fix(qoder): include actionable CLI_QODER_BIN hint in connection test when qodercli is not found (#9277 )

When the Qoder CLI (qodercli) is not detected by getCliRuntimeStatus after an OmniRoute restart (e.g. restricted launch context on Windows where APPDATA/PATH are not inherited), the connection test showed only the non-actionable 'Local CLI runtime is not installed'. Now it surfaces the same buildQoderCliNotFoundHint guidance already used in the executor path, telling the user to set CLI_QODER_BIN to the absolute path of qodercli. Closes #9277
2026-08-08 08:12:20 +03:00 · 2026-08-04 06:20:43 -03:00
340 changed files with 3443 additions and 31792 deletions
--- a/.dockerignore
+++ b/.dockerignore
@@ -18,7 +18,6 @@ coverage
 # Runtime data and logs
 data
 logs
-.sandbox

 # Local env files (inject at runtime via --env-file or -e)
 .env
--- a/.env.devin-bridge.example
+++ b/.env.devin-bridge.example
@@ -1,6 +0,0 @@
-ENABLE_LIVE_DEVIN_TESTS=0
-DEVIN_BRIDGE_MODEL=devin-cli-agentic/swe-1-7
-DEVIN_BRIDGE_SONNET_MODEL=devin-cli-agentic/swe-1-7
-DEVIN_BRIDGE_OPUS_MODEL=devin-cli-agentic/swe-1-7
-DEVIN_BRIDGE_HAIKU_MODEL=devin-cli-agentic/swe-1-7
-DEVIN_BRIDGE_SUBAGENT_MODEL=devin-cli-agentic/swe-1-7
--- a/.env.example
+++ b/.env.example
@@ -1839,18 +1839,6 @@ APP_LOG_TO_FILE=true
 # ── Devin CLI binary path ──
 # Used by: open-sse/executors/devin-cli.ts. Default: looked up via PATH.
 # CLI_DEVIN_BIN=devin
-# Agentic bridge-only binary override. The bridge still executes ACP stdio only.
-# CLI_DEVIN_AGENTIC_BIN=devin
-# Required isolated HOME for the agentic Devin child process.
-# DEVIN_AGENTIC_HOME=/home/bridge
-# Bounded ACP turn timeout in milliseconds. Default: 120000.
-# DEVIN_AGENTIC_ACP_TIMEOUT_MS=120000
-# Agentic bridge model aliases. Values must keep the devin-cli-agentic/ prefix.
-# DEVIN_BRIDGE_MODEL=devin-cli-agentic/swe-1-7
-# DEVIN_BRIDGE_SONNET_MODEL=devin-cli-agentic/swe-1-7
-# DEVIN_BRIDGE_OPUS_MODEL=devin-cli-agentic/swe-1-7
-# DEVIN_BRIDGE_HAIKU_MODEL=devin-cli-agentic/swe-1-7
-# DEVIN_BRIDGE_SUBAGENT_MODEL=devin-cli-agentic/swe-1-7

 # ── Command Code (custom CLI) callback ──
 # Local port used for OAuth-style callbacks from the Command Code CLI helper.
@@ -2312,18 +2300,6 @@ QUOTA_STORE_DRIVER=sqlite              # sqlite | redis
 # ─────────────────────────────────────────────────────────────────────────────
 # HYPERAGENT_USAGE_URL=https://hyperagent.com/api/settings/billing/usage

-# ─────────────────────────────────────────────────────────────────────────────
-# ChatGPT Web (Codex) headless browser and outbound tool tunnel
-# Used by: open-sse/executors/chatgpt-web-codex.ts
-# Connection values entered in the dashboard override these global defaults.
-# ─────────────────────────────────────────────────────────────────────────────
-# CHATGPT_WEB_CODEX_CHROME_PATH=/usr/bin/chromium
-# CHROME_PATH=/usr/bin/chromium
-# CHATGPT_WEB_CODEX_CDP_URL=http://chatgpt-web-codex-browser:9223
-# CHATGPT_WEB_CODEX_TUNNEL_ID=tunnel_0123456789abcdef0123456789abcdef
-# CHATGPT_WEB_CODEX_RUNTIME_KEY=
-# CHATGPT_WEB_CODEX_CONNECTOR_NAME=OmniRoute Codex
-
 # ─────────────────────────────────────────────────────────────────────────────
 # Browser-login VNC sessions (optional — src/lib/vncSession/manifest.ts)
 # Containerized Chromium+VNC used for interactive browser-login credential
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -27,7 +27,7 @@ env:
 jobs:
  changes:
    name: Change Classification
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    outputs:
      code: ${{ steps.classify.outputs.code }}
      docs: ${{ steps.classify.outputs.docs }}
@@ -35,10 +35,13 @@ jobs:
      workflow: ${{ steps.classify.outputs.workflow }}
      testsOnly: ${{ steps.classify.outputs.testsOnly }}
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1
        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.
@@ -74,7 +77,7 @@ jobs:

  lint:
    name: Lint
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: changes
    # P3 (plano mestre): a release-PR viva fica DRAFT o ciclo inteiro — jobs pesados pulam
    # drafts (ciclo v3.8.44: 123 runs pesados re-disparados por merges na release, 88 cancelados).
@@ -88,9 +91,13 @@ jobs:
      API_KEY_SECRET: ci-lint-api-key-secret-long
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@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 audit:deps
@@ -164,7 +171,7 @@ jobs:

  quality-gate:
    name: Quality Ratchet
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    # needs lint so eslint-results artifact is available (same inventory as the
    # blocking lint step). Allow lint failure so other ratchets still run.
    needs: [changes, test-coverage, lint]
@@ -184,9 +191,13 @@ jobs:
      contents: read
      security-events: read
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - uses: ./.github/actions/npm-ci-retry
      - name: Restore ESLint file cache
        uses: actions/cache@v6
@@ -278,7 +289,7 @@ jobs:
  # SonarQube needs SONAR_TOKEN/SONAR_HOST_URL secrets.
  quality-extended:
    name: Quality Gates (Extended)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: changes
    # P3 (plano mestre): a release-PR viva fica DRAFT o ciclo inteiro — jobs pesados pulam
    # drafts (ciclo v3.8.44: 123 runs pesados re-disparados por merges na release, 88 cancelados).
@@ -288,10 +299,14 @@ jobs:
      # fetch-depth: 0 — the OpenAPI breaking-change gate (oasdiff) reads the base
      # spec via `git show <base_ref>:docs/openapi.yaml`; a shallow clone
      # would lack the base ref and the gate would self-skip (base-unresolved).
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - 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
      # Dead-code, cognitive-complexity, type-coverage foram promovidos ao job
      # quality-gate (bloqueante) na Fase 7 INT — não rodam aqui para evitar duplo custo.
@@ -394,16 +409,20 @@ jobs:

  docs-sync-strict:
    name: Docs Sync (Strict)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: changes
    # P3 (plano mestre): a release-PR viva fica DRAFT o ciclo inteiro — jobs pesados pulam
    # drafts (ciclo v3.8.44: 123 runs pesados re-disparados por merges na release, 88 cancelados).
    # Run when docs OR code change: API/route code can break doc/OpenAPI contract gates.
    if: ${{ github.event_name != 'pull_request' || (github.event.pull_request.draft == false && (needs.changes.outputs.docs == 'true' || needs.changes.outputs.code == 'true')) }}
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - uses: ./.github/actions/npm-ci-retry
      - run: npm run check:docs-all
      # Previously-orphaned contract gates (existed as files, never wired anywhere).
@@ -423,7 +442,7 @@ jobs:

  docs-lint:
    name: Docs Lint (prose — advisory)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: changes
    # P3 (plano mestre): a release-PR viva fica DRAFT o ciclo inteiro — jobs pesados pulam
    # drafts (ciclo v3.8.44: 123 runs pesados re-disparados por merges na release, 88 cancelados).
@@ -433,9 +452,13 @@ jobs:
    # existing doc corpus is brought up to style. Promote to blocking once it converges.
    continue-on-error: true
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - name: markdownlint (docs + root, advisory)
        run: npx --yes markdownlint-cli2 "docs/**/*.md" "*.md" "!docs/i18n" "!docs/research" || true
      - name: Vale prose lint (Microsoft style, advisory)
@@ -450,7 +473,7 @@ jobs:

  i18n-ui-coverage:
    name: i18n UI Coverage
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: changes
    # P3 (plano mestre): a release-PR viva fica DRAFT o ciclo inteiro — jobs pesados pulam
    # drafts (ciclo v3.8.44: 123 runs pesados re-disparados por merges na release, 88 cancelados).
@@ -460,10 +483,14 @@ jobs:
      # 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@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - 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
@@ -479,13 +506,17 @@ jobs:
  # without needing app-boot/Playwright infra. Same gating as i18n-ui-coverage.
  i18n-glossary-zhcn:
    name: i18n Glossary (zh-CN)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: changes
    if: ${{ github.event_name != 'pull_request' || (github.event.pull_request.draft == false && (needs.changes.outputs.i18n == 'true' || needs.changes.outputs.code == 'true')) }}
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - 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-glossary-consistency.mjs --locale=zh-CN
      - run: node scripts/i18n/check-glossary-consistency.mjs --locale=zh-TW
@@ -497,7 +528,7 @@ jobs:
  # idioma (a matrix antiga subia 40 artifacts cujo result.txt colidia no merge-multiple).
  i18n:
    name: i18n Validation (all languages)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: changes
    # P3 (plano mestre): a release-PR viva fica DRAFT o ciclo inteiro — jobs pesados pulam
    # drafts (ciclo v3.8.44: 123 runs pesados re-disparados por merges na release, 88 cancelados).
@@ -505,7 +536,7 @@ jobs:
    if: ${{ github.event_name != 'pull_request' || (github.event.pull_request.draft == false && needs.changes.outputs.i18n == 'true') }}
    continue-on-error: true
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
      - uses: actions/setup-python@v7
@@ -540,12 +571,15 @@ jobs:
  pr-test-policy:
    name: PR Test Policy
    if: ${{ github.event_name == 'pull_request' && github.event.pull_request.draft == false }}
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
          fetch-depth: 0
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
      - name: Fetch base branch
        run: git fetch --no-tags origin "${GITHUB_BASE_REF}"
      - name: Validate source changes include tests
@@ -572,17 +606,17 @@ jobs:
    # online), the heavy jobs run on the dedicated 32-core VPS runners (label
    # omni-release) instead of queueing on the 20-concurrent-job hosted pool.
    # Safety: fork PRs NEVER reach the self-hosted runner — the expression falls
-    # back to ubuntu-26.04 unless the PR head repo is this repository (push /
+    # back to ubuntu-latest unless the PR head repo is this repository (push /
    # dispatch events are own-origin by definition). Any failure path (VM down,
-    # var unset/false) also falls back to ubuntu-26.04.
-    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-26.04' }}
+    # var unset/false) also falls back to 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' }}
    needs: changes
    if: ${{ github.event_name != 'pull_request' || (needs.changes.outputs.code == 'true' && github.event.pull_request.draft == false) }}
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: ${{ env.CI_NODE_VERSION }}
          cache: npm
@@ -622,14 +656,18 @@ jobs:

  package-artifact:
    name: Package Artifact
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: build
    env:
      JWT_SECRET: ci-build-secret-with-sufficient-length-for-validation
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - uses: ./.github/actions/npm-ci-retry
      - run: npm run check:node-runtime
      - name: Download Next.js build artifact
@@ -665,15 +703,15 @@ jobs:
    strategy:
      fail-fast: false
      matrix:
-        os: [ubuntu-26.04, windows-latest]
+        os: [ubuntu-latest, windows-latest]
    env:
      JWT_SECRET: ci-build-secret-with-sufficient-length-for-validation
      CSC_IDENTITY_AUTO_DISCOVERY: "false"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: ${{ env.CI_NODE_VERSION }}
          cache: npm
@@ -712,14 +750,14 @@ jobs:

  test-unit:
    name: Unit Tests (${{ matrix.shard }}/8)
-    # Same dynamic-runner rule as Build (own-origin only; fallback ubuntu-26.04).
+    # 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-26.04
+    runs-on: 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
@@ -737,9 +775,13 @@ jobs:
      API_KEY_SECRET: ci-test-api-key-secret-long
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - uses: ./.github/actions/npm-ci-retry
      - run: npm run check:node-runtime
      # QW-d (plano mestre): fonte única — o MESMO npm script dos runs locais (adiciona o
@@ -769,27 +811,31 @@ jobs:

  test-bun-sqlite:
    name: Bun SQLite Compatibility
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    timeout-minutes: 10
    needs: changes
    if: ${{ github.event_name != 'pull_request' || (needs.changes.outputs.code == 'true' && github.event.pull_request.draft == false) }}
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - uses: ./.github/actions/npm-ci-retry
      - run: npm run test:bun:db

  test-vitest:
    name: Vitest (MCP / autoCombo / UI components)
-    # Same dynamic-runner rule as Build (own-origin only; fallback ubuntu-26.04).
+    # 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-26.04
+    runs-on: ubuntu-latest
    timeout-minutes: 15
    # needs: changes (not build) — no artifact consumed; see test-unit note.
    needs: changes
@@ -799,9 +845,13 @@ jobs:
      API_KEY_SECRET: ci-test-api-key-secret-long
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - uses: ./.github/actions/npm-ci-retry
      # The second test runner (CLAUDE.md: "Both test runners must pass") — was never
      # wired into CI until the 2026-06-09 quality audit (Fase 6A.2).
@@ -829,7 +879,7 @@ jobs:
  # the release gate can still exercise them via workflow_dispatch when needed).
  test-coverage:
    name: Coverage
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    # 10min was sized before #7114 added the lcov reporter (Codecov/Sonar need it);
    # merging 8 shard JSONs + text+json+lcov now takes ~10-12min — three consecutive
    # release-tip runs died at exactly 10m as job-timeout "cancelled" (2026-07-15/16).
@@ -840,9 +890,13 @@ jobs:
      JWT_SECRET: ci-test-secret-with-sufficient-length-for-validation
      API_KEY_SECRET: ci-test-api-key-secret-long
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - uses: ./.github/actions/npm-ci-retry
      - name: Download all shard coverage
        uses: actions/download-artifact@v8
@@ -926,14 +980,14 @@ jobs:

  sonarqube:
    name: SonarQube
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: test-coverage
    if: ${{ !cancelled() && needs.test-coverage.result == 'success' }}
    env:
      SONAR_TOKEN: ${{ secrets.SONAR_TOKEN }}
      SONAR_HOST_URL: ${{ secrets.SONAR_HOST_URL }}
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
          fetch-depth: 0
@@ -978,7 +1032,7 @@ jobs:

  coverage-pr-comment:
    name: PR Coverage Comment
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    if: ${{ !cancelled() && github.event_name == 'pull_request' && github.event.pull_request.draft == false && github.event.pull_request.head.repo.fork == false && needs.changes.outputs.code == 'true' }}
    needs:
      - changes
@@ -1057,7 +1111,7 @@ jobs:

  test-e2e:
    name: E2E Tests (${{ matrix.shard }}/9)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    # Build artifact from the `build` job is downloaded instead of rebuilding
    # (~5min saved per shard). 9 shards (up from 6) reduces tests per shard by
    # ~33%. Playwright browser is cached across runs (~1.5min saved per shard).
@@ -1079,9 +1133,13 @@ jobs:
      DISABLE_SQLITE_AUTO_BACKUP: "true"
      OMNIROUTE_PLAYWRIGHT_SKIP_BUILD: "1"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - uses: ./.github/actions/npm-ci-retry
      - run: npm run check:node-runtime
      - name: Cache Playwright browsers
@@ -1130,7 +1188,7 @@ jobs:

  test-integration:
    name: Integration Tests (${{ matrix.shard }}/2)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    timeout-minutes: 15
    # needs: changes (not build) — no artifact consumed; see test-unit note.
    needs: changes
@@ -1146,9 +1204,13 @@ jobs:
      DATA_DIR: /tmp/omniroute-ci-${{ matrix.shard }}
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - uses: ./.github/actions/npm-ci-retry
      - run: npm run check:node-runtime
      # (tsx/esm = QW-b; o alinhamento de ESCOPO do integration com o npm script fica p/ follow-up)
@@ -1156,7 +1218,7 @@ jobs:

  test-security:
    name: Security Tests
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    # needs: changes (not build) — no artifact consumed; see test-unit note.
    needs: changes
    if: ${{ github.event_name != 'pull_request' || (needs.changes.outputs.code == 'true' && github.event.pull_request.draft == false) }}
@@ -1165,16 +1227,20 @@ jobs:
      API_KEY_SECRET: ci-test-api-key-secret-long
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@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 test:security

  ci-summary:
    name: CI Dashboard
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    if: ${{ !cancelled() }}
    needs:
      - changes
--- a/.github/workflows/claude.yml
+++ b/.github/workflows/claude.yml
@@ -21,7 +21,7 @@ jobs:
      (github.event_name == 'pull_request_review_comment' && contains(github.event.comment.body, '@claude')) ||
      (github.event_name == 'pull_request_review' && contains(github.event.review.body, '@claude')) ||
      (github.event_name == 'issues' && (contains(github.event.issue.body, '@claude') || contains(github.event.issue.title, '@claude')))
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: read
@@ -30,7 +30,7 @@ jobs:
      actions: read # Required for Claude to read CI results on PRs
    steps:
      - name: Checkout repository
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
        with:
          persist-credentials: false
          fetch-depth: 1
--- a/.github/workflows/codeql.yml
+++ b/.github/workflows/codeql.yml
@@ -13,7 +13,7 @@ permissions:
 jobs:
  analyze:
    name: Analyze (javascript-typescript)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    permissions:
      security-events: write
      actions: read
--- a/.github/workflows/dast-smoke.yml
+++ b/.github/workflows/dast-smoke.yml
@@ -18,7 +18,7 @@ concurrency:
  cancel-in-progress: true
 jobs:
  dast-smoke:
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    # ADVISORY while this new gate matures (repo convention: advisory -> blocking).
    # Flip to blocking (remove continue-on-error) once it's proven stable across a few PRs.
    continue-on-error: true
@@ -33,6 +33,10 @@ jobs:
      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
        with:
          persist-credentials: false
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
+        with:
+          node-version: "24"
+          cache: npm
      - run: npm ci
      - name: Build CLI bundle
        env:
--- a/.github/workflows/deploy-vps.yml
+++ b/.github/workflows/deploy-vps.yml
@@ -15,7 +15,7 @@ jobs:
      (github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success')
      && vars.DEPLOY_ENABLED == 'true'
    name: Deploy OmniRoute to VPS
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
      - name: Check VPS SSH reachability from runner
        id: reach
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -33,7 +33,7 @@ permissions:
 jobs:
  prepare:
    name: Resolve Docker release metadata
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    outputs:
      version: ${{ steps.version.outputs.version }}
      promote_latest: ${{ steps.version.outputs.promote_latest }}
@@ -42,7 +42,7 @@ jobs:
      IMAGE_NAME: diegosouzapw/omniroute
    steps:
      - name: Checkout
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
        with:
          persist-credentials: false
          ref: ${{ github.event_name == 'workflow_dispatch' && format('refs/tags/v{0}', inputs.version) || '' }}
@@ -145,7 +145,7 @@ jobs:
      GHCR_IMAGE_NAME: ghcr.io/diegosouzapw/omniroute
    steps:
      - name: Checkout
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
        with:
          persist-credentials: false
          ref: ${{ github.event_name == 'workflow_dispatch' && format('refs/tags/v{0}', inputs.version) || '' }}
@@ -233,7 +233,7 @@ jobs:
      - prepare
      - build
    if: needs.prepare.outputs.skip != 'true'
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    permissions:
      contents: read
      packages: write
@@ -245,7 +245,7 @@ jobs:
      PROMOTE_LATEST: ${{ needs.prepare.outputs.promote_latest }}
    steps:
      - name: Checkout
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
        with:
          persist-credentials: false
          ref: ${{ github.event_name == 'workflow_dispatch' && format('refs/tags/v{0}', inputs.version) || '' }}
--- a/.github/workflows/electron-release.yml
+++ b/.github/workflows/electron-release.yml
@@ -20,14 +20,14 @@ permissions:
 jobs:
  validate:
    name: Validate version
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    permissions:
      contents: read
    outputs:
      version: ${{ steps.validate.outputs.version }}
    steps:
      - name: Checkout code
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
        with:
          persist-credentials: false
          fetch-depth: 0
@@ -78,17 +78,17 @@ jobs:
            target: mac-arm64
            ext: -arm64.dmg
          - platform: linux
-            runner: ubuntu-26.04
+            runner: ubuntu-latest
            target: linux
            ext: .AppImage
            deb_ext: .deb

    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
      - name: Setup Node
-        uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+        uses: actions/setup-node@v7
        with:
          node-version: 24
          cache: npm
@@ -239,12 +239,12 @@ jobs:
    # 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-26.04
+    runs-on: ubuntu-latest
    permissions:
      contents: write # softprops/action-gh-release creates the GitHub Release
    steps:
      - name: Checkout
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
        with:
          persist-credentials: false
          fetch-depth: 0
@@ -329,7 +329,7 @@ jobs:
    # 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-26.04
+    runs-on: ubuntu-latest
    permissions:
      contents: read
    steps:
--- a/.github/workflows/lock-released-branch.yml
+++ b/.github/workflows/lock-released-branch.yml
@@ -40,7 +40,7 @@ jobs:
  # ─────────────────────────────────────────────────────────────────────────
  lock-branch:
    if: github.event_name == 'release' || github.event_name == 'workflow_dispatch'
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
      - name: Lock release/<tag> branch
        env:
@@ -97,7 +97,7 @@ jobs:
  # ─────────────────────────────────────────────────────────────────────────
  guard-no-push-after-release:
    if: github.event_name == 'push'
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
      - name: Reject push if matching release tag exists
        env:
--- a/.github/workflows/mutation-redundancy.yml
+++ b/.github/workflows/mutation-redundancy.yml
@@ -22,7 +22,7 @@ permissions:
 jobs:
  stryker-nobail:
    name: Stryker disableBail (batch ${{ matrix.batch.name }})
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    strategy:
      fail-fast: false
      matrix:
@@ -41,9 +41,13 @@ jobs:
            mutate: "open-sse/handlers/chatCore/telemetryHelpers.ts,open-sse/handlers/chatCore/memorySkillsInjection.ts,open-sse/handlers/chatCore/semanticCache.ts"
    timeout-minutes: 300
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: "24"
+          cache: npm
      - run: npm ci
      - name: Run Stryker (disableBail)
        env:
--- a/.github/workflows/nightly-compat.yml
+++ b/.github/workflows/nightly-compat.yml
@@ -28,11 +28,11 @@ concurrency:
 jobs:
  resolve-branch:
    name: Resolve active release branch
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    outputs:
      target: ${{ steps.branch.outputs.target }}
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          fetch-depth: 0
          persist-credentials: false
@@ -58,15 +58,15 @@ jobs:

  compat-build-26:
    name: Node 26 Compatibility Build
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    timeout-minutes: 25
    needs: resolve-branch
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          ref: ${{ needs.resolve-branch.outputs.target }}
          persist-credentials: false
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: "26"
          cache: npm
@@ -75,7 +75,7 @@ jobs:
      # 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 16 GB hosted runner. Turbopack's
+      # 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
@@ -88,7 +88,7 @@ jobs:

  compat-tests:
    name: Node ${{ matrix.node }} Compat Tests (${{ matrix.shard }}/4)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    timeout-minutes: 25
    needs: resolve-branch
    strategy:
@@ -102,11 +102,11 @@ jobs:
      DISABLE_SQLITE_AUTO_BACKUP: "true"
      TEST_SHARD: ${{ matrix.shard }}/4
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          ref: ${{ needs.resolve-branch.outputs.target }}
          persist-credentials: false
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: ${{ matrix.node }}
          cache: npm
@@ -116,7 +116,7 @@ jobs:

  report:
    name: Open / update tracking issue on failure
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    if: ${{ !cancelled() && (needs.compat-tests.result == 'failure' || needs.compat-build-26.result == 'failure') }}
    needs: [resolve-branch, compat-build-26, compat-tests]
    permissions:
--- a/.github/workflows/nightly-llm-security.yml
+++ b/.github/workflows/nightly-llm-security.yml
@@ -10,11 +10,13 @@ permissions:
 jobs:
  promptfoo-guard:
    name: promptfoo — injection guard (block mode, no secret)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with: { node-version: "24", cache: npm }
      - run: npm ci
      - name: Build CLI bundle
        env:
@@ -44,7 +46,7 @@ jobs:

  garak:
    name: garak probes (skip without provider secret)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    # NOTE: the `secrets` context is NOT available in a job-level `if:` — referencing
    # it there makes GitHub reject the file on push (startup_failure on every push).
    # Map the secret into a job-level env and gate each step on a presence check, so
@@ -61,10 +63,13 @@ jobs:
            echo "run=false" >> "$GITHUB_OUTPUT"
            echo "::notice::PROMPTFOO_PROVIDER_KEY not set — skipping garak probes (advisory)."
          fi
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
        if: steps.gate.outputs.run == 'true'
+      - uses: actions/setup-node@v7
+        if: steps.gate.outputs.run == 'true'
+        with: { node-version: "24", cache: npm }
      - run: npm ci
        if: steps.gate.outputs.run == 'true'
      - name: Build CLI bundle
--- a/.github/workflows/nightly-mutation.yml
+++ b/.github/workflows/nightly-mutation.yml
@@ -10,7 +10,7 @@ permissions:
 jobs:
  stryker:
    name: Stryker mutation (batch ${{ matrix.batch.name }} — advisory)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    # Mutation testing is expensive. History of the budget:
    #   - Full 8-module set TIMED OUT at the 180min cap (run 27705123780 = exactly 180min).
    #     The two god-files chatCore.ts/combo.ts dominated ~2/3 of the mutants and were
@@ -104,9 +104,13 @@ jobs:
    # scripts/quality/mutation-radiography.mjs both merge per file).
    timeout-minutes: ${{ matrix.batch.timeout || 180 }}
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: "24"
+          cache: npm
      - run: npm ci
      - name: Restore Stryker incremental cache
        uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
@@ -141,12 +145,15 @@ jobs:
    name: Mutation score ratchet (blocking)
    needs: stryker
    if: always()
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    timeout-minutes: 15
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: "24"
      - name: Download all mutation reports
        uses: actions/download-artifact@v8
        with:
--- a/.github/workflows/nightly-property.yml
+++ b/.github/workflows/nightly-property.yml
@@ -8,11 +8,15 @@ permissions:
  issues: write
 jobs:
  property-random-seed:
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: "24"
+          cache: npm
      - run: npm ci
      - name: fast-check random seed (high runs)
        id: prop
--- a/.github/workflows/nightly-release-green.yml
+++ b/.github/workflows/nightly-release-green.yml
@@ -68,13 +68,13 @@ jobs:
    # this runs on the dedicated VPS runner — clean env (no operator OMNIROUTE_API_KEY,
    # no local noauth CLIs => zero machine-specific false positives) and no contention.
    # Nightly cron normally finds the var false (VM off) and falls back to hosted.
-    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && fromJSON('["self-hosted","omni-release"]')) || 'ubuntu-26.04' }}
+    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && fromJSON('["self-hosted","omni-release"]')) || 'ubuntu-latest' }}
    env:
      JWT_SECRET: ci-nightly-secret-with-sufficient-length-for-validation
      API_KEY_SECRET: ci-nightly-api-key-secret-long
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          fetch-depth: 0
          persist-credentials: false
@@ -116,7 +116,7 @@ jobs:
          git checkout "$TARGET"
          git log -1 --oneline

