmniRoute/pull/6862)) - thanks @backryun
+- **chore(providers):** Align emitted Claude Code identity headers, bridge fingerprints, provider profiles, and documented defaults with claude-cli 2.1.207 ([#6862](https://github.com/diegosouzapw/OmniRoute/pull/6862)) - thanks @backryun

changelog.d/features/6872-relay-routing-fallback-reason-header.md

@@ -0,0 +1 @@
+- feat(api): add a structured `X-Routing-Fallback-Reason` header to relay routing responses, exposing a stable machine-readable reason code alongside the legacy `X-Routing-Fallback` detail string (#6872)

changelog.d/features/6873-model-latency-stats-api.md

@@ -0,0 +1 @@
+- **feat(api):** new **GET /api/usage/model-latency-stats** management endpoint exposes the existing rolling per-provider/model latency aggregate (avg/p50/p95/p99, success rate) already used internally by auto-combo routing — supports `windowHours`/`minSamples`/`maxRows`/`provider`/`model` filters (#6873).

changelog.d/features/6874-vibeproxy-openai-provider-node-preset.md

@@ -0,0 +1 @@
+- **feat(providers):** add a `vibeproxy-openai` provider-node preset to `POST /api/provider-nodes` — defaults name/prefix/apiType for VibeProxy's local OpenAI-compatible gateway and normalizes the caller-supplied base URL to its `/v1` root; `baseUrl` remains mandatory. (#6874, idea from #6137 by @KooshaPari)

changelog.d/features/6875-latency-stats-ttft.md

@@ -0,0 +1 @@
+- feat(usage): add avgTtftMs/avgE2ELatencyMs/avgTokensPerSecond to `getModelLatencyStats()` and feed them into auto-combo's speed-ranking factor (#6875)

changelog.d/features/6879-default-reasoning-effort.md

@@ -0,0 +1 @@
+- feat(sse): per-model default `reasoning_effort` (`ModelSpec.defaultReasoningEffort`, injected only when the request carries no reasoning field) and make `no-think/` express `reasoning_effort:"none"` instead of deleting the field on the OpenAI path, so thinks-by-default models actually stop thinking (#6879)

changelog.d/features/6880-connection-cache-override.md

@@ -0,0 +1 @@
+- **feat(providers):** let a custom/openai-compatible connection opt into prompt-cache behavior via a per-connection `cache` capability override, unblocking `prompt_cache_key` injection, the compression cache-aware guard, and `cache_control` passthrough for `openai-compatible-chat-<uuid>`-style connections. (thanks @andrea-kingautomation)

changelog.d/features/6915-free-rankings-auth-type-filter.md

@@ -0,0 +1 @@
+- **feat(dashboard):** add a Type filter (No Signup / OAuth Login / API Key) and an "Easiest first" sort toggle to Free Provider Rankings, so zero-setup NOAUTH providers can be surfaced without eyeballing the Type column. (#6915)

changelog.d/features/6928-comfyui-base-url-field.md

@@ -0,0 +1 @@
+- **feat(providers):** expose an editable base-URL field on the ComfyUI connection so Docker-network setups (e.g. `http://comfyui:8188`) work for image, video, and music generation ([#6928](https://github.com/diegosouzapw/OmniRoute/issues/6928))