-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: "24"
          cache: npm
@@ -217,19 +217,19 @@ jobs:
    # On a push, only run for a push to main — a push to release/* is handled by
    # release-green above. Schedule/dispatch always run (they also sweep main).
    if: ${{ github.event_name != 'push' || github.ref_name == 'main' }}
-    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && fromJSON('["self-hosted","omni-release"]')) || 'ubuntu-26.04' }}
+    runs-on: ${{ (vars.USE_VPS_RUNNER == 'true' && fromJSON('["self-hosted","omni-release"]')) || 'ubuntu-latest' }}
    env:
      JWT_SECRET: ci-nightly-secret-with-sufficient-length-for-validation
      API_KEY_SECRET: ci-nightly-api-key-secret-long
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          ref: main # literal — no injection surface; scheduled runs default to the repo default branch (a release/v*), so pin main explicitly
          fetch-depth: 0
          persist-credentials: false

-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: "24"
          cache: npm
@@ -331,12 +331,12 @@ jobs:
  bank-ratchet-shrinks:
    name: Bank ratchet shrinks
    if: ${{ github.event_name != 'push' }}
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    permissions:
      contents: write
      pull-requests: write
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          fetch-depth: 0

@@ -371,6 +371,11 @@ jobs:
          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
--- a/.github/workflows/nightly-resilience.yml
+++ b/.github/workflows/nightly-resilience.yml
@@ -10,31 +10,43 @@ permissions:
 jobs:
  heap:
    name: Heap-growth gate
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: "24"
+          cache: npm
      - run: npm ci
      - run: npm run test:heap

  chaos:
    name: Resilience chaos (fault injection)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: "24"
+          cache: npm
      - run: npm ci
      - run: npm run test:chaos

  k6-soak:
    name: k6 load/soak
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: "24"
+          cache: npm
      - run: npm ci
      - name: Build CLI bundle
        env:
@@ -66,7 +78,7 @@ jobs:

  a11y:
    name: A11y axe (nightly, freeze-and-alert)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    # The Playwright webServer (`start` mode) builds Next via build-next-isolated.mjs and
    # boots the standalone server itself (waits on /api/monitoring/health, 15min webServer
    # timeout). Unlike the per-PR test-e2e job, this nightly job has no pre-built artifact,
@@ -80,9 +92,13 @@ jobs:
      DISABLE_SQLITE_AUTO_BACKUP: "true"
      REQUIRE_AXE: "1"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: "24"
+          cache: npm
      - run: npm ci
      - name: Cache Playwright browsers
        uses: actions/cache@v6.1.0
--- a/.github/workflows/nightly-schemathesis.yml
+++ b/.github/workflows/nightly-schemathesis.yml
@@ -10,12 +10,14 @@ permissions:
 jobs:
  schemathesis:
    name: Schemathesis — OpenAPI contract fuzz (advisory)
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    timeout-minutes: 30
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with: { node-version: "24", cache: npm }
      - run: npm ci
      - name: Build CLI bundle
        env:
--- a/.github/workflows/npm-publish.yml
+++ b/.github/workflows/npm-publish.yml
@@ -62,7 +62,7 @@ jobs:
    # 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-26.04' }}
+    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' }}
    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)
@@ -70,7 +70,7 @@ jobs:
      packages: write # publish to npm.pkg.github.com
    steps:
      - name: Checkout
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
        with:
          persist-credentials: false
          # Need full tag history to compare against highest semver when
@@ -78,7 +78,7 @@ jobs:
          fetch-depth: 0

      - name: Setup Node.js
-        uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+        uses: actions/setup-node@v7
        with:
          node-version: ${{ env.NPM_PUBLISH_NODE_VERSION }}
          registry-url: https://registry.npmjs.org
@@ -339,20 +339,20 @@ jobs:
          echo "✅ Action finished for GitHub Packages"

  publish-opencode-plugin:
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    permissions:
      contents: read
      id-token: write # npm provenance
    steps:
      - name: Checkout
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
        with:
          persist-credentials: false
          fetch-depth: 0
          # Full history needed for auto-bump: git diff against previous release tag

      - name: Setup Node.js
-        uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+        uses: actions/setup-node@v7
        with:
          node-version: ${{ env.NPM_PUBLISH_NODE_VERSION }}
          registry-url: https://registry.npmjs.org
--- a/.github/workflows/opencode-plugin-ci.yml
+++ b/.github/workflows/opencode-plugin-ci.yml
@@ -26,16 +26,16 @@ defaults:
 jobs:
  test:
    name: Test (Node ${{ matrix.node }})
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    strategy:
      fail-fast: false
      matrix:
        node: ["22", "24"]
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: ${{ matrix.node }}
          cache: npm
@@ -46,13 +46,13 @@ jobs:

  build:
    name: Build
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: test
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: "22"
          cache: npm
--- a/.github/workflows/opencode-provider-ci.yml
+++ b/.github/workflows/opencode-provider-ci.yml
@@ -26,16 +26,16 @@ defaults:
 jobs:
  test:
    name: Test (Node ${{ matrix.node }})
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    strategy:
      fail-fast: false
      matrix:
        node: ["20", "22", "24"]
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: ${{ matrix.node }}
          cache: npm
@@ -45,13 +45,13 @@ jobs:

  build:
    name: Build
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    needs: test
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
-      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
+      - uses: actions/setup-node@v7
        with:
          node-version: "20"
          cache: npm
--- a/.github/workflows/quality.yml
+++ b/.github/workflows/quality.yml
@@ -25,17 +25,20 @@ jobs:
  # path filters share existence reasons: code / docs / i18n / workflow.
  changes:
    name: Change Classification
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    outputs:
      code: ${{ steps.classify.outputs.code }}
      docs: ${{ steps.classify.outputs.docs }}
      i18n: ${{ steps.classify.outputs.i18n }}
      workflow: ${{ steps.classify.outputs.workflow }}
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
          fetch-depth: 0
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
      - id: classify
        env:
          EVENT_NAME: ${{ github.event_name }}
@@ -58,16 +61,19 @@ jobs:
    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') }}
-    # Fork-safe fallback uses Ubuntu 26.04's bundled Node 24 without setup-node.
-    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-26.04' }}
+    # 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 # v7.0.1
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
        with:
          persist-credentials: false
-      - run: node -e 'if (process.versions.node.split(".")[0] !== process.env.CI_NODE_VERSION) throw new Error("Expected Node " + process.env.CI_NODE_VERSION)'
+      - 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
@@ -82,11 +88,15 @@ jobs:
    name: Docs Gates (fast-path)
    needs: changes
    if: ${{ github.event_name != 'pull_request' || ((github.event.pull_request.draft == false || startsWith(github.head_ref, 'mergify/merge-queue/')) && (needs.changes.outputs.docs == 'true' || needs.changes.outputs.code == 'true')) }}
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - run: npm ci
      # One walk of src/app/api for openapi-routes + docs-symbols (both still fail independently).
      - run: npm run check:api-docs-refs
@@ -101,7 +111,7 @@ jobs:
    # Dynamic runner (same rule as ci.yml): use the self-hosted VPS pool only when the
    # 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-26.04, so this is inert until the flag flips.
+    # 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
@@ -115,7 +125,7 @@ jobs:
    #
    # 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-26.04
+    runs-on: 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:
@@ -123,10 +133,14 @@ jobs:
      API_KEY_SECRET: ci-lint-api-key-secret-long
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          fetch-depth: 0
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - run: npm ci
      - name: Restore ESLint file cache
        uses: actions/cache@v6
@@ -265,7 +279,7 @@ jobs:
      # selector returns __RUN_ALL__ — full-suite authority is the parallel
      # `fast-unit` 4-shard job (test:unit:ci:shard; was 2-shard, #6781), NOT an
      # unsharded re-run here. Stacking unsharded test:unit:ci on top of fast-unit
-      # doubled wall time (~16 min extra on the hosted runner) without extra coverage.
+      # doubled wall time (~16 min extra on ubuntu-latest) without extra coverage.
      #
      # BLOCKING for the *impacted subset* (flipped 2026-06-17). Fail-safe full
      # coverage remains required via `Unit Tests fast-path` (fast-unit).
@@ -329,15 +343,36 @@ jobs:
          if-no-files-found: ignore
          retention-days: 30

-      # Share fast-gates' checkout + npm ci instead of spending ~80s preparing a
-      # separate runner for a ~13s Vitest invocation. !cancelled() preserves the
-      # independent test signal when an earlier fast gate fails.
-      - name: Vitest
-        if: ${{ !cancelled() }}
-        run: npm run test:vitest -- --reporter=default --reporter=junit --outputFile.junit=trunk-junit/vitest-fastpath.xml
-      # WS5.2/5.3: JUnit feeds Trunk Flaky Tests — the fast path runs on every PR.
-      # Advisory upload, own-origin only.
-      - name: Upload Vitest results to Trunk (advisory)
+  fast-vitest:
+    name: Vitest (fast-path)
+    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
+    env:
+      JWT_SECRET: ci-lint-secret-with-sufficient-length-for-validation
+      API_KEY_SECRET: ci-lint-api-key-secret-long
+      DISABLE_SQLITE_AUTO_BACKUP: "true"
+    steps:
+      - uses: actions/checkout@v7
+        with:
+          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
+      - run: npm ci
+      # WS5.2/5.3: JUnit feeds Trunk Flaky Tests — the fast-path runs on EVERY PR,
+      # which is where flaky-detection volume actually comes from (ci.yml's heavy
+      # jobs only run on the release PR). Advisory upload, own-origin only.
+      - run: npm run test:vitest -- --reporter=default --reporter=junit --outputFile.junit=trunk-junit/vitest-fastpath.xml
+      - name: Upload test results to Trunk (advisory)
        if: ${{ always() && (github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository) }}
        continue-on-error: true
        uses: trunk-io/analytics-uploader@385f1ccdf345b4532dc4b6c665dd432b702b8e28 # v2.1.2
@@ -350,9 +385,9 @@ jobs:
    name: Unit Tests fast-path (${{ matrix.shard }}/4)
    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-26.04).
+    # Dynamic runner — see fast-gates (own-origin + flag; fork/unset → ubuntu-latest).
    # This is the heaviest fast-path job; 4-way sharding (was 2, #6781) halves the
-    # critical path again (~8.5min → ~4.5min hosted; ~2min on the 8-slot
+    # 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
@@ -361,7 +396,7 @@ jobs:
    # 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-26.04
+    runs-on: ubuntu-latest
    strategy:
      fail-fast: false
      matrix:
@@ -371,9 +406,13 @@ jobs:
      API_KEY_SECRET: ci-lint-api-key-secret-long
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - run: npm ci
      # QW-d: fonte única — o mesmo npm script do CI pesado/local. Fecha dois drifts do
      # comando inline antigo: os dirs `memory` e `usage` estavam FORA do glob (testes
@@ -398,7 +437,7 @@ jobs:
    name: No new ESLint warnings
    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') }}
-    runs-on: ubuntu-26.04
+    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
@@ -407,9 +446,13 @@ jobs:
      contents: read
      security-events: read
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@v7
        with:
          persist-credentials: false
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - run: npm ci
      - name: Restore ESLint file cache
        uses: actions/cache@v6
@@ -462,17 +505,21 @@ jobs:
    name: Merge integrity (changelog + generated skills)
    # Always on non-draft PRs — CHANGELOG/skills can break on docs-only merges too.
    if: ${{ github.event_name != 'pull_request' || (github.event.pull_request.draft == false || startsWith(github.head_ref, 'mergify/merge-queue/')) }}
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    continue-on-error: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.repo.fork == true }}
    env:
      JWT_SECRET: ci-lint-secret-with-sufficient-length-for-validation
      API_KEY_SECRET: ci-lint-api-key-secret-long
      DISABLE_SQLITE_AUTO_BACKUP: "true"
    steps:
-      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
        with:
          fetch-depth: 0
          persist-credentials: false
+      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v6
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
      - run: npm ci
      - name: CHANGELOG integrity (nenhum bullet da base pode sumir no merge-result)
        run: npm run check:changelog-integrity
--- a/.github/workflows/scorecard.yml
+++ b/.github/workflows/scorecard.yml
@@ -11,7 +11,7 @@ permissions: read-all
 jobs:
  analysis:
    name: Scorecard analysis
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    permissions:
      # security-events: write removed — Scorecard findings are advisory and no longer
      # uploaded to the code-scanning Security tab (they are supply-chain/posture scores,
@@ -21,7 +21,7 @@ jobs:
      contents: read
    steps:
      - name: Checkout
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
        with:
          persist-credentials: false

--- a/.github/workflows/semgrep.yml
+++ b/.github/workflows/semgrep.yml
@@ -14,7 +14,7 @@ concurrency:
  cancel-in-progress: true
 jobs:
  semgrep:
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    container:
      image: semgrep/semgrep
    steps:
--- a/.github/workflows/wiki-sync.yml
+++ b/.github/workflows/wiki-sync.yml
@@ -34,10 +34,15 @@ concurrency:
 jobs:
  sync-wiki:
    name: Sync wiki with docs
-    runs-on: ubuntu-26.04
+    runs-on: ubuntu-latest
    steps:
      - name: Checkout repo
-        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        uses: actions/checkout@v7
+
+      - name: Setup Node
+        uses: actions/setup-node@v7
+        with:
+          node-version: "24"

      - name: Clone wiki
        env:
--- a/.gitignore
+++ b/.gitignore
@@ -72,7 +72,6 @@ yarn-error.log*
 # env files (can opt-in for committing if needed)
 .env*
 !.env.example
-!.env.devin-bridge.example
 !.env.homolog.example
 # Provider API keys (never commit)
 *.api-key
@@ -210,8 +209,6 @@ scripts/i18n/_pending-keys.json
 .agents/
 .antigravitycli/
 .claude/
-!tests/fixtures/devin-bridge/e2e-workspace/.claude/
-!tests/fixtures/devin-bridge/e2e-workspace/.claude/**

 # PR Reviews and local feedback files
 pr_reviews*.json
@@ -248,8 +245,6 @@ _artifacts/          # release-green artifacts

 # CI/local quality artifacts (eslint-results.json, quality-ratchet.md, etc.)
 .artifacts/
-# Isolated Devin bridge workspaces, evidence, and test databases
-.sandbox/

 # Homologation E2E suite (npm run homolog) — real-environment credentials + report output
 .env.homolog
--- a/@omniroute/opencode-plugin/package-lock.json
+++ b/@omniroute/opencode-plugin/package-lock.json
@@ -1,12 +1,12 @@
 {
  "name": "@omniroute/opencode-plugin",
-  "version": "0.2.1",
+  "version": "0.2.0",
  "lockfileVersion": 3,
  "requires": true,
  "packages": {
    "": {
      "name": "@omniroute/opencode-plugin",
-      "version": "0.2.1",
+      "version": "0.2.0",
      "license": "MIT",
      "dependencies": {
        "zod": "^4.4.3"
--- a/@omniroute/openhands-plugin/.gitignore
+++ b/@omniroute/openhands-plugin/.gitignore
@@ -1,5 +0,0 @@
-node_modules
-dist
-*.log
-.DS_Store
-.env
--- a/@omniroute/openhands-plugin/LICENSE
+++ b/@omniroute/openhands-plugin/LICENSE
@@ -1,21 +0,0 @@
-MIT License
-
-Copyright (c) 2026 OmniRoute contributors
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
--- a/@omniroute/openhands-plugin/README.md
+++ b/@omniroute/openhands-plugin/README.md
@@ -1,118 +0,0 @@
-# @omniroute/openhands-plugin
-
-OpenHands integration for the **OmniRoute AI Gateway**. Generates the OpenHands
-environment and Docker config that wires an OpenHands agent-server to a running
-OmniRoute instance — with the integration gotchas already handled.
-
-## Why
-
-Running OpenHands against OmniRoute directly hits several wall:
-
-1. **Model name mismatch** — OpenHands sends `model: "deepseek-chat"`, OmniRoute
-   uses provider-prefixed IDs (`ds/deepseek-v4-flash`) or combos.
-2. **Python 3.13 sandbox** — `socket.socketpair()` fails under Docker's default
-   seccomp profile; the agent-server needs `privileged: true`.
-3. **Host reachability** — the sandbox can't resolve `localhost` to the OmniRoute
-   host; needs `host.docker.internal:host-gateway`.
-4. **Lost state** — conversations die with the container unless
-   `OH_PERSISTENCE_DIR` is a host volume.
-5. **CORS** — the dashboard origin can't reach agent-server unless
-   `PERMITTED_CORS_ORIGINS` allows it.
-
-This plugin encodes all of that into one command.
-
-## Install
-
-```bash
-npm install -g @omniroute/openhands-plugin
-# or: npx @omniroute/openhands-plugin ...
-```
-
-## Quick start
-
-Generate the OpenHands `.env`:
-
-```bash
-omniroute-openhands env \
-  --api-key sk-... \
-  --model deepseek-chat \
-  --url http://192.168.3.106:20128
-```
-
-Generate a `docker-compose.yml` service:
-
-```bash
-omniroute-openhands compose \
-  --api-key sk-... \
-  --model glm-5.2 \
-  --persistence-dir /Users/me/.openhands-state \
-  --cors-origins http://100.73.44.17:3000
-```
-
-Or a plain `docker run`:
-
-```bash
-omniroute-openhands docker-run \
-  --api-key sk-... \
-  --model vivanta-core \
-  --persistence-dir /Users/me/.openhands-state
-```
-
-## Commands
-
-| Command | Description |
-|---------|-------------|
-| `env` | Print OpenHands `.env` contents |
-| `compose` | Print a Docker Compose service block |
-| `docker-run` | Print a full `docker run` command |
-| `models` | Print the default OpenHands → OmniRoute model map |
-
-### Common options
-
-| Flag | Description | Default |
-|------|-------------|---------|
-| `--api-key` | OmniRoute API key (`sk-...`) | — |
-| `--model` | OpenHands model name or OmniRoute combo | — |
-| `--url` | OmniRoute base URL | `http://localhost:20128` |
-| `--persistence-dir` | Host dir for conversation state | `.openhands-state` |
-| `--cors-origins` | Comma-separated allowed origins | `localhost:3000,3001` |
-| `--sandbox-image` | OpenHands sandbox base image | — |
-
-## Model mapping
-
-OpenHands-friendly names are mapped to OmniRoute IDs/combo names:
-
-| OpenHands sends | OmniRoute resolves to |
-|-----------------|----------------------|
-| `deepseek-chat` | `ds/deepseek-v4-flash` |
-| `deepseek-reasoner` | `ds/deepseek-v4-pro` |
-| `glm-5.2` | `nvidia/z-ai/glm-5.2` |
-| `gpt-4o` | `openai/gpt-4o` |
-| `claude-sonnet-4.5` | `anthropic/claude-sonnet-4.5` |
-| ... | ... |
-
-Or just pass an OmniRoute combo name (e.g. `--model vivanta-core`) — the Model
-Alias Resolver and combo router accept it directly.
-
-## Library usage
-
-```ts
-import {
-  buildOpenHandsEnv,
-  serializeOpenHandsEnv,
-  buildOpenHandsCompose,
-  resolveOpenHandsModel,
-} from "@omniroute/openhands-plugin";
-
-const env = buildOpenHandsEnv({
-  apiKey: "sk-...",
-  model: resolveOpenHandsModel("deepseek-chat"),
-  omnirouteUrl: "http://localhost:20128",
-  persistenceDir: "/Users/me/.openhands-state",
-});
-console.log(serializeOpenHandsEnv(env));
-```
-
-## License
-
-MIT — same as OmniRoute.
--- a/@omniroute/openhands-plugin/package-lock.json
+++ b/@omniroute/openhands-plugin/package-lock.json
--- a/@omniroute/openhands-plugin/package.json
+++ b/@omniroute/openhands-plugin/package.json
@@ -1,79 +0,0 @@
-{
-  "name": "@omniroute/openhands-plugin",
-  "version": "0.1.0",
-  "description": "OpenHands integration for the OmniRoute AI Gateway. Generates OpenHands env + Docker Compose config (model mapping, sandbox, CORS, persistence) so OpenHands agents talk to OmniRoute out of the box.",
-  "type": "module",
-  "main": "./dist/index.js",
-  "types": "./dist/index.d.ts",
-  "bin": {
-    "omniroute-openhands": "./dist/cli.js"
-  },
-  "exports": {
-    ".": {
-      "types": "./dist/index.d.ts",
-      "import": "./dist/index.js"
-    },
-    "./env": {
-      "types": "./dist/env.d.ts",
-      "import": "./dist/env.js"
-    },
-    "./docker": {
-      "types": "./dist/docker.d.ts",
-      "import": "./dist/docker.js"
-    },
-    "./model-map": {
-      "types": "./dist/model-map.d.ts",
-      "import": "./dist/model-map.js"
-    }
-  },
-  "files": [
-    "dist",
-    "README.md",
-    "LICENSE"
-  ],
-  "scripts": {
-    "build": "tsup",
-    "clean": "rm -rf dist",
-    "test": "node --import tsx/esm --test tests/env.test.ts tests/model-map.test.ts tests/docker.test.ts",
-    "prepublishOnly": "npm run clean && npm run build && npm test"
-  },
-  "keywords": [
-    "omniroute",
-    "openhands",
-    "open-hands",
-    "openhands-plugin",
-    "openai-compatible",
-    "docker",
-    "agent"
-  ],
-  "author": "OmniRoute contributors",
-  "license": "MIT",
-  "repository": {
-    "type": "git",
-    "url": "https://github.com/diegosouzapw/OmniRoute.git",
-    "directory": "@omniroute/openhands-plugin"
-  },
-  "bugs": {
-    "url": "https://github.com/diegosouzapw/OmniRoute/issues"
-  },
-  "homepage": "https://github.com/diegosouzapw/OmniRoute/tree/main/%40omniroute/openhands-plugin#readme",
-  "engines": {
-    "node": ">=18.0.0"
-  },
-  "peerDependencies": {
-    "@omniroute/open-sse": "*"
-  },
-  "peerDependenciesMeta": {
-    "@omniroute/open-sse": {
-      "optional": true
-    }
-  },
-  "publishConfig": {
-    "access": "public"
-  },
-  "devDependencies": {
-    "@types/node": "^22.19.19",
-    "tsup": "^8.5.1",
-    "tsx": "^4.22.3"
-  }
-}
--- a/@omniroute/openhands-plugin/src/cli.ts
+++ b/@omniroute/openhands-plugin/src/cli.ts
@@ -1,104 +0,0 @@
-#!/usr/bin/env node
-/**
- * @omniroute/openhands-plugin CLI — generate OpenHands .env / Docker config
- * for a running OmniRoute instance.
- *
- * Usage:
- *   omniroute-openhands env   --api-key sk-... --model deepseek-chat [--url http://localhost:20128]
- *   omniroute-openhands compose --api-key sk-... --model deepseek-chat [--persistence-dir /path]
- *   omniroute-openhands docker-run --api-key sk-... --model deepseek-chat
- *   omniroute-openhands models  (print the default model map)
- */
-import { buildOpenHandsEnv, serializeOpenHandsEnv } from "./env.ts";
-import { buildOpenHandsCompose, buildOpenHandsDockerRun } from "./docker.ts";
-import { DEFAULT_OPENHANDS_MODEL_MAP } from "./model-map.ts";
-
-function parseArgs(argv: string[]): Record<string, string> {
-  const out: Record<string, string> = {};
-  for (let i = 0; i < argv.length; i++) {
-    const arg = argv[i];
-    if (!arg.startsWith("--")) continue;
-    const key = arg.slice(2);
-    const next = argv[i + 1];
-    if (next !== undefined && !next.startsWith("--")) {
-      out[key] = next;
-      i++;
-    } else {
-      out[key] = "true";
-    }
-  }
-  return out;
-}
-
-function requireArgs(args: Record<string, string>, names: string[]): void {
-  for (const name of names) {
-    if (!args[name]) {
-      console.error(`Missing required --${name}`);
-      process.exit(2);
-    }
-  }
-}
-
-const [cmd, ...rest] = process.argv.slice(2);
-const args = parseArgs(rest);
-
-switch (cmd) {
-  case "env": {
-    requireArgs(args, ["api-key", "model"]);
-    const env = buildOpenHandsEnv({
-      apiKey: args["api-key"],
-      model: args.model,
-      omnirouteUrl: args.url,
-      persistenceDir: args["persistence-dir"],
-      corsOrigins: args["cors-origins"]?.split(","),
-    });
-    process.stdout.write(serializeOpenHandsEnv(env));
-    break;
-  }
-  case "compose": {
-    requireArgs(args, ["api-key", "model"]);
-    process.stdout.write(
-      buildOpenHandsCompose({
-        apiKey: args["api-key"],
-        model: args.model,
-        omnirouteUrl: args.url,
-        persistenceDir: args["persistence-dir"] ?? ".openhands-state",
-        corsOrigins: args["cors-origins"]?.split(","),
-        sandboxBaseImage: args["sandbox-image"],
-      })
-    );
-    break;
-  }
-  case "docker-run": {
-    requireArgs(args, ["api-key", "model"]);
-    process.stdout.write(
-      buildOpenHandsDockerRun({
-        apiKey: args["api-key"],
-        model: args.model,
-        omnirouteUrl: args.url,
-        persistenceDir: args["persistence-dir"] ?? ".openhands-state",
-        corsOrigins: args["cors-origins"]?.split(","),
-        sandboxBaseImage: args["sandbox-image"],
-      })
-    );
-    break;
-  }
-  case "models": {
-    for (const [name, target] of Object.entries(DEFAULT_OPENHANDS_MODEL_MAP)) {
-      process.stdout.write(`${name}\t->\t${target}\n`);
-    }
-    break;
-  }
-  default:
-    console.error(
-      "Usage: omniroute-openhands <env|compose|docker-run|models> [options]\n" +
-        "Options:\n" +
-        "  --api-key <sk-...>         OmniRoute API key (required for env/compose/docker-run)\n" +
-        "  --model <name>             OpenHands model name or OmniRoute combo\n" +
-        "  --url <base>               OmniRoute URL (default http://localhost:20128)\n" +
-        "  --persistence-dir <path>   Host dir for conversation state\n" +
-        "  --cors-origins <a,b,...>   Allowed CORS origins\n" +
-        "  --sandbox-image <image>    OpenHands sandbox base image"
-    );
-    process.exit(1);
-}
--- a/@omniroute/openhands-plugin/src/docker.ts
+++ b/@omniroute/openhands-plugin/src/docker.ts
@@ -1,92 +0,0 @@
-/**
- * OpenHands agent-server Docker Compose generator for OmniRoute.
- *
- * Bakes in the integration fixes that were needed to run OpenHands against
- * OmniRoute reliably:
- *   - `privileged: true`       — Python 3.13 socket.socketpair() needs it under
- *                                Docker's default seccomp profile
- *   - `extra_hosts`            — host.docker.internal → host-gateway so the
- *                                sandbox can reach OmniRoute on the host
- *   - host volume for OH_PERSISTENCE_DIR so conversations survive `docker rm`
- *   - PERMITTED_CORS_ORIGINS   — allow the dashboard origin to hit agent-server
- */
-
-export interface OpenHandsDockerOptions {
-  /** Agent-server image (default: the official OpenHands runtime image). */
-  image?: string;
-  /** Container name (default: openhands-agent). */
-  containerName?: string;
-  /** Model name to pass via LLM_MODEL. */
-  model: string;
-  /** OmniRoute API key. */
-  apiKey: string;
-  /** OmniRoute base URL reachable from the sandbox (default http://localhost:20128). */
-  omnirouteUrl?: string;
-  /** Host directory for OH_PERSISTENCE_DIR (must match env.ts persistenceDir). */
-  persistenceDir: string;
-  /** CORS origins to permit. */
-  corsOrigins?: string[];
-  /** Sandbox base image (defaults to OpenHands default). */
-  sandboxBaseImage?: string;
-  /** Set true to use host networking instead of extra_hosts. */
-  hostNetwork?: boolean;
-}
-
-export function buildOpenHandsCompose(opts: OpenHandsDockerOptions): string {
-  const image = opts.image ?? "docker.all-hands.dev/all-hands-ai/openhands:latest";
-  const containerName = opts.containerName ?? "openhands-agent";
-  const omnirouteHost = (opts.omnirouteUrl ?? "http://localhost:20128").replace(/\/+$/, "");
-  const cors =
-    opts.corsOrigins && opts.corsOrigins.length > 0
-      ? opts.corsOrigins
-      : ["http://localhost:3000", "http://localhost:3001"];
-
-  const lines: string[] = [];
-  lines.push(`services:`);
-  lines.push(`  openhands:`);
-  lines.push(`    image: ${image}`);
-  lines.push(`    container_name: ${containerName}`);
-  lines.push(`    privileged: true`);
-  lines.push(`    environment:`);
-  lines.push(`      LLM_MODEL: "${opts.model}"`);
-  lines.push(`      LLM_BASE_URL: "${omnirouteHost}/v1"`);
-  lines.push(`      LLM_API_KEY: "${opts.apiKey}"`);
-  lines.push(`      OH_PERSISTENCE_DIR: "/opt/.openhands-state"`);
-  lines.push(`      PERMITTED_CORS_ORIGINS: "${cors.join(",")}"`);
-  if (opts.sandboxBaseImage) {
-    lines.push(`      SANDBOX_BASE_IMAGE: "${opts.sandboxBaseImage}"`);
-  }
-  lines.push(`    volumes:`);
-  lines.push(`      - ${opts.persistenceDir}:/opt/.openhands-state`);
-  lines.push(`    extra_hosts:`);
-  lines.push(`      - "host.docker.internal:host-gateway"`);
-
-  return lines.join("\n") + "\n";
-}
-
-/**
- * docker run equivalent of {@link buildOpenHandsCompose} — returns the full
- * `docker run` command line.
- */
-export function buildOpenHandsDockerRun(opts: OpenHandsDockerOptions): string {
-  const image = opts.image ?? "docker.all-hands.dev/all-hands-ai/openhands:latest";
-  const omnirouteHost = (opts.omnirouteUrl ?? "http://localhost:20128").replace(/\/+$/, "");
-  const cors =
-    opts.corsOrigins && opts.corsOrigins.length > 0
-      ? opts.corsOrigins
-      : ["http://localhost:3000", "http://localhost:3001"];
-
-  const parts = [
-    "docker run",
-    "--privileged",
-    "--add-host host.docker.internal:host-gateway",
-    `-e LLM_MODEL="${opts.model}"`,
-    `-e LLM_BASE_URL="${omnirouteHost}/v1"`,
-    `-e LLM_API_KEY="${opts.apiKey}"`,
-    `-e OH_PERSISTENCE_DIR=/opt/.openhands-state`,
-    `-e PERMITTED_CORS_ORIGINS="${cors.join(",")}"`,
-    `-v "${opts.persistenceDir}:/opt/.openhands-state"`,
-    image,
-  ];
-  return parts.join(" ") + "\n";
-}
--- a/@omniroute/openhands-plugin/src/env.ts
+++ b/@omniroute/openhands-plugin/src/env.ts
@@ -1,63 +0,0 @@
-/**
- * OpenHands `.env` generator for the OmniRoute AI Gateway.
- *
- * Produces the OpenHands environment that points an OpenHands agent-server at
- * a running OmniRoute instance and fixes the integration gotchas found in the
- * field:
- *   - LLM_MODEL   — OpenHands-friendly model name → OmniRoute model/combo
- *   - LLM_BASE_URL — OmniRoute OpenAI-compatible endpoint
- *   - LLM_API_KEY  — OmniRoute key (sk-...)
- *   - OH_PERSISTENCE_DIR — host-mounted SQLite/conversation persistence
- *   - PERMITTED_CORS_ORIGINS — allow the dashboard origin to reach agent-server
- */
-
-export interface OpenHandsEnvOptions {
-  /** OmniRoute base URL as seen from the agent-server (default localhost:20128). */
-  omnirouteUrl?: string;
-  /** OmniRoute API key (sk-...). */
-  apiKey: string;
-  /** OpenHands model name (e.g. "deepseek-chat") or OmniRoute combo/model. */
-  model: string;
-  /** Host directory for OH_PERSISTENCE_DIR (default: current dir + .openhands-state). */
-  persistenceDir?: string;
-  /** CORS origins that must reach the agent-server (default dashboard origin + localhost). */
-  corsOrigins?: string[];
-  /** Optional OpenHands sandbox base image. */
-  sandboxBaseImage?: string;
-}
-
-export function buildOpenHandsEnv(opts: OpenHandsEnvOptions): Record<string, string> {
-  const omnirouteHost = (opts.omnirouteUrl ?? "http://localhost:20128").replace(/\/+$/, "");
-  const persistence = opts.persistenceDir ?? `${process.cwd()}/.openhands-state`;
-  const cors =
-    opts.corsOrigins && opts.corsOrigins.length > 0
-      ? opts.corsOrigins
-      : ["http://localhost:3000", "http://localhost:3001"];
-
-  const env: Record<string, string> = {
-    LLM_MODEL: opts.model,
-    LLM_BASE_URL: `${omnirouteHost}/v1`,
-    LLM_API_KEY: opts.apiKey,
-    OH_PERSISTENCE_DIR: persistence,
-    PERMITTED_CORS_ORIGINS: cors.join(","),
-  };
-
-  if (opts.sandboxBaseImage) {
-    env.SANDBOX_BASE_IMAGE = opts.sandboxBaseImage;
-  }
-
-  return env;
-}
-
-/**
- * Serialize the env record to `.env` file content (KEY=VALUE lines).
- * Values are not quoted unless they contain whitespace or `#`.
- */
-export function serializeOpenHandsEnv(env: Record<string, string>): string {
-  const lines: string[] = [];
-  for (const [key, value] of Object.entries(env)) {
-    const needsQuotes = /[\s#]/.test(value);
-    lines.push(needsQuotes ? `${key}="${value}"` : `${key}=${value}`);
-  }
-  return lines.join("\n") + "\n";
-}
--- a/@omniroute/openhands-plugin/src/index.ts
+++ b/@omniroute/openhands-plugin/src/index.ts
@@ -1,18 +0,0 @@
-/**
- * @omniroute/openhands-plugin — OpenHands integration for the OmniRoute AI Gateway.
- *
- * Generates the OpenHands environment and Docker Compose / docker run config
- * that wires an OpenHands agent-server to a running OmniRoute instance:
- * model mapping, sandbox privileges, host-gateway networking, persistent
- * conversation state and CORS.
- */
-export { buildOpenHandsEnv, serializeOpenHandsEnv } from "./env.ts";
-export type { OpenHandsEnvOptions } from "./env.ts";
-export { buildOpenHandsCompose, buildOpenHandsDockerRun } from "./docker.ts";
-export type { OpenHandsDockerOptions } from "./docker.ts";
-export {
-  DEFAULT_OPENHANDS_MODEL_MAP,
-  resolveOpenHandsModel,
-  buildOpenHandsModel,
-} from "./model-map.ts";
-export type { OpenHandsModelMap } from "./model-map.ts";
--- a/@omniroute/openhands-plugin/src/model-map.ts
+++ b/@omniroute/openhands-plugin/src/model-map.ts
@@ -1,64 +0,0 @@
-/**
- * OpenHands → OmniRoute model mapping.
- *
- * OpenHands sends `model: "<LLM_MODEL>"` and expects the OpenAI-compatible
- * endpoint to accept that exact string. OmniRoute uses provider-prefixed
- * model IDs (`ds/deepseek-v4-flash`) and combo names. This module maps
- * common OpenHands-friendly names to the OmniRoute model/combo they should
- * resolve to, and back-fills the `LLM_MODEL` value for OpenHands.
- */
-
-export interface OpenHandsModelMap {
-  /** OpenHands-friendly model name (e.g. "deepseek-chat") */
-  [openHandsName: string]: string;
-}
-
-/**
- * Default mapping for the model names OpenHands and the broader ecosystem
- * commonly send. Values are OmniRoute model IDs or combo names. Extend or
- * override via {@link resolveOpenHandsModel}.
- */
-export const DEFAULT_OPENHANDS_MODEL_MAP: OpenHandsModelMap = Object.freeze({
-  // DeepSeek
-  "deepseek-chat": "ds/deepseek-v4-flash",
-  "deepseek-reasoner": "ds/deepseek-v4-pro",
-  // Claude / Anthropic
-  "claude-sonnet-4.5": "anthropic/claude-sonnet-4.5",
-  "claude-opus-4.1": "anthropic/claude-opus-4.1",
-  "claude-haiku-4.5": "anthropic/claude-haiku-4.5",
-  // GPT / OpenAI
-  "gpt-4o": "openai/gpt-4o",
-  "gpt-4o-mini": "openai/gpt-4o-mini",
-  "gpt-5": "openai/gpt-5",
-  // Gemini
-  "gemini-2.5-flash": "gemini/gemini-2.5-flash",
-  "gemini-2.5-pro": "gemini/gemini-2.5-pro",
-  // GLM / Z.AI (NVIDIA NIM free endpoint)
-  "glm-5.2": "nvidia/z-ai/glm-5.2",
-});
-
-/**
- * Resolve the OmniRoute model ID for an OpenHands-friendly model name.
- * Returns the input unchanged when no mapping exists (OmniRoute will try to
- * resolve it as a literal model/combo).
- */
-export function resolveOpenHandsModel(
-  openHandsModel: string,
-  map: OpenHandsModelMap = DEFAULT_OPENHANDS_MODEL_MAP
-): string {
-  if (!openHandsModel) return openHandsModel;
-  const mapped = map[openHandsModel];
-  return mapped ?? openHandsModel;
-}
-
-/**
- * Build the `LLM_MODEL` value for OpenHands from an OmniRoute model ID/combo.
- *
- * OpenHands only surfaces the literal `LLM_MODEL` string in its UI, so for
- * OmniRoute combos (e.g. "vivanta-core") that's already the right value.
- * For provider-prefixed IDs, we return them as-is — the OmniRoute Model
- * Alias Resolver accepts both the raw ID and aliases on the `/v1` endpoint.
- */
-export function buildOpenHandsModel(omnirouteModelOrCombo: string): string {
-  return omnirouteModelOrCombo;
-}
--- a/@omniroute/openhands-plugin/tests/env.test.ts
+++ b/@omniroute/openhands-plugin/tests/env.test.ts
@@ -1,76 +0,0 @@
-import test from "node:test";
-import assert from "node:assert/strict";
-import { buildOpenHandsEnv, serializeOpenHandsEnv } from "../src/env.ts";
-import { resolveOpenHandsModel, buildOpenHandsModel } from "../src/model-map.ts";
-import { buildOpenHandsCompose, buildOpenHandsDockerRun } from "../src/docker.ts";
-
-test("buildOpenHandsEnv produces LLM vars pointing at OmniRoute", () => {
-  const env = buildOpenHandsEnv({
-    apiKey: "sk-test-123",
-    model: "deepseek-chat",
-    omnirouteUrl: "http://192.168.3.106:20128",
-    persistenceDir: "/opt/state",
-    corsOrigins: ["http://100.73.44.17:3000"],
-  });
-  assert.equal(env.LLM_MODEL, "deepseek-chat");
-  assert.equal(env.LLM_BASE_URL, "http://192.168.3.106:20128/v1");
-  assert.equal(env.LLM_API_KEY, "sk-test-123");
-  assert.equal(env.OH_PERSISTENCE_DIR, "/opt/state");
-  assert.equal(env.PERMITTED_CORS_ORIGINS, "http://100.73.44.17:3000");
-});
-
-test("serializeOpenHandsEnv quotes values with whitespace/#", () => {
-  const out = serializeOpenHandsEnv({ LLM_MODEL: "deepseek-chat", LLM_BASE_URL: "http://localhost:20128/v1" });
-  const lines = out.trim().split("\n");
-  assert.ok(lines.some((l) => l.startsWith("LLM_MODEL=deepseek-chat")));
-  assert.ok(lines.some((l) => l.startsWith("LLM_BASE_URL=http://localhost:20128/v1")));
-});
-
-test("resolveOpenHandsModel maps known names to OmniRoute IDs", () => {
-  assert.equal(resolveOpenHandsModel("deepseek-chat"), "ds/deepseek-v4-flash");
-  assert.equal(resolveOpenHandsModel("glm-5.2"), "nvidia/z-ai/glm-5.2");
-  assert.equal(resolveOpenHandsModel("gpt-4o"), "openai/gpt-4o");
-});
-
-test("resolveOpenHandsModel passes unknown names through unchanged", () => {
-  assert.equal(resolveOpenHandsModel("vivanta-core"), "vivanta-core");
-  assert.equal(resolveOpenHandsModel(""), "");
-});
-
-test("resolveOpenHandsModel accepts custom map overrides", () => {
-  const custom = { "my-alias": "nvidia/z-ai/glm-5.2" };
-  assert.equal(resolveOpenHandsModel("my-alias", custom), "nvidia/z-ai/glm-5.2");
-  assert.equal(resolveOpenHandsModel("deepseek-chat", custom), "deepseek-chat");
-});
-
-test("buildOpenHandsModel passes combo names through", () => {
-  assert.equal(buildOpenHandsModel("vivanta-core"), "vivanta-core");
-  assert.equal(buildOpenHandsModel("ds/deepseek-v4-flash"), "ds/deepseek-v4-flash");
-});
-
-test("buildOpenHandsCompose includes privileged, extra_hosts, volume, CORS", () => {
-  const compose = buildOpenHandsCompose({
-    apiKey: "sk-x",
-    model: "deepseek-chat",
-    persistenceDir: "/Users/me/.openhands-state",
-    corsOrigins: ["http://localhost:3000"],
-  });
-  assert.ok(compose.includes("privileged: true"), "privileged present");
-  assert.ok(compose.includes("host.docker.internal:host-gateway"), "host-gateway present");
-  assert.ok(compose.includes("/Users/me/.openhands-state"), "persistence volume present");
-  assert.ok(compose.includes("LLM_BASE_URL: \"http://localhost:20128/v1\""), "base url present");
-  assert.ok(compose.includes("PERMITTED_CORS_ORIGINS: \"http://localhost:3000\""), "cors present");
-});
-
-test("buildOpenHandsDockerRun produces a runnable docker command", () => {
-  const run = buildOpenHandsDockerRun({
-    apiKey: "sk-x",
-    model: "glm-5.2",
-    persistenceDir: "/opt/state",
-  });
-  assert.ok(run.startsWith("docker run"));
-  assert.ok(run.includes("--privileged"));
-  assert.ok(run.includes("--add-host host.docker.internal:host-gateway"));
-  assert.ok(run.includes("LLM_MODEL=\"glm-5.2\""));
-  assert.ok(run.includes("LLM_BASE_URL=\"http://localhost:20128/v1\""));
-});
--- a/@omniroute/openhands-plugin/tsconfig.json
+++ b/@omniroute/openhands-plugin/tsconfig.json
@@ -1,22 +0,0 @@
-{
-  "compilerOptions": {
-    "target": "ES2022",
-    "module": "ESNext",
-    "moduleResolution": "Bundler",
-    "lib": ["ES2022"],
-    "types": ["node"],
-    "ignoreDeprecations": "6.0",
-    "strict": true,
-    "esModuleInterop": true,
-    "skipLibCheck": true,
-    "allowImportingTsExtensions": true,
-    "declaration": true,
-    "isolatedModules": true,
-    "forceConsistentCasingInFileNames": true,
-    "noUncheckedIndexedAccess": false,
-    "outDir": "dist",
-    "rootDir": "src"
-  },
-  "include": ["src/**/*.ts"],
-  "exclude": ["dist", "node_modules", "tests"]
-}
--- a/@omniroute/openhands-plugin/tsup.config.ts
+++ b/@omniroute/openhands-plugin/tsup.config.ts
@@ -1,15 +0,0 @@
-import { defineConfig } from "tsup";
-
-export default defineConfig({
-  entry: ["src/index.ts", "src/cli.ts"],
-  format: ["esm"],
-  dts: true,
-  clean: true,
-  sourcemap: false,
-  splitting: false,
-  treeshake: false,
-  target: "node18",
-  outDir: "dist",
-  minify: false,
-  cjsInterop: false,
-});
--- a/5
+++ b/5
@@ -236,11 +236,6 @@ FROM runner-base AS runner-cli
 # runner-base runs.
 USER root

-# The CLI image can use the internal ChatGPT Web (Codex) Chromium sidecar over
-# CDP without installing a second browser in this container.
-COPY --from=builder /app/node_modules/playwright-core ./node_modules/playwright-core
-COPY --from=builder /app/node_modules/playwright ./node_modules/playwright
-
 # Install system dependencies required by openclaw (git+ssh references).
 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 \
--- a/THIRD_PARTY_NOTICES.md
+++ b/THIRD_PARTY_NOTICES.md
@@ -1,26 +0,0 @@
-# Third-Party Notices
-
-## codex-chatgpt-web
-
-Parts of `open-sse/vendor/codex-chatgpt-web/` are adapted from
-[`miuuyy/codex-chatgpt-web`](https://github.com/miuuyy/codex-chatgpt-web), commit
-`55592fca0ba19a27f1b769cec8fff61ff340a785`.
-
-MIT License
-
-Copyright (c) 2026 codex-chatgpt-web contributors
-
-Permission is hereby granted, free of charge, to any person obtaining a copy of this software and
-associated documentation files (the "Software"), to deal in the Software without restriction,
-including without limitation the rights to use, copy, modify, merge, publish, distribute,
-sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all copies or substantial
-portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT
-NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
-NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
-DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT
-OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
--- a/bin/chatgpt-web-codex-mcp.mjs
+++ b/bin/chatgpt-web-codex-mcp.mjs
@@ -1,56 +0,0 @@
-#!/usr/bin/env node
-
-import { existsSync } from "node:fs";
-import { dirname, join } from "node:path";
-import { fileURLToPath, pathToFileURL } from "node:url";
-
-const here = dirname(fileURLToPath(import.meta.url));
-const root = join(here, "..");
-
-export function resolveChatGptWebCodexMcpEntry(rootDir = root, exists = existsSync) {
-  const candidates = [
-    join(
-      rootDir,
-      "dist",
-      "open-sse",
-      "vendor",
-      "codex-chatgpt-web",
-      "adapters",
-      "chatgpt-web",
-      "mcp-server.js"
-    ),
-    join(
-      rootDir,
-      "open-sse",
-      "vendor",
-      "codex-chatgpt-web",
-      "adapters",
-      "chatgpt-web",
-      "mcp-server.ts"
-    ),
-  ];
-  return candidates.find((candidate) => exists(candidate)) ?? null;
-}
-
-export async function startChatGptWebCodexMcp(args = process.argv.slice(2), rootDir = root) {
-  const socketIndex = args.indexOf("--broker-socket");
-  const brokerSocketPath = socketIndex >= 0 ? args[socketIndex + 1] : undefined;
-  if (!brokerSocketPath) throw new Error("--broker-socket is required");
-  const entry = resolveChatGptWebCodexMcpEntry(rootDir);
-  if (!entry) throw new Error("ChatGPT Web (Codex) MCP entrypoint was not found");
-  if (entry.endsWith(".ts")) {
-    const { register } = await import("node:module");
-    register("tsx/esm", pathToFileURL(`${rootDir}/`));
-  }
-  const module = await import(pathToFileURL(entry).href);
-  await module.runChatGptMcpServer({ brokerSocketPath });
-}
-
-if (process.argv[1] && fileURLToPath(import.meta.url) === process.argv[1]) {
-  startChatGptWebCodexMcp().catch((error) => {
-    console.error(
-      `ChatGPT Web (Codex) MCP konnte nicht gestartet werden: ${error?.message || error}`
-    );
-    process.exit(1);
-  });
-}
--- a/changelog.d/fixes/8909-vertex-claude-rawpredict-streaming.md
+++ b/changelog.d/fixes/8909-vertex-claude-rawpredict-streaming.md
@@ -1 +0,0 @@
- **fix(executors):** Vertex AI now routes Claude models through the native Anthropic `rawPredict` endpoint instead of the generic OpenAI-compatible partner endpoint, and synthesizes a real streaming response so Claude-via-Vertex works with `stream: true` ([#8909](https://github.com/diegosouzapw/OmniRoute/pull/8909)) — thanks @wgordon17
--- a/changelog.d/fixes/9006-vertex-claude-catalog-dispatch.md
+++ b/changelog.d/fixes/9006-vertex-claude-catalog-dispatch.md
@@ -1,14 +0,0 @@
- **fix(sse):** Claude reasoning-effort suffix ids (`-high`/`-low`/`-medium`/`-xhigh`) now strip
-  correctly on any provider serving a real Claude model, not just the direct Anthropic provider
-  ([#9006](https://github.com/diegosouzapw/OmniRoute/pull/9006))
- **fix(sse):** the no-thinking (`no-think/`) catalog variant's provider-qualification bug — which
-  made it unusable outside the direct provider, both in the discovery catalog and the dashboard
-  playground — is fixed ([#9006](https://github.com/diegosouzapw/OmniRoute/pull/9006))
- **fix(sse):** a single unrecognized model id on a Vertex connection no longer cools down every
-  other model on that connection for 2 minutes — Vertex 404s are now scoped to a per-model
-  lockout via `passthroughModels` instead of a connection-wide cooldown
-  ([#9006](https://github.com/diegosouzapw/OmniRoute/pull/9006))
- **fix(sse):** Vertex `PERMISSION_DENIED` 403s are now disambiguated using Google's own
-  documented error format — a genuinely connection-wide cause (API disabled, project-level IAM
-  denial) still cools the whole connection, while a model-specific denial locks out only that
-  model ([#9006](https://github.com/diegosouzapw/OmniRoute/pull/9006))
--- a/changelog.d/fixes/9013-model-param-filter-save.md
+++ b/changelog.d/fixes/9013-model-param-filter-save.md
@@ -1 +0,0 @@
- **fix(dashboard):** model-level allowed/blocked param edits now persist when the compatibility popover is closed by clicking outside, and a failed save no longer clears the edit or reports success ([#9013](https://github.com/diegosouzapw/OmniRoute/pull/9013))
--- a/changelog.d/fixes/9277-fix.plan.md
+++ b/changelog.d/fixes/9277-fix.plan.md
@@ -0,0 +1 @@
+- fix(qoder): include actionable CLI_QODER_BIN hint in connection test when qodercli is not found (#9277)
--- a/compression-core/.gitignore
+++ b/compression-core/.gitignore
@@ -1,2 +0,0 @@
-/target
-Cargo.lock
--- a/compression-core/Cargo.toml
+++ b/compression-core/Cargo.toml
@@ -1,27 +0,0 @@
-[workspace]
-resolver = "2"
-members = [
-    "crates/core-api",
-    "crates/tokenizer",
-    "crates/tests",
-    "crates/bench",
-    "crates/ffi",
-]
-
-[workspace.package]
-version = "0.1.0"
-edition = "2021"
-license = "MIT"
-repository = "https://github.com/Egorich-print/OmniRoute"
-
-[workspace.dependencies]
-core-api = { path = "crates/core-api" }
-tokenizer = { path = "crates/tokenizer" }
-tiktoken-rs = "0.6"
-serde = { version = "1", features = ["derive"] }
-serde_json = "1"
-thiserror = "2"
-
-[profile.release]
-lto = true
-codegen-units = 1
--- a/compression-core/README.md
+++ b/compression-core/README.md
@@ -1,61 +0,0 @@
-# compression-core
-
-Standalone Rust core for AI context optimization — tokenization, compression,
-hashing, translation primitives. Independent OSS library usable by OmniRoute,
-OpenCode, Cline, Roo, and any AI proxy. No OmniRoute imports anywhere.
-
-## Layout
-
-```
-compression-core/
-├── Cargo.toml            # workspace
-├── crates/
-│   ├── core-api/         # stable public API (traits + types) — no host deps
-│   ├── tokenizer/        # tiktoken (cl100k_base, o200k_base) — PORTED
-│   ├── tests/            # golden tests against fixtures/expected/
-│   ├── bench/            # criterion benchmarks
-│   └── ffi/              # N-API adapter (integration phase)
-├── fixtures/
-│   ├── tokenizer/        # JS-generated token counts (13 samples)
-│   └── expected/         # manifests
-└── scripts/
-    ├── generate-fixtures.ts  # JS reference output (source of truth)
-    └── verify-golden.ts      # regen + cargo test
-```
-
-## Porting order (per design)
-
-1. tiktoken (done — golden 100%)
-2. ionizer
-3. headroom
-4. caveman
-5. RTK (last — biggest, requires proven harness)
-
-## Golden pipeline
-
-```text
-fixtures → JS implementation → expected.json → Rust → assert_eq!
-```
-
-`node scripts/verify-golden.ts` regenerates fixtures from the current JS code
-and runs `cargo test -p compression-tests`. Until 100% match, JS stays in prod.
-
-## Measured baseline
-
-| Impl | Input | Cost |
-|---|---|---|
-| JS js-tiktoken (cl100k) | 230K chars | 37.9 ms |
-| Rust tiktoken-rs (cl100k) | ~440K chars | 21.4 ms |
-
-Per-char Rust is ~3x faster; golden output is byte-identical on all fixtures.
-
-## Status
-
- [x] workspace + stable API (`core-api`)
- [x] tokenizer port + golden tests (100% match)
- [x] bench harness (criterion)
- [ ] ionizer
- [ ] headroom
- [ ] caveman
- [ ] RTK
- [ ] N-API adapter
--- a/compression-core/crates/bench/Cargo.toml
+++ b/compression-core/crates/bench/Cargo.toml
@@ -1,17 +0,0 @@
-[package]
-name = "compression-bench"
-version.workspace = true
-edition.workspace = true
-license.workspace = true
-publish = false
-
-[dependencies]
-core-api = { workspace = true }
-tokenizer = { workspace = true }
-
-[dev-dependencies]
-criterion = "0.5"
-
-[[bench]]
-name = "tokenizer"
-harness = false
--- a/compression-core/crates/bench/benches/tokenizer.rs
+++ b/compression-core/crates/bench/benches/tokenizer.rs
@@ -1,19 +0,0 @@
-//! Criterion bench for the tokenizer. Baseline target: < 5 ms per 57K tokens
-//! (JS js-tiktoken measures ~38 ms on the same input).
-
-use core_api::TokenCounter;
-use criterion::{criterion_group, criterion_main, Criterion};
-use tokenizer::TiktokenCounter;
-
-fn bench_tokenizer(c: &mut Criterion) {
-    let counter = TiktokenCounter::default();
-    // ~230K chars ≈ 57K cl100k tokens (mirrors the measured JS baseline).
-    let text = "Hello world! This is a test of tokenization performance. \
-                The quick brown fox jumps over the lazy dog. "
-        .repeat(4000);
-
-    c.bench_function("cl100k_57k_tokens", |b| b.iter(|| counter.count(&text)));
-}
-
-criterion_group!(benches, bench_tokenizer);
-criterion_main!(benches);
--- a/compression-core/crates/core-api/Cargo.toml
+++ b/compression-core/crates/core-api/Cargo.toml
@@ -1,10 +0,0 @@
-[package]
-name = "core-api"
-version.workspace = true
-edition.workspace = true
-license.workspace = true
-
-[dependencies]
-serde = { workspace = true }
-serde_json = { workspace = true }
-thiserror = { workspace = true }
--- a/compression-core/crates/core-api/src/lib.rs
+++ b/compression-core/crates/core-api/src/lib.rs
@@ -1,76 +0,0 @@
-//! Stable public API of the compression core.
-//!
-//! This crate is intentionally free of any OmniRoute-specific types.
-//! It defines the contracts that every adapter (N-API, sidecar, CLI)
-//! implements, so algorithms stay independent of the host project.
-
-use serde::{Deserialize, Serialize};
-
-/// Role of a message in a conversation.
-#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
-#[serde(rename_all = "lowercase")]
-pub enum Role {
-    System,
-    User,
-    Assistant,
-    Tool,
-}
-
-/// One chat message. Field-compatible with OpenAI `messages[]` entries.
-#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
-pub struct Message {
-    pub role: Role,
-    #[serde(skip_serializing_if = "Option::is_none")]
-    pub content: Option<String>,
-    #[serde(skip_serializing_if = "Option::is_none")]
-    pub name: Option<String>,
-    #[serde(skip_serializing_if = "Option::is_none")]
-    pub tool_call_id: Option<String>,
-}
-
-/// Tokenizer encodings supported by the core.
-#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
-pub enum Encoding {
-    #[serde(rename = "cl100k_base")]
-    Cl100kBase,
-    #[serde(rename = "o200k_base")]
-    O200kBase,
-}
-
-/// Configuration for a compression pass.
-#[derive(Debug, Clone, PartialEq, Serialize, Deserialize, Default)]
-pub struct CompressionConfig {
-    /// Target token budget for the compressed messages.
-    pub budget_tokens: Option<u64>,
-    /// Engine stack priority hint (rtk=10, ionizer=13, headroom=15, ...).
-    pub stack_priority: Option<u32>,
-}
-
-/// Outcome of a compression pass.
-#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
-pub struct CompressionResult {
-    pub messages: Vec<Message>,
-    pub compressed: bool,
-    pub stats: Option<CompressionStats>,
-}
-
-#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
-pub struct CompressionStats {
-    pub saved_tokens: Option<u64>,
-    pub input_tokens: Option<u64>,
-    pub output_tokens: Option<u64>,
-}
-
-/// A token counter. Pure, stateless, thread-safe.
-pub trait TokenCounter {
-    fn count(&self, text: &str) -> usize;
-}
-
-/// A compressor. Pure, deterministic, stateless per call.
-pub trait Compressor {
-    fn compress(
-        &self,
-        messages: &[Message],
-        config: &CompressionConfig,
-    ) -> CompressionResult;
-}
--- a/compression-core/crates/ffi/Cargo.toml
+++ b/compression-core/crates/ffi/Cargo.toml
@@ -1,13 +0,0 @@
-[package]
-name = "compression-ffi"
-version.workspace = true
-edition.workspace = true
-license.workspace = true
-publish = false
-
-[dependencies]
-core-api = { workspace = true }
-tokenizer = { workspace = true }
-
-# napi-rs bindings are added in the integration phase. This crate exists to
-# keep the N-API adapter out of the algorithm crates.
--- a/compression-core/crates/ffi/src/lib.rs
+++ b/compression-core/crates/ffi/src/lib.rs
@@ -1,8 +0,0 @@
-//! N-API binding crate (integration phase).
-//!
-//! This crate is intentionally empty until the N-API phase. It will expose
-//! `count_tokens` / `compress` over napi-rs using the core-api traits, so the
-//! algorithms in `tokenizer` and the future `compression` crates stay free of
-//! any Node bindings.
-
-pub use core_api;
--- a/compression-core/crates/tests/Cargo.toml
+++ b/compression-core/crates/tests/Cargo.toml
@@ -1,11 +0,0 @@
-[package]
-name = "compression-tests"
-version.workspace = true
-edition.workspace = true
-license.workspace = true
-publish = false
-
-[dependencies]
-core-api = { workspace = true }
-tokenizer = { workspace = true }
-serde_json = { workspace = true }
--- a/compression-core/crates/tests/src/lib.rs
+++ b/compression-core/crates/tests/src/lib.rs
@@ -1,66 +0,0 @@
-//! Golden tests: run the Rust implementations against fixtures and compare
-//! byte-for-byte with the JS-produced `expected/` files.
-//!
-//! The `verify-golden.ts` script regenerates fixtures from the OmniRoute JS
-//! implementation. Until this crate passes 100% of golden fixtures, the JS
-//! implementation must NOT be replaced in production.
-
-use core_api::{Encoding, TokenCounter};
-use std::path::Path;
-use tokenizer::TiktokenCounter;
-
-const FIXTURES_DIR: &str = concat!(env!("CARGO_MANIFEST_DIR"), "/../../fixtures");
-
-fn fixture_path(relative: &str) -> String {
-    Path::new(FIXTURES_DIR).join(relative).to_string_lossy().into_owned()
-}
-
-#[test]
-fn tokenizer_golden_cl100k() {
-    let counter = TiktokenCounter::default();
-    let dir = fixture_path("tokenizer");
-    let entries = std::fs::read_dir(&dir).expect("fixtures/tokenizer must exist");
-    let mut checked = 0;
-    for entry in entries {
-        let path = entry.unwrap().path();
-        if path.extension().map(|e| e == "json").unwrap_or(false) {
-            let input: serde_json::Value =
-                serde_json::from_str(&std::fs::read_to_string(&path).unwrap()).unwrap();
-            let text = input["text"].as_str().unwrap();
-            let expected = input["cl100k_tokens"].as_u64().unwrap() as usize;
-            assert_eq!(
-                counter.count(text),
-                expected,
-                "cl100k mismatch on {}",
-                path.display()
-            );
-            checked += 1;
-        }
-    }
-    assert!(checked > 0, "no tokenizer fixtures found");
-}
-
-#[test]
-fn tokenizer_golden_o200k() {
-    let counter = TiktokenCounter::default();
-    let dir = fixture_path("tokenizer");
-    let entries = std::fs::read_dir(&dir).expect("fixtures/tokenizer must exist");
-    let mut checked = 0;
-    for entry in entries {
-        let path = entry.unwrap().path();
-        if path.extension().map(|e| e == "json").unwrap_or(false) {
-            let input: serde_json::Value =
-                serde_json::from_str(&std::fs::read_to_string(&path).unwrap()).unwrap();
-            let text = input["text"].as_str().unwrap();
-            let expected = input["o200k_tokens"].as_u64().unwrap() as usize;
-            assert_eq!(
-                counter.count_with_encoding(text, Encoding::O200kBase),
-                expected,
-                "o200k mismatch on {}",
-                path.display()
-            );
-            checked += 1;
-        }
-    }
-    assert!(checked > 0, "no tokenizer fixtures found");
-}
--- a/compression-core/crates/tokenizer/Cargo.toml
+++ b/compression-core/crates/tokenizer/Cargo.toml
@@ -1,13 +0,0 @@
-[package]
-name = "tokenizer"
-version.workspace = true
-edition.workspace = true
-license.workspace = true
-
-[dependencies]
-core-api = { workspace = true }
-tiktoken-rs = { workspace = true }
-anyhow = "1"
-
-[dev-dependencies]
-serde_json = { workspace = true }
--- a/compression-core/crates/tokenizer/src/lib.rs
+++ b/compression-core/crates/tokenizer/src/lib.rs
@@ -1,71 +0,0 @@
-//! Tiktoken token counter backed by `tiktoken-rs`.
-//!
-//! Port target: `src/shared/utils/tiktokenCounter.ts` in OmniRoute.
-//! Encodings: cl100k_base (default), o200k_base (Codex).
-
-use core_api::{Encoding, TokenCounter};
-use tiktoken_rs::tokenizer::Tokenizer;
-
-pub struct TiktokenCounter {
-    cl100k: tiktoken_rs::CoreBPE,
-    o200k: tiktoken_rs::CoreBPE,
-}
-
-impl TiktokenCounter {
-    pub fn new() -> Result<Self, anyhow::Error> {
-        let cl100k = tiktoken_rs::get_bpe_from_tokenizer(Tokenizer::Cl100kBase)?;
-        let o200k = tiktoken_rs::get_bpe_from_tokenizer(Tokenizer::O200kBase)?;
-        Ok(Self { cl100k, o200k })
-    }
-
-    pub fn count_with_encoding(&self, text: &str, encoding: Encoding) -> usize {
-        let bpe = match encoding {
-            Encoding::Cl100kBase => &self.cl100k,
-            Encoding::O200kBase => &self.o200k,
-        };
-        // CoreBPE::encode_with_special_tokens requires allocation; the
-        // plain encode is the closest equivalent to the JS byte-pair count.
-        bpe.encode_ordinary(text).len()
-    }
-}
-
-impl Default for TiktokenCounter {
-    fn default() -> Self {
-        Self::new().expect("tiktoken rank tables must load")
-    }
-}
-
-impl TokenCounter for TiktokenCounter {
-    fn count(&self, text: &str) -> usize {
-        self.count_with_encoding(text, Encoding::Cl100kBase)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn counts_known_tokens_cl100k() {
-        let counter = TiktokenCounter::default();
-        // "Hello world" is 2 tokens in cl100k_base.
-        assert_eq!(counter.count("Hello world"), 2);
-    }
-
-    #[test]
-    fn empty_string_is_zero() {
-        let counter = TiktokenCounter::default();
-        assert_eq!(counter.count(""), 0);
-    }
-
-    #[test]
-    fn o200k_differs_from_cl100k_on_emoji() {
-        let counter = TiktokenCounter::default();
-        let emoji = "🎉";
-        let cl100k = counter.count_with_encoding(emoji, Encoding::Cl100kBase);
-        let o200k = counter.count_with_encoding(emoji, Encoding::O200kBase);
-        // o200k has dedicated emoji tokens; counts may differ. Just assert both are > 0.
-        assert!(cl100k > 0);
-        assert!(o200k > 0);
-    }
-}
--- a/compression-core/fixtures/expected/tokenizer-manifest.json
+++ b/compression-core/fixtures/expected/tokenizer-manifest.json
@@ -1,80 +0,0 @@
-[
-  {
-    "id": "000",
-    "chars": 28,
-    "cl100k_tokens": 8,
-    "o200k_tokens": 8
-  },
-  {
-    "id": "001",
-    "chars": 44,
-    "cl100k_tokens": 10,
-    "o200k_tokens": 10
-  },
-  {
-    "id": "002",
-    "chars": 40,
-    "cl100k_tokens": 15,
-    "o200k_tokens": 12
-  },
-  {
-    "id": "003",
-    "chars": 38,
-    "cl100k_tokens": 15,
-    "o200k_tokens": 15
-  },
-  {
-    "id": "004",
-    "chars": 47,
-    "cl100k_tokens": 15,
-    "o200k_tokens": 15
-  },
-  {
-    "id": "005",
-    "chars": 100,
-    "cl100k_tokens": 100,
-    "o200k_tokens": 50
-  },
-  {
-    "id": "006",
-    "chars": 100,
-    "cl100k_tokens": 41,
-    "o200k_tokens": 23
-  },
-  {
-    "id": "007",
-    "chars": 10000,
-    "cl100k_tokens": 1250,
-    "o200k_tokens": 1250
-  },
-  {
-    "id": "008",
-    "chars": 1,
-    "cl100k_tokens": 1,
-    "o200k_tokens": 1
-  },
-  {
-    "id": "009",
-    "chars": 0,
-    "cl100k_tokens": 0,
-    "o200k_tokens": 0
-  },
-  {
-    "id": "010",
-    "chars": 49,
-    "cl100k_tokens": 18,
-    "o200k_tokens": 14
-  },
-  {
-    "id": "011",
-    "chars": 69,
-    "cl100k_tokens": 24,
-    "o200k_tokens": 24
-  },
-  {
-    "id": "012",
-    "chars": 405000,
-    "cl100k_tokens": 90001,
-    "o200k_tokens": 90001
-  }
-]
--- a/compression-core/fixtures/tokenizer/sample-000.json
+++ b/compression-core/fixtures/tokenizer/sample-000.json
@@ -1,6 +0,0 @@
-{
-  "id": "000",
-  "text": "Hello world! This is a test.",
-  "cl100k_tokens": 8,
-  "o200k_tokens": 8
-}
--- a/compression-core/fixtures/tokenizer/sample-001.json
+++ b/compression-core/fixtures/tokenizer/sample-001.json
@@ -1,6 +0,0 @@
-{
-  "id": "001",
-  "text": "The quick brown fox jumps over the lazy dog.",
-  "cl100k_tokens": 10,
-  "o200k_tokens": 10
-}
--- a/compression-core/fixtures/tokenizer/sample-002.json
+++ b/compression-core/fixtures/tokenizer/sample-002.json
@@ -1,6 +0,0 @@
-{
-  "id": "002",
-  "text": "🎉🎊 party time! emoji heavy sentence 🚀",
-  "cl100k_tokens": 15,
-  "o200k_tokens": 12
-}
--- a/compression-core/fixtures/tokenizer/sample-003.json
+++ b/compression-core/fixtures/tokenizer/sample-003.json
@@ -1,6 +0,0 @@
-{
-  "id": "003",
-  "text": "JSON:\n{\"name\":\"test\",\"values\":[1,2,3]}",
-  "cl100k_tokens": 15,
-  "o200k_tokens": 15
-}
--- a/compression-core/fixtures/tokenizer/sample-004.json
+++ b/compression-core/fixtures/tokenizer/sample-004.json
@@ -1,6 +0,0 @@
-{
-  "id": "004",
-  "text": "Code:\n```rust\nfn main() { println!(\"hi\"); }\n```",
-  "cl100k_tokens": 15,
-  "o200k_tokens": 15
-}
--- a/compression-core/fixtures/tokenizer/sample-005.json
+++ b/compression-core/fixtures/tokenizer/sample-005.json
@@ -1,6 +0,0 @@
-{
-  "id": "005",
-  "text": "😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀😀",
-  "cl100k_tokens": 100,
-  "o200k_tokens": 50
-}
--- a/compression-core/fixtures/tokenizer/sample-006.json
+++ b/compression-core/fixtures/tokenizer/sample-006.json
@@ -1,6 +0,0 @@
-{
-  "id": "006",
-  "text": "Поддерживается ли русский текст корректно? Проверяем длинное предложение с кириллицей и пунктуацией!",
-  "cl100k_tokens": 41,
-  "o200k_tokens": 23
-}
--- a/compression-core/fixtures/tokenizer/sample-007.json
+++ b/compression-core/fixtures/tokenizer/sample-007.json
--- a/compression-core/fixtures/tokenizer/sample-008.json
+++ b/compression-core/fixtures/tokenizer/sample-008.json
@@ -1,6 +0,0 @@
-{
-  "id": "008",
-  "text": "t",
-  "cl100k_tokens": 1,
-  "o200k_tokens": 1
-}
--- a/compression-core/fixtures/tokenizer/sample-009.json
+++ b/compression-core/fixtures/tokenizer/sample-009.json
@@ -1,6 +0,0 @@
-{
-  "id": "009",
-  "text": "",
-  "cl100k_tokens": 0,
-  "o200k_tokens": 0
-}
--- a/compression-core/fixtures/tokenizer/sample-010.json
+++ b/compression-core/fixtures/tokenizer/sample-010.json
@@ -1,6 +0,0 @@
-{
-  "id": "010",
-  "text": "Mixed 🎯 unicode 中文 한국어 + english + numbers 12345",
-  "cl100k_tokens": 18,
-  "o200k_tokens": 14
-}
--- a/compression-core/fixtures/tokenizer/sample-011.json
+++ b/compression-core/fixtures/tokenizer/sample-011.json
@@ -1,6 +0,0 @@
-{
-  "id": "011",
-  "text": "function foo(a,b){return a+b*2;}\n\nconst x = foo(1,2);\nconsole.log(x);",
-  "cl100k_tokens": 24,
-  "o200k_tokens": 24
-}
--- a/compression-core/fixtures/tokenizer/sample-012.json
+++ b/compression-core/fixtures/tokenizer/sample-012.json
--- a/compression-core/scripts/generate-fixtures.ts
+++ b/compression-core/scripts/generate-fixtures.ts
@@ -1,77 +0,0 @@
-#!/usr/bin/env node
-/**
- * Generates golden fixtures for compression-core from the OmniRoute JS
- * implementation. Every fixture records: input text + expected token counts
- * (cl100k / o200k) computed by the JS tokenizer.
- *
- * Usage:  node --import tsx/esm scripts/generate-fixtures.ts
- * Output: fixtures/tokenizer/*.json, fixtures/conversations/*.json
- *
- * The Rust side (crates/tests) reads these and asserts equality. Until 100%
- * of fixtures pass, the JS implementation must not be replaced.
- */
-import { countTextTokens } from "../../src/shared/utils/tiktokenCounter.ts";
-import { mkdirSync, writeFileSync } from "node:fs";
-import { dirname, join } from "node:path";
-import { fileURLToPath } from "node:url";
-
-const HERE = dirname(fileURLToPath(import.meta.url));
-const ROOT = join(HERE, "..");
-const TOKENIZER_DIR = join(ROOT, "fixtures", "tokenizer");
-const EXPECTED_DIR = join(ROOT, "fixtures", "expected");
-
-mkdirSync(TOKENIZER_DIR, { recursive: true });
-mkdirSync(EXPECTED_DIR, { recursive: true });
-
-const SAMPLES = [
-  "Hello world! This is a test.",
-  "The quick brown fox jumps over the lazy dog.",
-  "🎉🎊 party time! emoji heavy sentence 🚀",
-  "JSON:\n{\"name\":\"test\",\"values\":[1,2,3]}",
-  "Code:\n```rust\nfn main() { println!(\"hi\"); }\n```",
-  "😀".repeat(50),
-  "Поддерживается ли русский текст корректно? Проверяем длинное предложение с кириллицей и пунктуацией!",
-  "a".repeat(10000),
-  "t".repeat(1),
-  "",
-  "Mixed 🎯 unicode 中文 한국어 + english + numbers 12345",
-  "function foo(a,b){return a+b*2;}\n\nconst x = foo(1,2);\nconsole.log(x);",
-];
-
-// A longer realistic conversation-style text (~230K chars) to mirror the
-// measured baseline and to stress the counter on large inputs.
-const LONG = ("The quick brown fox jumps over the lazy dog. ").repeat(9000);
-SAMPLES.push(LONG);
-
-const cl100k = (t) => countTextTokens(t);
-const o200k = (t) => countTextTokens(t, { provider: "codex", model: "codex/gpt-5.5" });
-
-let count = 0;
-for (const [idx, text] of SAMPLES.entries()) {
-  const id = String(idx).padStart(3, "0");
-  const record = {
-    id,
-    text,
-    cl100k_tokens: cl100k(text),
-    o200k_tokens: o200k(text),
-  };
-  writeFileSync(join(TOKENIZER_DIR, `sample-${id}.json`), JSON.stringify(record, null, 2));
-  count++;
-}
-
-// Also emit a combined manifest for quick scanning.
-writeFileSync(
-  join(EXPECTED_DIR, "tokenizer-manifest.json"),
-  JSON.stringify(
-    SAMPLES.map((t, idx) => ({
-      id: String(idx).padStart(3, "0"),
-      chars: t.length,
-      cl100k_tokens: cl100k(t),
-      o200k_tokens: o200k(t),
-    })),
-    null,
-    2
-  )
-);
-
-console.log(`Generated ${count} tokenizer fixtures + manifest in fixtures/`);
--- a/compression-core/scripts/verify-golden.ts
+++ b/compression-core/scripts/verify-golden.ts
@@ -1,45 +0,0 @@
-#!/usr/bin/env node
-/**
- * Verifies golden equivalence between the JS implementation and the Rust
- * implementation.
- *
- * Rust side: runs `cargo test -p compression-tests` which asserts byte-level
- * equality against fixtures/expected/. This script:
- *   1. regenerates fixtures from the current JS implementation
- *   2. runs cargo tests
- *   3. reports pass/fail per fixture family
- *
- * Usage: node scripts/verify-golden.ts [--skip-generate]
- */
-import { spawnSync } from "node:child_process";
-import { fileURLToPath } from "node:url";
-import { dirname, join } from "node:path";
-
-const HERE = dirname(fileURLToPath(import.meta.url));
-const CORE_DIR = join(HERE, "..");
-
-const skipGenerate = process.argv.includes("--skip-generate");
-
-if (!skipGenerate) {
-  console.log("[verify-golden] regenerating fixtures from JS implementation...");
-  const gen = spawnSync("node", ["--import", "tsx/esm", "scripts/generate-fixtures.ts"], {
-    cwd: CORE_DIR,
-    stdio: "inherit",
-  });
-  if (gen.status !== 0) {
-    console.error("FAIL: fixture generation exited with", gen.status);
-    process.exit(1);
-  }
-}
-
-console.log("[verify-golden] running Rust golden tests...");
-const run = spawnSync("cargo", ["test", "-p", "compression-tests"], {
-  cwd: CORE_DIR,
-  stdio: "inherit",
-});
-if (run.status !== 0) {
-  console.error("FAIL: Rust golden tests exited with", run.status);
-  process.exit(1);
-}
-
-console.log("[verify-golden] ALL GOLDEN TESTS PASSED ✅");
--- a/config/quality/dashboard-typecheck-baseline.json
+++ b/config/quality/dashboard-typecheck-baseline.json
@@ -127,6 +127,12 @@
  "src/app/(dashboard)/dashboard/providers/[id]/components/ConnectionsListPanel.tsx": {
    "TS2322": 2
  },
+  "src/app/(dashboard)/dashboard/providers/[id]/components/CustomModelsSection.tsx": {
+    "TS2739": 1
+  },
+  "src/app/(dashboard)/dashboard/providers/[id]/components/ModelCompatPopover.tsx": {
+    "TS2304": 5
+  },
  "src/app/(dashboard)/dashboard/providers/[id]/components/ProviderModalsPanel.tsx": {
    "TS2322": 3,
    "TS2739": 1,
@@ -141,6 +147,9 @@
  "src/app/(dashboard)/dashboard/providers/[id]/components/ProviderPlaygroundPanel.tsx": {
    "TS2503": 1
  },
+  "src/app/(dashboard)/dashboard/providers/[id]/components/__tests__/phase1d.test.tsx": {
+    "TS2739": 2
+  },
  "src/app/(dashboard)/dashboard/providers/[id]/components/modals/EditConnectionModal.tsx": {
    "TS2322": 1
  },
--- a/config/quality/dependency-allowlist.json
+++ b/config/quality/dependency-allowlist.json
@@ -121,8 +121,6 @@
    "tailwind-merge",
    "tailwindcss",
    "tls-client-node",
-    "turndown",
-    "turndown-plugin-gfm",
    "tsup",
    "tsx",
    "type-coverage",
--- a/config/quality/file-size-baseline.json
+++ b/config/quality/file-size-baseline.json
@@ -1,5 +1,4 @@
 {
-  "_rebaseline_2026_07_30_9006_vertex_claude_catalog_dispatch": "PR #9006 (fix/vertex-claude-catalog-dispatch): three files, two causes. (1) src/sse/handlers/chat.ts 1845->1846 (+1): NOT this PR's own growth — this PR never touches chat.ts at all. Measured 1846 (split(\"\\n\").length) at this PR's own merge-base (before any of its 11 commits), so the drift was already inherited from already-merged PRs on release/v3.8.50 (fast-gates PR->release do not run check:file-size, same root cause as _rebaseline_2026_07_25_v3849_basered_filesize and _rebaseline_2026_07_02_5798_release_green) — no offending branch left to fix. (2) src/sse/services/auth.ts 2508->2512 (+4 net, after extraction — see below) and open-sse/handlers/chatCore.ts 5020->5023 (+3, comment-only): genuine own growth. auth.ts adds Vertex 403 PERMISSION_DENIED disambiguation (Google's google.rpc.ErrorInfo proto distinguishes a connection-wide cause — SERVICE_DISABLED, or IAM_PERMISSION_DENIED against a project-level resource — from a model-specific one scoped to a .../models/<id> resource), added mid-PR after a quality-gate reviewer flagged the plan's originally-accepted \"Vertex 403 always -> per-model lockout\" trade-off. The actual classification logic (~40 lines) was EXTRACTED into a new leaf module src/sse/services/vertexErrorClassifier.ts (mirrors the googApiKeyAuth.ts precedent, _rebaseline_2026_07_14_7034_goog_api_key), leaving only the irreducible call-site wiring in the frozen file: a 1-line import plus widening the existing #3027 per-model-403 guard condition. chatCore.ts's +3 is a pure comment expansion (no functional change) clarifying that the adjacent effort-suffix strip is no longer unconditional for every provider, requested by a separate quality-gate code-reviewer finding; not extractable (it's a comment). Auth.ts's disambiguation logic covered by 3 new test cases in tests/unit/vertex-passthrough-model-lockout.test.ts (SERVICE_DISABLED, IAM_PERMISSION_DENIED+model-resource, IAM_PERMISSION_DENIED+project-resource) plus a 4th regression test for a multi-detail-body correlation bug (reason and resource must be read from the SAME ErrorInfo detail, not independently regexed across the whole body) found by an adversarial quality-gate pass and fixed before merge.",
  "_rebaseline_2026_07_24_8470_hyperagent_sticky_thread": "PR #8470 (artickc, fix/hyperagent-tool-loop-thread-sticky) own growth: open-sse/executors/hyperagent.ts 936->1025 (wc -l; check-file-size.mjs counts via split(\"\\n\").length so the gate sees 937->1026, +89, crosses the 1000 cap). Fixes a real bug where a reverse-conversion proxy (text-Intent/JSON to Claude Code native tool_calls) rewrites assistant messages between agentic tool-loop turns, breaking HyperAgent's conversation-prefix fingerprint and cold-starting the thread mid tool-loop. Adds Anthropic tool_use/tool_result flattening to extractMessageText() plus a new rootUserFingerprint()/root-key lookup tier in resolveHyperAgentThreadBinding()/storeHyperAgentThreadAfterTurn() so the thread stays sticky across the tool loop. Cohesive additions inside the existing single-file executor; not extractable without splitting the executor mid-request-flow. Covered by tests/unit/executor-hyperagent.test.ts (19/19, +5 new cases for tool_result/tool_use flattening + root-key stickiness). Pre-merge review flagged a cross-conversation root-key collision risk (tracked in the PR's own mandatory pre-merge checklist, not yet addressed) — unrelated to this file-size ratchet, tracked separately by /fix-prs.",
  "_rebaseline_2026_07_25_8494_capability_filter_fail_closed": "PR #8494 (fix/capability-filters-fail-closed, #8488) own growth: open-sse/services/combo.ts 3640->3693 (+53) adds a fail-closed guard after filterTargetsByRequestCompatibility() — when every eligible target is excluded by request-capability filtering (vision/tools/etc) instead of quota/health, the combo now returns an explicit `capability_mismatch` 400 (describeCapabilityFilterExhaustion, imported from combo/comboStructure.ts) rather than silently falling through to a generic no-targets error, plus a `compatFilterFailOpen` escape hatch (combo config OR settings) mirrored at both the main/auto and round-robin call sites for symmetry. combo/comboStructure.ts (previously under cap, un-frozen) grows 794->918 (+124) — new home for describeCapabilityFilterExhaustion + providerSupportsEmulatedToolCalling (#5240 emulated tool-calling exemption so fail-closed does not regress prompt-emulation-only combos like all-chatgpt-web). Irreducible orchestration wiring at the existing filter chokepoint (same precedent as #7301's universal-cooldown-retry generalization). Companion test tests/unit/combo-routing-engine.test.ts 3409->3449 (+40, fail-closed/fail-open coverage across both call sites) also rebaselined. Covered by tests/unit/8488-capability-filter-fail-closed.test.ts (new) + 95/95 passing across both files. Structural shrink of combo.ts tracked in #3501.",
  "_rebaseline_2026_07_25_8499_ts7_result_union_predicates": "PR #8499 (backryun, chore/ts7-types-executor-scattered) own growth: muse-spark-web.ts 1396->1405 (+9, irreducible). Under this workspace's `strictNullChecks: false`, the boolean-literal discriminant on `GraphqlResult` (`{ ok: true } | { ok: false; error: string }`) narrows the positive `.ok===true` branch but leaves `!result.ok` at the full union under TS7, making `.error` unreachable to the checker at the two call sites (warmup, mode-switch). Fixed by adding a single `isGraphqlFailure()` type-predicate helper (doc comment + 3-line body) reused at both call sites instead of duplicating the predicate inline — not extractable to a shared module without splitting a single-file executor's local narrowing helper out of its own file. Covered by the existing muse-spark-web executor test suite (no behavior change, pure narrowing fix).",
@@ -350,7 +349,7 @@
    "open-sse/executors/deepseek-web.ts": 1148,
    "open-sse/executors/grok-web.ts": 1044,
    "open-sse/executors/muse-spark-web.ts": 1405,
-    "open-sse/handlers/chatCore.ts": 5023,
+    "open-sse/handlers/chatCore.ts": 5020,
    "open-sse/handlers/imageGeneration.ts": 3101,
    "open-sse/handlers/responseSanitizer.ts": 1115,
    "open-sse/handlers/search.ts": 1536,
@@ -401,8 +400,8 @@
    "src/shared/components/RequestLoggerV2.tsx": 1629,
    "src/shared/components/analytics/charts.tsx": 1035,
    "src/shared/services/cliRuntime.ts": 1122,
-    "src/sse/handlers/chat.ts": 1846,
-    "src/sse/services/auth.ts": 2512,
+    "src/sse/handlers/chat.ts": 1845,
+    "src/sse/services/auth.ts": 2508,
    "tests/unit/account-fallback-service.test.ts": 1572,
    "tests/unit/provider-validation-specialty.test.ts": 2980,
    "open-sse/executors/hyperagent.ts": 1026
--- a/docker-compose.prod.yml
+++ b/docker-compose.prod.yml
@@ -46,8 +46,6 @@ services:
    depends_on:
      redis:
        condition: service_healthy
-      chatgpt-web-codex-browser:
-        condition: service_started
    build:
      context: .
      target: runner-cli
@@ -69,7 +67,6 @@ services:
      - HOSTNAME=0.0.0.0
      - DATA_DIR=/app/data
      - OMNIROUTE_BASE_PATH=${OMNIROUTE_BASE_PATH:-}
-      - CHATGPT_WEB_CODEX_CDP_URL=http://chatgpt-web-codex-browser:9223
    ports:
      - "${PROD_DASHBOARD_PORT:-20130}:${DASHBOARD_PORT:-${PORT:-20128}}"
      - "${PROD_API_PORT:-20131}:${API_PORT:-20129}"
@@ -83,19 +80,7 @@ services:
      retries: 3
      start_period: 15s

-  chatgpt-web-codex-browser:
-    build:
-      context: .
-      dockerfile: docker/chatgpt-web-codex-browser/Dockerfile
-    image: omniroute:chatgpt-web-codex-browser
-    restart: unless-stopped
-    shm_size: "2gb"
-    volumes:
-      - chatgpt-web-codex-browser-prod-data:/browser-profile
-
 volumes:
-  chatgpt-web-codex-browser-prod-data:
-    name: omniroute-chatgpt-web-codex-browser-prod-data
  omniroute-prod-data:
    name: omniroute-prod-data
  redis-prod-data:
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -98,21 +98,6 @@ services:
      args:
        OMNIROUTE_BASE_PATH: ${OMNIROUTE_BASE_PATH:-}
    image: omniroute:web
-    depends_on:
-      chatgpt-web-codex-browser:
-        condition: service_started
-    environment:
-      - DATA_DIR=/app/data
-      - PORT=${PORT:-20128}
-      - DASHBOARD_PORT=${DASHBOARD_PORT:-20128}
-      - API_PORT=${API_PORT:-20129}
-      - API_HOST=${API_HOST:-0.0.0.0}
-      - LIVE_WS_PORT=${LIVE_WS_PORT:-20132}
-      - LIVE_WS_HOST=${LIVE_WS_HOST:-0.0.0.0}
-      - LIVE_WS_ALLOWED_ORIGINS=${LIVE_WS_ALLOWED_ORIGINS:-http://localhost:20128,http://127.0.0.1:20128}
-      - REDIS_URL=${REDIS_URL:-redis://redis:6379}
-      - OMNIROUTE_BASE_PATH=${OMNIROUTE_BASE_PATH:-}
-      - CHATGPT_WEB_CODEX_CDP_URL=http://chatgpt-web-codex-browser:9223
    ports:
      - "${DASHBOARD_PORT:-20128}:${DASHBOARD_PORT:-20128}"
      - "${API_PORT:-20129}:${API_PORT:-20129}"
@@ -120,20 +105,6 @@ services:
    profiles:
      - web

-  # Internal-only Chromium runtime for ChatGPT Web (Codex). No CDP or browser
-  # UI port is published to the host.
-  chatgpt-web-codex-browser:
-    build:
-      context: .
-      dockerfile: docker/chatgpt-web-codex-browser/Dockerfile
-    image: omniroute:chatgpt-web-codex-browser
-    restart: unless-stopped
-    shm_size: "2gb"
-    volumes:
-      - chatgpt-web-codex-browser-data:/browser-profile
-    profiles:
-      - web
-
  # ── Profile: cli (CLIs installed inside container) ─────────────────
  omniroute-cli:
    <<: *common
@@ -281,8 +252,6 @@ services:
      - cliproxyapi

 volumes:
-  chatgpt-web-codex-browser-data:
-    name: omniroute-chatgpt-web-codex-browser-data
  cliproxyapi-data:
    name: cliproxyapi-data
  redis-data:
--- a/docker/chatgpt-web-codex-browser/Dockerfile
+++ b/docker/chatgpt-web-codex-browser/Dockerfile
@@ -1,10 +0,0 @@
-FROM mcr.microsoft.com/playwright:v1.62.0-noble
-
-USER root
-RUN mkdir -p /browser-profile && chown -R pwuser:pwuser /browser-profile
-COPY --chown=pwuser:pwuser docker/chatgpt-web-codex-browser/cdp-proxy.mjs /opt/cdp-proxy.mjs
-USER pwuser
-
-EXPOSE 9223
-
-CMD ["/bin/sh", "-lc", "node /opt/cdp-proxy.mjs & exec $(find /ms-playwright -path '*/chrome-linux/chrome' -type f | head -n 1) --headless=new --no-sandbox --disable-dev-shm-usage --remote-debugging-port=9222 --user-data-dir=/browser-profile about:blank"]
--- a/docker/chatgpt-web-codex-browser/cdp-proxy.mjs
+++ b/docker/chatgpt-web-codex-browser/cdp-proxy.mjs
@@ -1,72 +0,0 @@
-import http from "node:http";
-import net from "node:net";
-
-const listenPort = 9223;
-const upstreamHost = "127.0.0.1";
-const upstreamPort = 9222;
-
-function proxyHeaders(headers) {
-  const next = { ...headers, host: `${upstreamHost}:${upstreamPort}` };
-  delete next.connection;
-  delete next.upgrade;
-  return next;
-}
-
-const server = http.createServer((request, response) => {
-  const upstream = http.request(
-    {
-      host: upstreamHost,
-      port: upstreamPort,
-      method: request.method,
-      path: request.url,
-      headers: proxyHeaders(request.headers),
-    },
-    (upstreamResponse) => {
-      const chunks = [];
-      upstreamResponse.on("data", (chunk) => chunks.push(chunk));
-      upstreamResponse.on("end", () => {
-        let body = Buffer.concat(chunks);
-        const contentType = String(upstreamResponse.headers["content-type"] || "");
-        if (contentType.includes("application/json")) {
-          body = Buffer.from(
-            body
-              .toString("utf8")
-              .replaceAll(`ws://${upstreamHost}:${upstreamPort}`, `ws://${request.headers.host}`)
-          );
-        }
-        const headers = { ...upstreamResponse.headers, "content-length": String(body.length) };
-        response.writeHead(upstreamResponse.statusCode || 502, headers);
-        response.end(body);
-      });
-    }
-  );
-  upstream.on("error", () => {
-    response.writeHead(503, { "content-type": "application/json" });
-    response.end(JSON.stringify({ error: "CDP browser is starting" }));
-  });
-  request.pipe(upstream);
-});
-
-server.on("upgrade", (request, socket, head) => {
-  const upstream = net.connect(upstreamPort, upstreamHost, () => {
-    const upgradeHeaders = {
-      ...request.headers,
-      host: `${upstreamHost}:${upstreamPort}`,
-      connection: "Upgrade",
-      upgrade: "websocket",
-    };
-    const headers = Object.entries(upgradeHeaders)
-      .flatMap(([name, value]) =>
-        Array.isArray(value) ? value.map((item) => `${name}: ${item}`) : [`${name}: ${value}`]
-      )
-      .join("\r\n");
-    upstream.write(
-      `${request.method} ${request.url} HTTP/${request.httpVersion}\r\n${headers}\r\n\r\n`
-    );
-    if (head.length > 0) upstream.write(head);
-    socket.pipe(upstream).pipe(socket);
-  });
-  upstream.on("error", () => socket.destroy());
-});
-
-server.listen(listenPort, "0.0.0.0");
--- a/docker/devin-bridge/Dockerfile
+++ b/docker/devin-bridge/Dockerfile
@@ -1,57 +0,0 @@
-FROM node:26.0.0-bookworm-slim
-
-ARG CLAUDE_CODE_VERSION=2.1.220
-ARG DEVIN_CLI_VERSION=3000.2.17
-ARG TARGETARCH
-
-RUN apt-get update \
-  && apt-get install -y --no-install-recommends ca-certificates curl git bash python3 make g++ tini \
-  && rm -rf /var/lib/apt/lists/* \
-  && npm install --global "@anthropic-ai/claude-code@${CLAUDE_CODE_VERSION}"
-
-RUN set -eu; \
-  case "${TARGETARCH}" in \
-    amd64) devin_arch=x86_64-unknown-linux; devin_sha=f0e1e9363afc6ee68c4ef87bab4aeb7ff5cc08a5fa838350ef3ceefdbb2a2be2 ;; \
-    arm64) devin_arch=aarch64-unknown-linux; devin_sha=116dc71ef085a922bc3ff0ea0377d4b26c529a431d58246e36572913e2d25624 ;; \
-    *) echo "Unsupported TARGETARCH=${TARGETARCH}" >&2; exit 1 ;; \
-  esac; \
-  curl -fsSL "https://static.devin.ai/cli/${DEVIN_CLI_VERSION}/devin-${DEVIN_CLI_VERSION}-${devin_arch}.tar.gz" -o /tmp/devin.tar.gz; \
-  echo "${devin_sha}  /tmp/devin.tar.gz" | sha256sum -c -; \
-  tar -xzf /tmp/devin.tar.gz -C /tmp; \
-  install -m 0755 "$(find /tmp -type f -name devin | head -1)" /usr/local/bin/devin; \
-  rm -rf /tmp/devin.tar.gz /tmp/devin-*
-
-RUN groupadd --gid 10001 bridge \
-  && useradd --uid 10001 --gid bridge --create-home --home-dir /home/bridge --shell /bin/bash bridge \
-  && mkdir -p /opt/omniroute /workspace \
-  && chown -R bridge:bridge /opt/omniroute /workspace
-
-WORKDIR /opt/omniroute
-USER bridge
-COPY --chown=bridge:bridge package.json package-lock.json .npmrc ./
-RUN npm ci --ignore-scripts --no-audit --fund=false
-COPY --chown=bridge:bridge . .
-RUN npm rebuild better-sqlite3 || true
-
-ENV HOME=/home/bridge \
-  CLAUDE_CONFIG_DIR=/home/bridge/.claude-devin-isolated \
-  DEVIN_AGENTIC_HOME=/home/bridge \
-  DATA_DIR=/home/bridge/.omniroute-isolated \
-  SQLITE_FILE=/home/bridge/.omniroute-isolated/storage.sqlite \
-  CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1 \
-  DISABLE_TELEMETRY=1 \
-  DISABLE_ERROR_REPORTING=1 \
-  DISABLE_AUTOUPDATER=1 \
-  CLAUDE_CODE_ENABLE_GATEWAY_MODEL_DISCOVERY=1 \
-  NEXT_TELEMETRY_DISABLED=1
-
-RUN mkdir -p /home/bridge/.claude-devin-isolated /home/bridge/.local/share/devin \
-  /home/bridge/.omniroute-isolated
-
-RUN DATA_DIR=/tmp/omniroute-build-data \
-  SQLITE_FILE=/tmp/omniroute-build-data/storage.sqlite \
-  npm run build \
-  && rm -rf /tmp/omniroute-build-data
-
-ENTRYPOINT ["/usr/bin/tini", "--"]
-CMD ["bash"]
--- a/docker/devin-bridge/compose.yml
+++ b/docker/devin-bridge/compose.yml
@@ -1,218 +0,0 @@
-name: omniroute-devin-bridge
-
-x-isolated-environment: &isolated-environment
-  HOME: /home/bridge
-  CLAUDE_CONFIG_DIR: /home/bridge/.claude-devin-isolated
-  DEVIN_AGENTIC_HOME: /home/bridge
-  DATA_DIR: /home/bridge/.omniroute-isolated
-  SQLITE_FILE: /home/bridge/.omniroute-isolated/storage.sqlite
-  ANTHROPIC_BASE_URL: http://omniroute:20128
-  ANTHROPIC_AUTH_TOKEN: sk-local-devin-gateway
-  CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC: "1"
-  DISABLE_TELEMETRY: "1"
-  DISABLE_ERROR_REPORTING: "1"
-  DISABLE_AUTOUPDATER: "1"
-  CLAUDE_CODE_ENABLE_GATEWAY_MODEL_DISCOVERY: "1"
-  DEVIN_BRIDGE_MODEL: ${DEVIN_BRIDGE_MODEL:-devin-cli-agentic/swe-1-7}
-  ANTHROPIC_MODEL: ${DEVIN_BRIDGE_MODEL:-devin-cli-agentic/swe-1-7}
-  ANTHROPIC_DEFAULT_SONNET_MODEL: ${DEVIN_BRIDGE_SONNET_MODEL:-devin-cli-agentic/swe-1-7}
-  ANTHROPIC_DEFAULT_OPUS_MODEL: ${DEVIN_BRIDGE_OPUS_MODEL:-devin-cli-agentic/swe-1-7}
-  ANTHROPIC_DEFAULT_HAIKU_MODEL: ${DEVIN_BRIDGE_HAIKU_MODEL:-devin-cli-agentic/swe-1-7}
-  CLAUDE_CODE_SUBAGENT_MODEL: ${DEVIN_BRIDGE_SUBAGENT_MODEL:-devin-cli-agentic/swe-1-7}
-  REQUIRE_API_KEY: "true"
-  OMNIROUTE_API_KEY: sk-local-devin-gateway
-
-x-runtime: &runtime
-  image: omniroute-devin-bridge:local
-  build:
-    context: ../..
-    dockerfile: docker/devin-bridge/Dockerfile
-    args:
-      CLAUDE_CODE_VERSION: 2.1.220
-      DEVIN_CLI_VERSION: 3000.2.17
-  user: "10001:10001"
-  read_only: true
-  tmpfs:
-    - /tmp:rw,noexec,nosuid,nodev,size=256m
-    - /opt/omniroute/.source:rw,nosuid,nodev,size=16m,uid=10001,gid=10001
-  cap_drop: [ALL]
-  security_opt: [no-new-privileges:true]
-  environment: *isolated-environment
-  networks: [bridge-internal]
-
-services:
-  omniroute:
-    <<: *runtime
-    profiles: [offline]
-    hostname: omniroute
-    environment:
-      <<: *isolated-environment
-      CLI_DEVIN_AGENTIC_BIN: /opt/omniroute/docker/devin-bridge/mock-devin.mjs
-      DEVIN_BRIDGE_MOCK_LOG: /evidence/mock-acp.jsonl
-    command: ["npm", "run", "start"]
-    healthcheck:
-      test:
-        [
-          "CMD",
-          "node",
-          "-e",
-          "fetch('http://127.0.0.1:20128/healthz').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))",
-        ]
-      interval: 2s
-      timeout: 2s
-      retries: 60
-    volumes:
-      - omniroute-offline-data:/home/bridge/.omniroute-isolated
-      - ../../.sandbox/evidence:/evidence
-      - ./mock-devin.mjs:/opt/omniroute/docker/devin-bridge/mock-devin.mjs:ro
-
-  claude:
-    <<: *runtime
-    profiles: [offline]
-    depends_on:
-      omniroute:
-        condition: service_healthy
-      claude-egress-guard:
-        condition: service_healthy
-    working_dir: /workspace
-    command: ["bash", "/opt/omniroute/docker/devin-bridge/run-claude-e2e.sh"]
-    environment:
-      <<: *isolated-environment
-      NODE_USE_ENV_PROXY: "1"
-      HTTP_PROXY: http://claude-egress-guard:8080
-      HTTPS_PROXY: http://claude-egress-guard:8080
-      NO_PROXY: omniroute
-    volumes:
-      - claude-isolated-config:/home/bridge/.claude-devin-isolated
-      - ../../.sandbox/e2e-workspace:/workspace
-      - ../../.sandbox/evidence:/evidence
-      - ./run-claude-e2e.sh:/opt/omniroute/docker/devin-bridge/run-claude-e2e.sh:ro
-
-  contract:
-    <<: *runtime
-    profiles: [offline]
-    depends_on:
-      omniroute:
-        condition: service_healthy
-    command: ["node", "/opt/omniroute/docker/devin-bridge/run-contract.mjs"]
-    volumes:
-      - ./run-contract.mjs:/opt/omniroute/docker/devin-bridge/run-contract.mjs:ro
-
-  claude-egress-guard:
-    image: node:26.0.0-bookworm-slim
-    profiles: [offline, live-devin]
-    user: "10001:10001"
-    read_only: true
-    cap_drop: [ALL]
-    security_opt: [no-new-privileges:true]
-    command: ["node", "/guard/proxy.mjs"]
-    environment:
-      GUARD_LISTEN: 0.0.0.0:8080
-      GUARD_POLICY: deny-all
-      GUARD_LOG: /guard-audit/egress.jsonl
-    healthcheck:
-      test:
-        [
-          "CMD",
-          "node",
-          "-e",
-          "require('net').connect(8080,'127.0.0.1').on('connect',()=>process.exit(0)).on('error',()=>process.exit(1))",
-        ]
-      interval: 1s
-      timeout: 1s
-      retries: 15
-    volumes:
-      - ./network-guard:/guard:ro
-      - ../../.sandbox/guard-audit/claude:/guard-audit
-    networks: [bridge-internal]
-
-  network-guard:
-    image: node:26.0.0-bookworm-slim
-    profiles: [live-devin]
-    user: "10001:10001"
-    read_only: true
-    cap_drop: [ALL]
-    security_opt: [no-new-privileges:true]
-    command: ["node", "/guard/proxy.mjs"]
-    environment:
-      GUARD_LISTEN: 0.0.0.0:8080
-      GUARD_POLICY: devin
-      GUARD_LOG: /guard-audit/egress.jsonl
-    healthcheck:
-      test:
-        [
-          "CMD",
-          "node",
-          "-e",
-          "require('net').connect(8080,'127.0.0.1').on('connect',()=>process.exit(0)).on('error',()=>process.exit(1))",
-        ]
-      interval: 1s
-      timeout: 1s
-      retries: 15
-    volumes:
-      - ./network-guard:/guard:ro
-      - ../../.sandbox/guard-audit/devin:/guard-audit
-    networks: [devin-guard-internal, guard-egress]
-
-  omniroute-live:
-    <<: *runtime
-    profiles: [live-devin]
-    hostname: omniroute
-    depends_on:
-      network-guard:
-        condition: service_healthy
-    environment:
-      <<: *isolated-environment
-      CLI_DEVIN_AGENTIC_BIN: /usr/local/bin/devin
-      DEVIN_BRIDGE_PROXY_URL: http://network-guard:8080
-    networks: [bridge-internal, devin-guard-internal]
-    command: ["npm", "run", "start"]
-    healthcheck:
-      test:
-        [
-          "CMD",
-          "node",
-          "-e",
-          "fetch('http://127.0.0.1:20128/healthz').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))",
-        ]
-      interval: 2s
-      timeout: 2s
-      retries: 60
-    volumes:
-      - devin-auth:/home/bridge/.local/share/devin
-      - omniroute-live-data:/home/bridge/.omniroute-isolated
-
-  claude-live:
-    <<: *runtime
-    profiles: [live-devin]
-    depends_on:
-      omniroute-live:
-        condition: service_healthy
-      claude-egress-guard:
-        condition: service_healthy
-    working_dir: /workspace
-    command: ["bash", "/opt/omniroute/docker/devin-bridge/run-claude-live-e2e.sh"]
-    environment:
-      <<: *isolated-environment
-      NODE_USE_ENV_PROXY: "1"
-      HTTP_PROXY: http://claude-egress-guard:8080
-      HTTPS_PROXY: http://claude-egress-guard:8080
-      NO_PROXY: omniroute
-    volumes:
-      - claude-isolated-config:/home/bridge/.claude-devin-isolated
-      - ../../.sandbox/live-workspace:/workspace
-      - ../../.sandbox/evidence:/evidence
-      - ./run-claude-live-e2e.sh:/opt/omniroute/docker/devin-bridge/run-claude-live-e2e.sh:ro
-
-networks:
-  bridge-internal:
-    internal: true
-  devin-guard-internal:
-    internal: true
-  guard-egress: {}
-
-volumes:
-  claude-isolated-config: {}
-  devin-auth: {}
-  omniroute-offline-data: {}
-  omniroute-live-data: {}
--- a/docker/devin-bridge/mock-devin.mjs
+++ b/docker/devin-bridge/mock-devin.mjs
@@ -1,229 +0,0 @@
-#!/usr/bin/env node
-import fs from "node:fs";
-import readline from "node:readline";
-
-if (
-  process.argv[2] !== "acp" ||
-  process.argv[3] !== "--agent-type" ||
-  process.argv[4] !== "summarizer" ||
-  process.argv.length !== 5
-) {
-  process.exit(64);
-}
-
-const logFile = process.env.DEVIN_BRIDGE_MOCK_LOG || "/evidence/mock-acp.jsonl";
-const rl = readline.createInterface({ input: process.stdin });
-const send = (value) => process.stdout.write(`${JSON.stringify(value)}\n`);
-const log = (value) => fs.appendFileSync(logFile, `${JSON.stringify(value)}\n`);
-
-const actions = [
-  {
-    name: "Skill",
-    arguments: { skill: "bridge-proof" },
-  },
-  {
-    name: "Bash",
-    arguments: {
-      command: "find . -maxdepth 2 -type f -print",
-      description: "Locate the fixture files",
-    },
-  },
-  {
-    name: "Read",
-    arguments: { file_path: "/workspace/math.js" },
-  },
-  {
-    name: "Edit",
-    arguments: {
-      file_path: "/workspace/math.js",
-      old_string: "return a - b;",
-      new_string: "return a * b;",
-    },
-  },
-  {
-    name: "Bash",
-    arguments: { command: "npm test", description: "Run the fixture tests" },
-  },
-  {
-    name: "Edit",
-    arguments: {
-      file_path: "/workspace/math.js",
-      old_string: "return a * b;",
-      new_string: "return a + b;",
-    },
-  },
-  {
-    name: "Bash",
-    arguments: { command: "npm test", description: "Confirm the corrected fixture" },
-  },
-];
-
-rl.on("line", (line) => {
-  const message = JSON.parse(line);
-  if (message.method === "initialize") {
-    if (message.params?.protocolVersion !== 1) {
-      send({ jsonrpc: "2.0", id: message.id, error: { code: -32602, message: "ACP v1 required" } });
-      return;
-    }
-    send({ jsonrpc: "2.0", id: message.id, result: { protocolVersion: 1 } });
-  } else if (message.method === "session/new") {
-    if (message.params?.cwd !== "/home/bridge" || !Array.isArray(message.params?.mcpServers)) {
-      send({ jsonrpc: "2.0", id: message.id, error: { code: -32602, message: "unsafe session" } });
-      return;
-    }
-    send({
-      jsonrpc: "2.0",
-      id: message.id,
-      result: { sessionId: "offline" },
-    });
-  } else if (message.method === "session/set_config_option") {
-    send({
-      jsonrpc: "2.0",
-      id: message.id,
-      error: { code: -32602, message: "summarizer mode must not be mutated" },
-    });
-  } else if (message.method === "session/prompt") {
-    const prompt = String(message.params?.prompt?.[0]?.text || "");
-    if (!prompt.includes("[Devin Summarizer Bridge]") || !prompt.includes("[Execution Trace]")) {
-      send({
-        jsonrpc: "2.0",
-        id: message.id,
-        error: { code: -32602, message: "summarizer bridge framing required" },
-      });
-      return;
-    }
-    if (prompt.includes("CONTRACT_AFTER_TOOL")) {
-      log({ provider: "devin-cli-agentic", scenario: "after-tool" });
-      send({
-        jsonrpc: "2.0",
-        method: "session/update",
-        params: {
-          sessionId: "offline",
-          update: {
-            sessionUpdate: "agent_message_chunk",
-            content: { type: "text", text: "contract continued" },
-          },
-        },
-      });
-      send({ jsonrpc: "2.0", id: message.id, result: { stopReason: "end_turn" } });
-      return;
-    }
-    if (prompt.includes("CONTRACT_EXIT")) {
-      log({ provider: "devin-cli-agentic", scenario: "exit" });
-      process.exit(7);
-    }
-    if (prompt.includes("CONTRACT_ERROR")) {
-      log({ provider: "devin-cli-agentic", scenario: "error" });
-      send({
-        jsonrpc: "2.0",
-        id: message.id,
-        error: { code: -32000, message: "deterministic upstream failure" },
-      });
-      return;
-    }
-    if (prompt.includes("CONTRACT_TEXT")) {
-      log({ provider: "devin-cli-agentic", scenario: "text" });
-      send({
-        jsonrpc: "2.0",
-        method: "session/update",
-        params: {
-          sessionId: "offline",
-          update: {
-            sessionUpdate: "agent_message_chunk",
-            content: { type: "text", text: "contract text" },
-          },
-        },
-      });
-      send({ jsonrpc: "2.0", id: message.id, result: { stopReason: "end_turn" } });
-      return;
-    }
-    if (prompt.includes("CONTRACT_NARRATIVE_REPAIR")) {
-      const isRepair = prompt.includes("[Single Repair Attempt]");
-      log({
-        provider: "devin-cli-agentic",
-        scenario: "narrative-repair",
-        stage: isRepair ? "repair" : "initial",
-      });
-      send({
-        jsonrpc: "2.0",
-        method: "session/update",
-        params: {
-          sessionId: "offline",
-          update: {
-            sessionUpdate: "agent_message_chunk",
-            content: {
-              type: "text",
-              text: isRepair
-                ? '<tool>{"name":"Read","arguments":{"file_path":"/workspace/math.js"}}</tool>'
-                : "I'll start by reading the math.js file, then run the tests.",
-            },
-          },
-        },
-      });
-      send({ jsonrpc: "2.0", id: message.id, result: { stopReason: "end_turn" } });
-      return;
-    }
-    if (prompt.includes("CONTRACT_TOOL")) {
-      log({ provider: "devin-cli-agentic", scenario: "tool" });
-      send({
-        jsonrpc: "2.0",
-        method: "session/update",
-        params: {
-          sessionId: "offline",
-          update: {
-            sessionUpdate: "agent_message_chunk",
-            content: {
-              type: "text",
-              text: '<tool>{"name":"Read","arguments":{"file_path":"/workspace/math.js"}}</tool>',
-            },
-          },
-        },
-      });
-      send({ jsonrpc: "2.0", id: message.id, result: { stopReason: "end_turn" } });
-      return;
-    }
-    const resultCount = (prompt.match(/\[Tool Result\]/g) || []).length;
-    if (!prompt.includes("CLAUDE_MD_BRIDGE_ACTIVE") || !prompt.includes("COMMAND_BRIDGE_ACTIVE")) {
-      send({
-        jsonrpc: "2.0",
-        id: message.id,
-        error: { code: -32602, message: "Claude project context missing" },
-      });
-      return;
-    }
-
-    const action = actions[resultCount];
-    const text = action
-      ? `<tool>${JSON.stringify(action)}</tool>`
-      : "BRIDGE_E2E_COMPLETE CLAUDE_MD_BRIDGE_ACTIVE SKILL_BRIDGE_ACTIVE COMMAND_BRIDGE_ACTIVE";
-    if (!action && !prompt.includes("SKILL_BRIDGE_ACTIVE")) {
-      send({
-        jsonrpc: "2.0",
-        id: message.id,
-        error: { code: -32602, message: "Skill result missing" },
-      });
-      return;
-    }
-    log({
-      provider: "devin-cli-agentic",
-      model: message.params?.model || "swe-1-7",
-      resultCount,
-      action: action?.name || "final",
-    });
-    const midpoint = Math.max(1, Math.floor(text.length / 2));
-    for (const chunk of [text.slice(0, midpoint), text.slice(midpoint)]) {
-      send({
-        jsonrpc: "2.0",
-        method: "session/update",
-        params: {
-          sessionId: "offline",
-          update: {
-            sessionUpdate: "agent_message_chunk",
-            content: { type: "text", text: chunk },
-          },
-        },
-      });
-    }
-    send({ jsonrpc: "2.0", id: message.id, result: { stopReason: "end_turn" } });
-  }
-});
--- a/docker/devin-bridge/network-guard/policy.mjs
+++ b/docker/devin-bridge/network-guard/policy.mjs
@@ -1,130 +0,0 @@
-export const DEVIN_ALLOWED_SUFFIXES = Object.freeze([".devin.ai", ".cognition.ai"]);
-export const DEVIN_ALLOWED_EXACT_HOSTS = Object.freeze([
-  "server.codeium.com",
-  "unleash.codeium.com",
-]);
-
-function normalizeHostname(hostname) {
-  return String(hostname || "")
-    .trim()
-    .toLowerCase()
-    .replace(/\.$/, "");
-}
-
-export function isAllowedGuardHostname(hostname, policy = "deny-all") {
-  if (policy !== "devin") return false;
-  const value = normalizeHostname(hostname);
-  if (!value) return false;
-  if (DEVIN_ALLOWED_EXACT_HOSTS.includes(value)) return true;
-  return DEVIN_ALLOWED_SUFFIXES.some(
-    (suffix) => value === suffix.slice(1) || value.endsWith(suffix)
-  );
-}
-
-const HOP_BY_HOP_HEADERS = new Set([
-  "connection",
-  "keep-alive",
-  "proxy-authenticate",
-  "proxy-authorization",
-  "proxy-connection",
-  "te",
-  "trailer",
-  "transfer-encoding",
-  "upgrade",
-]);
-
-export function sanitizeForwardHeaders(headers, target) {
-  const connectionTokens = String(headers.connection || "")
-    .split(",")
-    .map((value) => value.trim().toLowerCase())
-    .filter(Boolean);
-  const blocked = new Set([...HOP_BY_HOP_HEADERS, ...connectionTokens]);
-  const sanitized = {};
-  for (const [name, value] of Object.entries(headers)) {
-    if (value === undefined || blocked.has(name.toLowerCase()) || name.toLowerCase() === "host") {
-      continue;
-    }
-    sanitized[name] = value;
-  }
-  sanitized.host = target.host;
-  return sanitized;
-}
-
-export function parseConnectAuthority(authority) {
-  const value = String(authority || "");
-  const match = value.match(/^(?:\[([^\]]+)\]|([^:]+)):(\d+)$/);
-  if (!match) return null;
-  const hostname = normalizeHostname(match[1] || match[2]);
-  const port = Number(match[3]);
-  if (!hostname || port !== 443) return null;
-  return { hostname, port };
-}
-
-function readUint24(buffer, offset) {
-  return (buffer[offset] << 16) | (buffer[offset + 1] << 8) | buffer[offset + 2];
-}
-
-export function parseTlsClientHelloSni(buffer) {
-  if (!Buffer.isBuffer(buffer)) return { status: "invalid", reason: "not_buffer" };
-  let offset = 0;
-  const handshakeParts = [];
-  while (offset < buffer.length) {
-    if (buffer.length - offset < 5) return { status: "need-more" };
-    if (buffer[offset] !== 22) return { status: "invalid", reason: "not_handshake_record" };
-    const recordLength = buffer.readUInt16BE(offset + 3);
-    if (recordLength <= 0 || recordLength > 18432) {
-      return { status: "invalid", reason: "invalid_record_length" };
-    }
-    if (buffer.length - offset - 5 < recordLength) return { status: "need-more" };
-    handshakeParts.push(buffer.subarray(offset + 5, offset + 5 + recordLength));
-    offset += 5 + recordLength;
-  }
-  const handshake = Buffer.concat(handshakeParts);
-  if (handshake.length < 4) return { status: "need-more" };
-  if (handshake[0] !== 1) return { status: "invalid", reason: "not_client_hello" };
-  const helloLength = readUint24(handshake, 1);
-  if (helloLength > 65531) return { status: "invalid", reason: "client_hello_too_large" };
-  if (handshake.length - 4 < helloLength) return { status: "need-more" };
-  const hello = handshake.subarray(4, 4 + helloLength);
-  let cursor = 34;
-  if (hello.length < cursor + 1) return { status: "invalid", reason: "truncated_hello" };
-  const sessionLength = hello[cursor++];
-  cursor += sessionLength;
-  if (hello.length < cursor + 2) return { status: "invalid", reason: "truncated_ciphers" };
-  const cipherLength = hello.readUInt16BE(cursor);
-  cursor += 2 + cipherLength;
-  if (hello.length < cursor + 1) return { status: "invalid", reason: "truncated_compression" };
-  const compressionLength = hello[cursor++];
-  cursor += compressionLength;
-  if (hello.length < cursor + 2) return { status: "invalid", reason: "missing_extensions" };
-  const extensionsLength = hello.readUInt16BE(cursor);
-  cursor += 2;
-  const extensionsEnd = cursor + extensionsLength;
-  if (extensionsEnd > hello.length) return { status: "invalid", reason: "truncated_extensions" };
-  while (cursor < extensionsEnd) {
-    if (extensionsEnd - cursor < 4) return { status: "invalid", reason: "truncated_extension" };
-    const type = hello.readUInt16BE(cursor);
-    const length = hello.readUInt16BE(cursor + 2);
-    cursor += 4;
-    if (cursor + length > extensionsEnd) {
-      return { status: "invalid", reason: "invalid_extension_length" };
-    }
-    if (type === 0) {
-      const data = hello.subarray(cursor, cursor + length);
-      if (data.length < 5 || data.readUInt16BE(0) !== data.length - 2 || data[2] !== 0) {
-        return { status: "invalid", reason: "invalid_server_name" };
-      }
-      const nameLength = data.readUInt16BE(3);
-      if (nameLength !== data.length - 5) {
-        return { status: "invalid", reason: "invalid_server_name_length" };
-      }
-      const serverName = normalizeHostname(data.subarray(5).toString("ascii"));
-      if (!/^[a-z0-9.-]+$/.test(serverName)) {
-        return { status: "invalid", reason: "invalid_server_name_value" };
-      }
-      return { status: "ok", serverName };
-    }
-    cursor += length;
-  }
-  return { status: "invalid", reason: "missing_sni" };
-}
--- a/docker/devin-bridge/network-guard/proxy.mjs
+++ b/docker/devin-bridge/network-guard/proxy.mjs
@@ -1,136 +0,0 @@
-import fs from "node:fs";
-import http from "node:http";
-import net from "node:net";
-import { pathToFileURL } from "node:url";
-
-import {
-  isAllowedGuardHostname,
-  parseConnectAuthority,
-  parseTlsClientHelloSni,
-  sanitizeForwardHeaders,
-} from "./policy.mjs";
-
-const MAX_CLIENT_HELLO_BYTES = 64 * 1024;
-const CLIENT_HELLO_TIMEOUT_MS = 3000;
-
-export function createGuardProxy({
-  policy = "deny-all",
-  logPath = "/tmp/egress.jsonl",
-  allowHostname = (hostname) => isAllowedGuardHostname(hostname, policy),
-  connectSocket = (port, hostname, onConnect) => net.connect(port, hostname, onConnect),
-} = {}) {
-  if (!new Set(["deny-all", "devin"]).has(policy)) {
-    throw new Error(`Unknown network guard policy: ${policy}`);
-  }
-
-  function audit(hostname, decision, reason) {
-    fs.appendFileSync(
-      logPath,
-      `${JSON.stringify({ at: new Date().toISOString(), hostname, decision, reason })}\n`
-    );
-  }
-
-  const server = http.createServer((req, res) => {
-    let target;
-    try {
-      target = new URL(req.url);
-    } catch {
-      res.writeHead(400).end("invalid proxy target\n");
-      return;
-    }
-    if (target.protocol !== "http:" || target.username || target.password) {
-      audit(target.hostname, "deny", "invalid_http_target");
-      res.writeHead(403).end("egress denied\n");
-      return;
-    }
-    if (!allowHostname(target.hostname)) {
-      audit(target.hostname, "deny", "host_policy");
-      res.writeHead(403).end("egress denied\n");
-      return;
-    }
-    audit(target.hostname, "allow", "host_policy");
-    const upstream = http.request(
-      target,
-      {
-        method: req.method,
-        headers: sanitizeForwardHeaders(req.headers, target),
-      },
-      (reply) => {
-        res.writeHead(reply.statusCode || 502, reply.headers);
-        reply.pipe(res);
-      }
-    );
-    req.pipe(upstream);
-    upstream.on("error", () => res.writeHead(502).end("upstream error\n"));
-  });
-
-  server.on("connect", (req, client, head) => {
-    const authority = parseConnectAuthority(req.url);
-    if (!authority) {
-      audit(req.url, "deny", "invalid_connect_authority");
-      client.end("HTTP/1.1 403 Forbidden\r\n\r\n");
-      return;
-    }
-    const { hostname, port } = authority;
-    if (!allowHostname(hostname)) {
-      audit(hostname, "deny", "host_policy");
-      client.end("HTTP/1.1 403 Forbidden\r\n\r\n");
-      return;
-    }
-
-    let buffer = Buffer.from(head);
-    let settled = false;
-    const timer = setTimeout(() => fail("client_hello_timeout"), CLIENT_HELLO_TIMEOUT_MS);
-    timer.unref?.();
-
-    const cleanup = () => {
-      clearTimeout(timer);
-      client.removeListener("data", onData);
-    };
-    const fail = (reason) => {
-      if (settled) return;
-      settled = true;
-      cleanup();
-      audit(hostname, "deny", reason);
-      client.destroy();
-    };
-    const inspect = () => {
-      if (buffer.length > MAX_CLIENT_HELLO_BYTES) return fail("client_hello_too_large");
-      const parsed = parseTlsClientHelloSni(buffer);
-      if (parsed.status === "need-more") return;
-      if (parsed.status !== "ok") return fail(parsed.reason || "invalid_client_hello");
-      if (parsed.serverName !== hostname) return fail("sni_mismatch");
-      settled = true;
-      cleanup();
-      client.pause();
-      const upstream = connectSocket(port, hostname, () => {
-        audit(hostname, "allow", "sni_match");
-        if (buffer.length) upstream.write(buffer);
-        upstream.pipe(client);
-        client.pipe(upstream);
-        client.resume();
-      });
-      upstream.on("error", () => client.destroy());
-    };
-    const onData = (chunk) => {
-      buffer = Buffer.concat([buffer, chunk]);
-      inspect();
-    };
-
-    client.write("HTTP/1.1 200 Connection Established\r\n\r\n");
-    client.on("data", onData);
-    if (buffer.length) inspect();
-    client.resume();
-  });
-
-  return server;
-}
-
-if (process.argv[1] && pathToFileURL(process.argv[1]).href === import.meta.url) {
-  const [host, portText] = (process.env.GUARD_LISTEN || "0.0.0.0:8080").split(":");
-  const server = createGuardProxy({
-    policy: process.env.GUARD_POLICY || "deny-all",
-    logPath: process.env.GUARD_LOG || "/tmp/egress.jsonl",
-  });
-  server.listen(Number(portText), host);
-}
--- a/docker/devin-bridge/run-claude-e2e.sh
+++ b/docker/devin-bridge/run-claude-e2e.sh
@@ -1,27 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-unset ANTHROPIC_API_KEY CLAUDE_CODE_OAUTH_TOKEN ANTHROPIC_BEDROCK_BASE_URL ANTHROPIC_VERTEX_BASE_URL
-unset CLAUDE_CODE_USE_BEDROCK CLAUDE_CODE_USE_VERTEX CLAUDE_CODE_USE_FOUNDRY
-
-set -o pipefail
-check() {
-  "$@"
-  printf 'E2E check passed: %s\n' "$*"
-}
-
-claude -p --output-format stream-json --verbose --max-turns 12 \
-  --permission-mode bypassPermissions \
-  "/bridge-check" | tee /evidence/claude-stream.jsonl
-
-if grep -Eqi 'log[ -]?in|authenticate.*anthropic|claude\.ai' /evidence/claude-stream.jsonl; then
-  echo "Claude Code requested forbidden authentication" >&2
-  exit 1
-fi
-check grep -q 'return a + b;' /workspace/math.js
-npm test
-check grep -q 'Skill' /workspace/.e2e-hook.log
-check grep -q 'Read' /workspace/.e2e-hook.log
-check grep -q 'Edit' /workspace/.e2e-hook.log
-check grep -q 'Bash' /workspace/.e2e-hook.log
-check grep -q 'BRIDGE_E2E_COMPLETE' /evidence/claude-stream.jsonl
--- a/docker/devin-bridge/run-claude-live-e2e.sh
+++ b/docker/devin-bridge/run-claude-live-e2e.sh
@@ -1,52 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-unset ANTHROPIC_API_KEY CLAUDE_CODE_OAUTH_TOKEN ANTHROPIC_BEDROCK_BASE_URL ANTHROPIC_VERTEX_BASE_URL
-unset CLAUDE_CODE_USE_BEDROCK CLAUDE_CODE_USE_VERTEX CLAUDE_CODE_USE_FOUNDRY
-
-bridge_system_prompt="You are a coding agent inside Claude Code. Use only the client-owned tools supplied in the request. Never execute or request a Devin-owned tool. When work requires a tool, select the appropriate client tool and wait for its result before continuing."
-scenario_cooldown_seconds="${DEVIN_BRIDGE_LIVE_SCENARIO_COOLDOWN_SECONDS:-15}"
-
-run_scenario() {
-  local evidence_file="$1"
-  local prompt="$2"
-  claude -p --output-format stream-json --verbose --max-turns 12 \
-    --tools Read,Edit,Bash \
-    --system-prompt "$bridge_system_prompt" \
-    --permission-mode bypassPermissions "$prompt" | tee "$evidence_file"
-  if grep -Eqi 'log[ -]?in|authenticate.*anthropic|claude\.ai' "$evidence_file"; then
-    echo "Claude Code requested forbidden authentication" >&2
-    exit 1
-  fi
-}
-
-validate_scenario() {
-  local evidence_file="$1"
-  local marker="$2"
-  local required_tools="$3"
-  local require_npm_test="$4"
-  local required_slash_command="${5:-}"
-  local required_skill="${6:-}"
-  local accept_explicit_completion="${7:-false}"
-  node /opt/omniroute/scripts/devin-bridge/validate-claude-evidence.mjs \
-    "$evidence_file" "$marker" "$required_tools" "$require_npm_test" \
-    "$required_slash_command" "$required_skill" "$accept_explicit_completion"
-}
-
-run_scenario /evidence/live-analysis.jsonl \
-  "Read /workspace/CLAUDE.md, /workspace/math.js, and /workspace/math.test.js directly without searching or editing. Explain the defect, then end with LIVE_ANALYSIS_COMPLETE."
-validate_scenario /evidence/live-analysis.jsonl LIVE_ANALYSIS_COMPLETE Read false
-sleep "$scenario_cooldown_seconds"
-
-run_scenario /evidence/live-fix.jsonl \
-  "Use Edit now to replace 'return a - b;' with 'return a + b;' in /workspace/math.js. Then use Bash to run npm test. Do not summarize before npm test succeeds. End with LIVE_FIX_COMPLETE only after the test passes."
-grep -q 'return a + b;' /workspace/math.js
-npm test
-validate_scenario /evidence/live-fix.jsonl LIVE_FIX_COMPLETE Edit,Bash true
-sleep "$scenario_cooldown_seconds"
-
-run_scenario /evidence/live-command.jsonl "/bridge-check"
-validate_scenario /evidence/live-command.jsonl BRIDGE_E2E_COMPLETE Bash true \
-  bridge-check bridge-proof true
-
-printf 'PASS: three live Devin-backed Claude Code scenarios completed\n'
--- a/docker/devin-bridge/run-contract.mjs
+++ b/docker/devin-bridge/run-contract.mjs
@@ -1,135 +0,0 @@
-#!/usr/bin/env node
-import assert from "node:assert/strict";
-
-const endpoint = "http://omniroute:20128/v1/messages";
-const headers = {
-  "anthropic-version": "2023-06-01",
-  "content-type": "application/json",
-  "x-api-key": "sk-local-devin-gateway",
-};
-const model = process.env.DEVIN_BRIDGE_MODEL || "devin-cli-agentic/swe-1-7";
-
-async function request(prompt, extra = {}) {
-  return fetch(endpoint, {
-    method: "POST",
-    headers,
-    body: JSON.stringify({
-      model,
-      max_tokens: 256,
-      messages: [{ role: "user", content: prompt }],
-      ...extra,
-    }),
-  });
-}
-
-const textReply = await request("CONTRACT_TEXT");
-assert.equal(textReply.status, 200);
-assert.match(textReply.headers.get("content-type") || "", /application\/json/);
-const textBody = await textReply.json();
-assert.equal(textBody.type, "message");
-assert.equal(textBody.role, "assistant");
-assert.equal(textBody.stop_reason, "end_turn");
-assert.deepEqual(textBody.content, [{ type: "text", text: "contract text" }]);
-
-const toolReply = await request("CONTRACT_TOOL", {
-  stream: true,
-  tools: [
-    {
-      name: "Read",
-      description: "Read a file",
-      input_schema: {
-        type: "object",
-        properties: { file_path: { type: "string" } },
-        required: ["file_path"],
-        additionalProperties: false,
-      },
-    },
-  ],
-});
-assert.equal(toolReply.status, 200);
-assert.match(toolReply.headers.get("content-type") || "", /text\/event-stream/);
-const toolStream = await toolReply.text();
-const eventNames = toolStream
-  .split("\n")
-  .filter((line) => line.startsWith("event: "))
-  .map((line) => line.slice(7));
-assert.deepEqual(eventNames, [
-  "message_start",
-  "content_block_start",
-  "content_block_delta",
-  "content_block_stop",
-  "message_delta",
-  "message_stop",
-]);
-const toolEvents = toolStream
-  .split("\n")
-  .filter((line) => line.startsWith("data: "))
-  .map((line) => JSON.parse(line.slice(6)));
-const toolUse = toolEvents.find((event) => event.type === "content_block_start")?.content_block;
-assert.equal(toolUse?.type, "tool_use");
-assert.equal(toolUse?.name, "Read");
-assert.match(toolUse?.id || "", /^tool_devin_/);
-
-const repairedNarrativeReply = await request("CONTRACT_NARRATIVE_REPAIR", {
-  tools: [
-    {
-      name: "Read",
-      description: "Read a file",
-      input_schema: {
-        type: "object",
-        properties: { file_path: { type: "string" } },
-        required: ["file_path"],
-        additionalProperties: false,
-      },
-    },
-  ],
-});
-assert.equal(repairedNarrativeReply.status, 200);
-const repairedNarrativeBody = await repairedNarrativeReply.json();
-assert.equal(repairedNarrativeBody.stop_reason, "tool_use");
-assert.equal(repairedNarrativeBody.content?.[0]?.type, "tool_use");
-assert.equal(repairedNarrativeBody.content?.[0]?.name, "Read");
-
-const continuationReply = await fetch(endpoint, {
-  method: "POST",
-  headers,
-  body: JSON.stringify({
-    model,
-    max_tokens: 256,
-    tools: [
-      {
-        name: "Read",
-        description: "Read a file",
-        input_schema: { type: "object", properties: {}, additionalProperties: true },
-      },
-    ],
-    messages: [
-      { role: "user", content: "CONTRACT_TOOL" },
-      { role: "assistant", content: [toolUse] },
-      {
-        role: "user",
-        content: [
-          {
-            type: "tool_result",
-            tool_use_id: toolUse.id,
-            content: "CONTRACT_AFTER_TOOL",
-          },
-        ],
-      },
-    ],
-  }),
-});
-assert.equal(continuationReply.status, 200);
-const continuationBody = await continuationReply.json();
-assert.equal(continuationBody.stop_reason, "end_turn");
-assert.deepEqual(continuationBody.content, [{ type: "text", text: "contract continued" }]);
-
-for (const marker of ["CONTRACT_ERROR", "CONTRACT_EXIT"]) {
-  const failedReply = await request(marker);
-  assert.equal(failedReply.status, 502);
-  const failedBody = await failedReply.json();
-  assert.equal(failedBody.error?.type, "server_error");
-  assert.doesNotMatch(JSON.stringify(failedBody), /stack|anthropic|openai/i);
-}
-
-console.log("PASS: Anthropic Messages wire contracts and fail-closed errors passed");
--- a/docs/DEVIN_CLAUDE_BRIDGE.md
+++ b/docs/DEVIN_CLAUDE_BRIDGE.md
@@ -1,181 +0,0 @@
-# Devin Claude Bridge
-
-`devin-cli-agentic` lets the real Claude Code runtime use OmniRoute's local Anthropic
-Messages endpoint while the official Devin CLI supplies model responses over ACP stdio. It
-does not modify the existing Anthropic, Claude OAuth, Claude Web, or `devin-cli` providers.
-
-> **Current status: offline and live validated.** The pinned Claude Code `2.1.220` completed
-> three isolated scenarios through Devin CLI `3000.2.17` and model
-> `swe-1-7-lightning`. The final live run proved client-owned `Read`, `Edit`, and `Bash`
-> turns, successful `npm test` results, project command and skill discovery, Devin-only
-> routing, and zero Claude egress.
-
-## Architecture
-
-```text
-Claude Code 2.1.220 (isolated non-root Linux container)
-  -> http://omniroute:20128/v1/messages
-  -> devin-cli-agentic (Claude-format, no-auth provider)
-  -> devin acp --agent-type summarizer (official ACP stdio, no Devin tools)
-  -> Devin account in the dedicated devin-auth volume
-```
-
-The official CLI's default ACP agent can execute its own tools, so this bridge does not use
-it. It starts the fixed `summarizer` ACP agent, whose official CLI mode has no tools, and
-frames the serialized Anthropic request as an execution trace. When another Claude-owned
-action is needed, the response must contain exactly one client tool envelope. Any ACP
-`tool_call` or `tool_call_update` is rejected before a response can be reported as
-successful.
-
-The serializer in `open-sse/executors/devin-agentic/serializer.ts` preserves `system`,
-`text`, `tool_use`, `tool_result`, `thinking`, `redacted_thinking`, `tool_choice`, and the
-tools supplied by Claude Code. Images and unknown blocks fail explicitly. Large tool results
-use a visible truncation marker.
-
-The parser accepts one standalone `<tool>{...}</tool>` envelope per model turn. It checks
-the name against the request's tool list, validates arguments against that tool's JSON
-Schema, rejects mixed narrative/actions, and permits one bounded repair. Claude Code then
-executes the resulting Anthropic `tool_use` locally and sends the `tool_result` back through
-OmniRoute.
-
-## Isolation and threat model
-
-The host's Claude installation, account, and configuration are out of scope and treated as
-forbidden. The Compose services:
-
- run as UID/GID `10001:10001`, with a read-only root filesystem, dropped capabilities, and
-  `no-new-privileges`;
- use a private `/home/bridge`, a dedicated Claude config volume, isolated OmniRoute data,
-  and a separate `devin-auth` volume;
- mount only disposable `.sandbox` workspaces/evidence;
- do not mount the host home, Keychain, SSH, cloud credentials, or Docker socket;
- construct explicit environments and remove Anthropic API/OAuth/routing variables;
- direct Claude Code inference only to `http://omniroute:20128` with a local-only key.
-
-The offline profile uses an internal network. In the live profile, OmniRoute reaches the
-official Devin endpoints only through `network-guard`; unrelated destinations are denied.
-Claude Code has a separate deny-all egress guard and can reach only the local OmniRoute
-service through `NO_PROXY`. Guard audit files are mounted only by their guard process. The
-scripts verify file ownership, mode, link count, and every decision before exporting
-token-free evidence.
-
-Run the isolation proof independently:
-
-```bash
-./scripts/devin-bridge/verify-anthropic-isolation
-```
-
-It validates topology, named mounts, non-root/read-only settings, explicit local routing,
-absence of sensitive environment variables, absence of the Docker socket, blocked access to
-`api.anthropic.com` and `claude.ai`, Devin-only provider selection, and explicit failure when
-the ACP backend is unavailable.
-
-## First-time setup and normal use
-
-Build the pinned image:
-
-```bash
-./scripts/devin-bridge/build
-```
-
-Authenticate only the isolated Devin volume:
-
-```bash
-ENABLE_LIVE_DEVIN_TESTS=1 ./scripts/devin-bridge/login-devin
-```
-
-The login command uses the official manual-token flow intended for remote/container
-environments. The value is entered directly into the CLI prompt; it is not passed as a
-process argument, written to Git, or copied from the host.
-
-Launch the isolated Claude Code runtime:
-
-```bash
-./scripts/devin-bridge/launch
-```
-
-`launch` rechecks isolation, Devin authentication, and model discovery before starting the
-containerized Claude Code. It never runs the host's Claude executable. Model aliases can be
-set in `.env.devin-bridge`; every configured value must keep the
-`devin-cli-agentic/` prefix.
-
-## Validation commands
-
-The reproducible offline path requires no Devin account and has no runtime Internet:
-
-```bash
-./scripts/devin-bridge/test-unit
-./scripts/devin-bridge/test-contract
-./scripts/devin-bridge/test-e2e-mock
-./scripts/devin-bridge/verify-anthropic-isolation
-```
-
-The authenticated opt-in live path is:
-
-```bash
-ENABLE_LIVE_DEVIN_TESTS=1 ./scripts/devin-bridge/test-live-devin
-```
-
-The live runner waits between scenarios to avoid opening ACP sessions in a burst and
-validates structured Claude stream events instead of trusting textual claims. Its three
-scenarios prove:
-
-1. direct project reads and defect analysis;
-2. a real `Edit`, a client-owned `Bash` `npm test`, and a terminal result;
-3. `/bridge-check` plus `bridge-proof` discovery, project reads, another successful
-   client-owned `npm test`, and completion without pending work.
-
-The final gate also checks the Devin network audit and requires the Claude egress audit to
-remain empty.
-
-## Updating pinned tools
-
-The image pins Node, Claude Code, and Devin CLI in
-`docker/devin-bridge/Dockerfile`. To update:
-
-1. change the explicit versions;
-2. replace both architecture-specific Devin archive checksums with values for the official
-   artifact;
-3. rebuild and run every offline validation command;
-4. confirm the versions inside the image;
-5. rerun the authenticated three-scenario live suite.
-
-Do not install either CLI globally on the host or replace checksum verification with an
-unverified download.
-
-## Diagnosis and cleanup
-
- `docker compose -f docker/devin-bridge/compose.yml --profile offline logs omniroute`
-  shows local routing and sanitized executor errors.
- `.sandbox/evidence/mock-acp.jsonl` records deterministic mock ACP actions.
- `.sandbox/evidence/claude-stream.jsonl` records the real Claude Code offline run.
- `.sandbox/evidence/live-*.jsonl` records the three validated live streams.
- `.sandbox/evidence/egress.jsonl` and `.sandbox/evidence/claude-egress.jsonl` are validated,
-  token-free copies of the guard audits.
-
-Stop owned containers and networks while preserving login/config volumes:
-
-```bash
-./scripts/devin-bridge/clean
-```
-
-Remove the complete bridge-owned environment, including named volumes:
-
-```bash
-./scripts/devin-bridge/clean --all
-```
-
-## Limits
-
- The bridge relies on the fixed no-tools `summarizer` role because Devin CLI `3000.2.17`
-  does not expose a neutral no-tools ACP agent. The adapter compensates for summary-shaped
-  intermediate responses, but one bounded repair can still fail explicitly.
- Live ACP calls can return transient `502`/`504` responses. The harness spaces scenarios;
-  persistent failure remains fail-closed and never selects another provider.
- ACP context is reconstructed from each Anthropic request; there is no process/session
-  affinity.
- One tool call is supported per model response; parallel calls are rejected.
- Images are explicitly unsupported. Vision, thinking output, effort controls, and a 1M
-  context window are not advertised.
- SSE uses valid Anthropic lifecycle events but is emitted after the bounded ACP turn is
-  collected; ACP chunks are not forwarded incrementally.
--- a/docs/DEVIN_CLAUDE_BRIDGE_PROGRESS.md
+++ b/docs/DEVIN_CLAUDE_BRIDGE_PROGRESS.md
@@ -1,115 +0,0 @@
-# Devin Claude Bridge Progress
-
-Updated: 2026-07-28
-
-## Baseline
-
- Fork version: `3.8.49`.
- Starting branch: `release/v3.8.49`.
- Starting commit: `ed7db3ee5f89a144b2d931d8605534522f83de30`.
- Fixed runtime artifacts: Node `26.0.0`, Claude Code `2.1.220`, Devin CLI `3000.2.17`.
- Existing `devin-cli` remains unchanged; the new path is the separate
-  `devin-cli-agentic` provider.
-
-## Implemented architecture
-
- Claude Code runs only inside the non-root bridge container with its own empty config
-  volume and local OmniRoute base URL.
- `devin-cli-agentic` preserves Anthropic messages, tool schemas, `tool_use`, and
-  `tool_result`, then calls the official Devin CLI over ACP stdio.
- The executor starts `devin acp --agent-type summarizer`. This is the only fixed official
-  ACP role in the pinned CLI that has no Devin-owned tools.
- The request is framed as an execution trace. Devin can return one strict client tool
-  envelope; Claude Code executes that tool locally.
- Internal ACP `tool_call` events, unsupported blocks, invalid schemas, narrative actions,
-  timeouts, cancellation, and process failure all fail closed.
- Provider and network policy prevent combo/auto/Anthropic fallback.
-
-## Offline proof
-
- Focused serializer, parser, executor, ACP lifecycle, wire-format, environment, and audit
-  tests pass (39/39).
- The contract suite covers Anthropic JSON/SSE, `tool_use`, `tool_result` continuation,
-  fragmented ACP frames, stderr, early exit, timeout, cancellation, and fail-closed provider
-  loss.
- The production bridge image builds with the pinned CLIs.
- Real Claude Code offline E2E loads `CLAUDE.md`, the project skill and slash command, fires
-  hooks, executes local tools over multiple turns, observes a failed test, repairs the file,
-  reruns the test, and completes.
- The isolation verifier proves non-root/read-only execution, isolated mounts and config,
-  blocked Anthropic/Claude access, no host credential mounts, local-only inference, and no
-  fallback.
-
-Evidence is generated under `.sandbox/evidence` and ignored by Git.
-
-## Regression status
-
- `typecheck:core`, focused ESLint, Prettier, shell/Node syntax, and the complete documentation
-  accuracy suite pass.
- The broad `npm run check` is not reported as passed: after its lint phase, the repository
-  test runner remained alive while an existing `ioredis` client repeatedly retried an
-  unavailable local Redis endpoint after `quota-redis-store.test.ts`. The bridge-focused
-  suites, production image build, offline E2E, isolation proof, and live gate do not use that
-  Redis service and all pass.
-
-## Live Devin proof
-
-Passed with the official in-container login and discovered model
-`swe-1-7-lightning`. The terminal live run completed all three scenarios:
-
-1. Claude Code loaded the fixture instructions, issued client-owned `Read` calls, and
-   returned a correct defect analysis.
-2. Claude Code issued a real `Edit` changing subtraction to addition, then a client-owned
-   `Bash` call running `npm test`; the test reported one pass and zero failures.
-3. Claude Code initialization listed `bridge-check` and `bridge-proof`, read the corrected
-   source and test, executed another client-owned `npm test`, and completed successfully.
-
-The live evidence validator parses stream JSON and requires successful tool results. It does
-not accept a textual claim that a tool ran. It also rejects terminal summaries that report a
-blocker, incomplete work, or required next steps.
-
-The final live gate reported:
-
-```text
-PASS: validated Claude evidence for LIVE_ANALYSIS_COMPLETE
-PASS: validated Claude evidence for LIVE_FIX_COMPLETE
-PASS: validated Claude evidence for BRIDGE_E2E_COMPLETE
-PASS: three live Devin-backed Claude Code scenarios completed
-PASS: live model swe-1-7-lightning was discovered and validated by three scenarios
-```
-
-The same gate validated the network audit: only the Devin guard path was used, no internal
-Devin tool event was accepted, and the Claude egress audit remained empty.
-
-## Investigation conclusion
-
-The initial default-agent hypothesis failed because ACP permission modes do not turn the
-default Devin agent into a raw inference backend. Even `ask` mode can emit Devin-owned
-`tool_call` events. A discovered `allowed-tools: []` agent configuration was not consumed by
-`devin acp` in CLI `3000.2.17`.
-
-The working adaptation uses the official `summarizer` agent because it is structurally
-no-tools. Its fixed summarization behavior can produce intermediate prose, so the bridge
-frames requests as execution traces, detects future-action narration, performs at most one
-strict repair, and otherwise fails. Live validation also exposed transient ACP timeouts;
-the harness now spaces independent scenarios rather than weakening routing or retrying into
-another provider.
-
-## Safety record
-
-No host Claude executable, configuration, login, OAuth token, Keychain, or Anthropic API was
-used. The dedicated Docker volumes remain role-separated. No credential value is written to
-the repository or evidence output.
-
-During the early baseline, a focused test without isolated `DATA_DIR` initialized the
-repository's normal OmniRoute database at `/Users/lucasisrael/.omniroute/storage.sqlite`.
-It was not rolled back or touched again. Every bridge command now pins database and temporary
-paths under the worktree's `.sandbox` directory.
-
-## Remaining limits
-
- The no-tools backend has a summarizer system role rather than a neutral generation role.
- One client tool call per response is supported; parallel tool calls are rejected.
- ACP processes are per-turn and stateless.
- Live Devin availability can still produce explicit `502`/`504` failures.
- Images and unadvertised vision/effort/large-context capabilities remain unsupported.
--- a/docs/README.md
+++ b/docs/README.md
@@ -64,8 +64,6 @@ How the system is put together — read these to understand the runtime, code la
 - [QUALITY_GATES.md](architecture/QUALITY_GATES.md) — quality-gate scripts and CI jobs inventory.
 - [MONITORING_SECTIONS.md](architecture/MONITORING_SECTIONS.md) — monitoring/costs dashboard navigation.
 - [cluster-decisions.md](architecture/cluster-decisions.md) — optional sidecar/cluster profile decisions.
- [rust-port-research.md](rust-port-research.md) — feasibility study for porting the CPU-bound compression/tokenization core to a standalone Rust library (fork research, 2026-07-31).
- [infrastructure.md](infrastructure.md) — deployment topology (Proxmox / LXC 101 / LXC 106 / Forgejo / Docker) and push flows (fork research, 2026-07-31).

 ## reference/

--- a/docs/infrastructure.md
+++ b/docs/infrastructure.md
@@ -1,99 +0,0 @@
-# OmniRoute Deployment & Infrastructure
-
-> **Date**: 2026-07-31
-> **Scope**: Local Mac dev instance + Proxmox/LXC production layout. Context for anyone resuming work.
-
-## Topology
-
-```
-┌─────────────────────────────────────────────────────────┐
-│ HOST Proxmox 100.73.44.17 (pve-n150.tailad1b81.ts.net)   │
-│                                                           │
-│  tailscaled → holds :443 (Funnel)                         │
-│    ├─ /         → http://192.168.3.101:80/   (Forgejo)    │
-│    └─ /omniroute → http://192.168.3.106:20128/ (OmniRoute)│
-│                                                           │
-│  Docker:                                                  │
-│    ├─ openhands  :3000  (host-network, --privileged)      │
-│    └─ amnezia-awg2 :48243/udp (WireGuard, do NOT touch)   │
-└───────┬───────────────────────────────────────────────────┘
-        │ LXC (lxc-attach -n <id>)
-┌───────┴──────────┐  ┌────────────────────────────────────┐
-│ LXC 101           │  │ LXC 106 (agent-node, 192.168.3.106)│
-│ Forgejo :80       │  │  Docker:                           │
-│ (git-repositories)│  │   ├─ omniroute :20128 (data→/opt/  │
-└───────────────────┘  │  │        omniroute/data)          │
-                       │  │   └─ openhands :8000/18000/8002 │
-                       │  │      (OLD duplicate — DELETE)   │
-                       │  ├─ systemd project-history :43128 │
-                       │  ├─ component-vault :43133 (old)   │
-                       │  └─ iptables: INPUT DROP + ACCEPT  │
-                       │     for 22,20128,43128,43133,...   │
-                       └────────────────────────────────────┘
-```
-
-## Components
-
-| Component | Where | Details |
-|---|---|---|
-| **Tailscale** | host | `tailscale serve` with Funnel; certs `/var/lib/tailscale/certs/pve-n150.*` |
-| **Caddy** | — | **not installed** (no package, no Caddyfile) — HTTPS handled by Tailscale Serve |
-| **OpenHands** | host, docker | image `openhands:fixed` (`4e631813f208`), host-network, privileged; DB in `/opt/openhands/workspace/.openhands-state`; created via `docker run -e LLM_MODEL=ds/deepseek-v4-flash -e LLM_BASE_URL=http://192.168.3.106:20128/v1 -e LLM_API_KEY=sk-d146...` (backup: `/opt/openhands/container-config-backup.txt`) |
-| **OmniRoute** | LXC 106, docker | image `diegosouzapw/omniroute`, mount `/opt/omniroute/data→/app/data`, cmd `node dev/run-standalone.mjs`; sources/build: `/opt/omniroute-build` (git + Dockerfile + compose) |
-| **Forgejo** | LXC 101 | git server, `http://192.168.3.101`, external `https://pve-n150.tailad1b81.ts.net/` (path prefix `/git/`; Gitea 15.0.1) |
-| **project-history** | LXC 106, systemd | Rust, `/opt/project-history` (src + binary + data), port 43128 |
-
-## Pushing changes
-
-**1. To Forgejo (any session):** remote `http://192.168.3.101/egorich/<repo>.git`.
-   From Mac: `https://egorich:<token>@pve-n150.tailad1b81.ts.net/git/egorich/<repo>.git`
-   ⚠️ URL-encode `@` in the password as `%40`.
-
-**2. To OpenHands (code/fixes):** image built via `docker commit`, so change = edit inside container + commit image:
-```bash
-docker exec -it openhands bash          # edit /app/openhands/...
-docker commit openhands openhands:fixed  # fix patch into image
-docker restart openhands                 # apply
-```
-Env config (model, key, CORS): recreate container with same command from `/opt/openhands/container-config-backup.txt` + new `-e`.
-
-**3. To OmniRoute (LXC 106):**
-```bash
-lxc-attach -n 106 -- bash
-cd /opt/omniroute-build            # git pull / checkout pr/fix-pack
-docker compose -f docker-compose.prod.yml build
-docker compose -f docker-compose.prod.yml up -d
-```
-Data (providers, keys) in `/opt/omniroute/data` — survives rebuild (volume).
-
-**4. To project-history (LXC 106):**
-```bash
-lxc-attach -n 106 -- bash
-cd /opt/project-history            # or git clone from Forgejo (no .git there!)
-# edit src/, then:
-cargo build --release
-systemctl restart project-history
-curl http://127.0.0.1:43128/api/health
-```
-
-## Exposing a new path over HTTPS
-```bash
-tailscale serve --bg --set-path /history http://192.168.3.106:43128/
-```
-
-## Do NOT touch
- iptables in LXC 106 (INPUT DROP, persistent rules)
- `omniroute` (needed by OpenHands)
- `amnezia-awg2`
- DB `data/project_history.sqlite3`
-
-## Access notes (Mac)
-
- SSH to Proxmox/LXC **does not work** from this Mac (Tailscale is stopped here; ports time out).
- Forgejo API works over `https://pve-n150.tailad1b81.ts.net/git/api/v1/` (Basic auth `egorich`).
- Everything else reachable only from the Proxmox host / LXC sessions.
-
-## Forgejo repo (created 2026-07-31)
-
- `egorich/OmniRoute` — branches `pr/fix-pack` (PR-ready), `feat/personal-build` (full history)
- GitHub PR: https://github.com/diegosouzapw/OmniRoute/pull/9058
--- a/docs/providers/CHATGPT_WEB_CODEX.md
+++ b/docs/providers/CHATGPT_WEB_CODEX.md
@@ -1,94 +0,0 @@
-# ChatGPT Web (Codex)
-
-`ChatGPT Web (Codex)` ist ein zusätzlicher Provider. Der bestehende Provider
-`ChatGPT Web (Plus/Pro)` bleibt für normale Chats, Bilder und dessen bisherige
-Tool-Emulation unverändert.
-
-## Voraussetzungen
-
- ein vollständiger Cookie-Header einer angemeldeten ChatGPT-Sitzung;
- Chrome oder Chromium bei npm-, systemd- und PM2-Installationen;
- beim Docker-Profil `web` der interne Chromium-Dienst aus `docker-compose.yml`;
- ein OpenAI-Tunnel und ein ChatGPT-Custom-Connector für lokale Codex-Tools.
-
-Der Tunnel ist nur für Tool-Runden nötig. `pro` ist read-only und benötigt keinen
-lokalen Tool-Connector.
-
-## Einrichtung in der Weboberfläche
-
-1. Öffne den Provider `ChatGPT Web (Codex)` und füge eine Connection hinzu.
-2. Füge den vollständigen ChatGPT-Cookie, die Tunnel-ID, den Runtime-Key und den
-   Namen des Custom Connectors ein.
-3. Starte die Prüfung. OmniRoute öffnet headless einen Temporary Chat und erkennt
-   dabei auch, ob `pro` für das Konto verfügbar ist.
-4. Speichere die Connection. OmniRoute ersetzt den eingegebenen Cookie durch den
-   geprüften Playwright-Storage-State und speichert ihn zusammen mit dem Runtime-Key
-   über die verschlüsselte Credential-Abstraktion.
-
-Der rohe Cookie wird nach erfolgreichem Speichern nicht zusätzlich aufbewahrt.
-Wenn die Sitzung abläuft, öffne die Connection, gib einen frischen vollständigen
-Cookie ein und prüfe sie erneut. Der Doctor-Status im Edit-Dialog zeigt Browser,
-Storage-State, Anmeldung, Temporary Chat, Tunnel, Connector und Tool-Roundtrip
-getrennt an.
-
-## Modelle und Combos
-
-Die festen Modelle sind:
-
- `chatgpt-web-codex/instant`
- `chatgpt-web-codex/medium`
- `chatgpt-web-codex/high`
- `chatgpt-web-codex/extra-high`
- `chatgpt-web-codex/pro`
-
-Füge eines davon wie jedes andere Modell zu einer Combo hinzu. Die Codex-App
-sendet nur den Combo-Namen als `model` an den normalen Responses-Endpunkt
-`/v1/responses`. Es gibt keinen Sonderendpoint und keinen Codex-Modus-Schalter.
-
-`pro` führt keine lokalen Tools aus. Ein erzwungenes Tool macht dieses Combo-Ziel
-inkompatibel; bei optionalen Tools läuft der Turn read-only und meldet diese
-Einschränkung als Commentary.
-
-## Sicherheitsmodell
-
- Der native Pfad verlangt einen Responses-Request, einen erkannten Codex-Client
-  sowie zusammenpassende Thread- und Turn-Identitäten.
- Workspace, Sandbox, Approval-Policy und Toolkatalog stammen aus der nativen
-  Codex-Hülle. Freier Prompttext ist dafür keine Autorität.
- ChatGPT erhält pro Turn nur eine kurzlebige Capability. Der MCP-Broker akzeptiert
-  ausschließlich Tools, die Codex in genau diesem Turn angeboten hat.
- Das automatische Bestätigen von „Allow once“ gibt nur den Tool-Wunsch an Codex
-  zurück. Codex allein entscheidet über Freigabe und Ausführung.
- Vor dem ersten Output darf die Combo auf ein anderes kompatibles Ziel fallen.
-  Danach bleiben Provider, Modell, Connection und Browserturn bis zum Abschluss
-  gepinnt.
- Cookies, Runtime-Keys, Storage-State und Capability-Tokens erscheinen nicht in
-  Providerantworten oder Request-Logs.
-
-## Headless VPS und Docker
-
-Bei npm-, systemd- und PM2-Betrieb erkennt OmniRoute übliche Chrome- und
-Chromium-Pfade. Alternativ kann `CHATGPT_WEB_CODEX_CHROME_PATH` gesetzt werden.
-
-Das Docker-Profil `web` startet `chatgpt-web-codex-browser` im internen
-Compose-Netz. Sein CDP-Port wird nicht auf dem Host veröffentlicht. Das geschützte
-Profilvolume bleibt getrennt vom OmniRoute-Datenvolume und der Browser erhält
-ausreichend Shared Memory. Der interne CDP-Proxy lauscht nur im Compose-Netz auf
-Port `9223`; Chrome selbst bleibt im Sidecar an Loopback gebunden.
-
-Eine Supervisor-Lease unter `DATA_DIR` verhindert, dass mehrere OmniRoute-Prozesse
-denselben Tunnel- und Brokerzustand besitzen. Ein Konflikt erscheint im Doctor.
-
-## Interaktive Wiederherstellung
-
-Der normale Pfad ist vollständig headless. Wenn ChatGPT eine interaktive
-Anmeldung oder Challenge verlangt, kann die bestehende VNC-Browser-Infrastruktur
-als Recovery-Weg verwendet werden. Browser-UI und CDP dürfen dabei nur über
-Loopback, eine authentifizierte Managementverbindung oder einen SSH-Tunnel
-erreichbar sein; noVNC bleibt im normalen Betrieb deaktiviert.
-
-## WebSocket-Fallback
-
-Enthält eine Combo `ChatGPT Web (Codex)`, fordert die Responses-WebSocket-Brücke
-vor der Upstream-Verbindung den HTTP/SSE-Fallback an. Die eigentliche Übertragung
-erfolgt dann über `/v1/responses`.
--- a/docs/reference/ENVIRONMENT.md
+++ b/docs/reference/ENVIRONMENT.md
@@ -377,14 +377,6 @@ Controls how OmniRoute discovers and launches CLI sidecars (Claude Code, Codex,
 | `CLI_QODER_BIN`           | `qoder`     | `src/shared/services/cliRuntime.ts`                 | Custom path to Qoder CLI binary.                                                                                                                                               |
 | `CLI_QWEN_BIN`            | `qwen`      | `src/shared/services/cliRuntime.ts`                 | Custom path to the Qwen Code CLI binary.                                                                                                                                       |
 | `CLI_DEVIN_BIN`           | `devin`     | `open-sse/executors/devin-cli.ts`                   | Custom path to the Devin CLI binary (v3.8.0). Used by the Windsurf/Devin executor.                                                                                             |
-| `CLI_DEVIN_AGENTIC_BIN`   | `devin`     | `open-sse/executors/devin-cli-agentic.ts`           | Agentic bridge-only Devin CLI override. The executor accepts only the local ACP stdio upstream.                                                                               |
-| `DEVIN_AGENTIC_HOME`      | _(required)_ | `open-sse/executors/devin-cli-agentic.ts`           | Absolute isolated home for the agentic Devin subprocess; accepted bridge paths are `/home/bridge` and task-local `.sandbox` paths.                                            |
-| `DEVIN_AGENTIC_ACP_TIMEOUT_MS` | `120000` | `open-sse/executors/devin-cli-agentic.ts`           | Maximum duration of one Devin ACP turn before the bridge terminates the child and returns an explicit timeout.                                                                |
-| `DEVIN_BRIDGE_MODEL` | `devin-cli-agentic/swe-1-7` | `docker/devin-bridge/compose.yml` | Main Claude Code model alias for the isolated bridge. The live harness replaces the example with a model returned by the current Devin account. |
-| `DEVIN_BRIDGE_SONNET_MODEL` | `DEVIN_BRIDGE_MODEL` | `docker/devin-bridge/compose.yml` | Isolated bridge alias used when Claude Code requests its Sonnet default. |
-| `DEVIN_BRIDGE_OPUS_MODEL` | `DEVIN_BRIDGE_MODEL` | `docker/devin-bridge/compose.yml` | Isolated bridge alias used when Claude Code requests its Opus default. |
-| `DEVIN_BRIDGE_HAIKU_MODEL` | `DEVIN_BRIDGE_MODEL` | `docker/devin-bridge/compose.yml` | Isolated bridge alias used when Claude Code requests its Haiku default. |
-| `DEVIN_BRIDGE_SUBAGENT_MODEL` | `DEVIN_BRIDGE_MODEL` | `docker/devin-bridge/compose.yml` | Isolated bridge alias used for Claude Code subagents. |
 | `AUGGIE_BIN`              | `auggie`    | `open-sse/executors/auggie.ts`                      | Absolute-path override for the Augment (Auggie) CLI binary used by the local `auggie` provider. Falls back to `CLI_AUGGIE_BIN`, then a PATH lookup.                            |
 | `CLI_AUGGIE_BIN`          | `auggie`    | `open-sse/executors/auggie.ts`                      | Alias override for the Augment (Auggie) CLI binary path (checked after `AUGGIE_BIN`).                                                                                          |
 | `HERMES_HOME`             | `~/.hermes` | `src/lib/cli-helper/config-generator/hermesHome.ts` | Hermes Agent home directory where OmniRoute reads/writes the Hermes CLI config. Matches the env var the Hermes PowerShell installer sets on Windows (`%LOCALAPPDATA%\hermes`). |
@@ -1307,16 +1299,3 @@ Used by `src/lib/vncSession/manifest.ts` to configure Docker-based headless Chro
 | `OMNIROUTE_VNC_READY_MS`              | `45000`                       | `src/lib/vncSession/manifest.ts`  | Browser readiness timeout (ms).                                                |
 | `OMNIROUTE_VNC_HARVEST_MS`            | `20000`                       | `src/lib/vncSession/manifest.ts`  | Harvest/cleanup timeout (ms).                                                  |
 | `VIBEPROXY_DATA_DIR`                  | _(unset)_                     | `open-sse/services/notionThreadSessions.ts` | Directory for Notion thread session persistence.                               |
-
-### ChatGPT Web (Codex)
-
-Globale Defaults für den headless Browser und den ausgehenden Tool-Tunnel. Im Dashboard gesetzte Connection-Werte haben Vorrang.
-
-| Variable                             | Default                          | Source File                                      | Description                                                                 |
-| ------------------------------------ | -------------------------------- | ------------------------------------------------ | --------------------------------------------------------------------------- |
-| `CHATGPT_WEB_CODEX_CHROME_PATH`      | _(auto-detect)_                  | `open-sse/executors/chatgpt-web-codex.ts`        | Expliziter Chrome-/Chromium-Pfad für npm-, systemd- und PM2-Betrieb.       |
-| `CHROME_PATH`                        | _(auto-detect)_                  | `open-sse/executors/chatgpt-web-codex.ts`        | Gemeinsamer Fallback für einen expliziten Chrome-/Chromium-Pfad.            |
-| `CHATGPT_WEB_CODEX_CDP_URL`          | _(unset)_                        | `open-sse/executors/chatgpt-web-codex.ts`        | Interner CDP-Endpunkt; Docker verwendet den Sidecar auf Port `9223`.        |
-| `CHATGPT_WEB_CODEX_TUNNEL_ID`        | _(unset)_                        | `open-sse/executors/chatgpt-web-codex.ts`        | Globale OpenAI-Tunnel-ID für lokale Codex-Tool-Runden.                      |
-| `CHATGPT_WEB_CODEX_RUNTIME_KEY`      | _(unset)_                        | `open-sse/executors/chatgpt-web-codex.ts`        | Globaler Tunnel Runtime-Key; niemals in Logs ausgeben.                      |
-| `CHATGPT_WEB_CODEX_CONNECTOR_NAME`   | _(unset)_                        | `open-sse/executors/chatgpt-web-codex.ts`        | Name des ChatGPT-Custom-Connectors für die MCP-Brücke.                      |
--- a/docs/rust-port-research.md
+++ b/docs/rust-port-research.md
@@ -1,201 +0,0 @@
-# Rust Port Research — OmniRoute Compute Engine Extraction
-
-> **Date**: 2026-07-31
-> **Status**: Feasibility study (research only, no code written yet)
-> **Author**: Egor (fork `Egorich-print/OmniRoute`, branch `feat/personal-build`)
-> **Reviewer feedback**: ChatGPT architecture review incorporated below
-
-## TL;DR
-
-OmniRoute's latency-critical path is the **CPU-bound deterministic compression + tokenization layer** — not the backend plumbing. Port these pure algorithms to a **standalone Rust library** (`compression-core`) with a thin N-API binding as the primary integration path. Ship in this order: **tiktoken → ionizer → headroom → caveman → RTK** (RTK last — thousands of rules, highest risk). Golden-test JS↔Rust byte-in-byte before replacing anything.
-
-Target: an independent OSS crate (`ai-compression-engine` / `context-engine`) usable by OmniRoute, OpenCode, Cline, Roo, and any AI proxy — not `omniroute-rust`.
-
---
-
-## 1. Measured Baseline
-
-Benchmark on the dev Mac (2026-07-31):
-
-| Operation | Input | Cost | Notes |
-|---|---|---|---|
-| `countTextTokens()` — js-tiktoken `cl100k_base` | 230K chars (~57K tokens) | **37.9 ms/count** | Runs per chat request |
-| Rust `tiktoken-rs` (est.) | same | ~1-3 ms | 10-30x faster |
-
-The token counter runs on **every** chat request. Compression runs per-request when the conversation exceeds budget.
-
-## 2. Hot Path Map (chat streaming request)
-
-All operations below are synchronous and block the Node event loop.
-
-| # | Operation | File:Line | CPU | Freq | Rust portability |
-|---|---|---|---|---|---|
-| 1 | SSE chunk JSON parse/stringify | `open-sse/utils/stream.ts:2391` | Expensive | per-chunk | High (serde) |
-| 2 | Tiktoken token counting | `src/shared/utils/tiktokenCounter.ts:54` | Moderate | per-req | **High** |
-| 3 | RTK compression (regex/line filtering) | `open-sse/services/compression/engines/rtk/index.ts:525` | Expensive | per-req | **High** |
-| 4 | Headroom tabular compaction | `open-sse/services/compression/engines/headroom/index.ts:114` | Moderate | per-req | High |
-| 5 | Request format translation | `open-sse/translator/registry.ts:23` | Moderate | per-req | Moderate |
-| 6 | SQLite usage persistence | `src/lib/usage/usageHistory.ts:675` | Moderate | per-req | Low (DB coupling) |
-| 7 | PII sanitization (SSE transform) | `open-sse/handlers/chatCore/streamingPipeline.ts:91` | Moderate | per-chunk | High (regex) |
-| 8 | Memory/skills injection (context merge) | `open-sse/handlers/chatCore.ts:1065` | Cheap | per-req | Moderate |
-| 9 | Idempotency/request hashing | `open-sse/handlers/chatCore.ts:608` | Cheap | per-req | High (crypto) |
-| 10 | Usage estimation (fallback counting) | `open-sse/utils/usageTracking.ts:560` | Cheap | per-chunk | High |
-
-**Where time goes (estimate):** network wait ≫ CPU (compression + tiktoken) > DB > per-chunk overhead.
-
-## 3. Compression Engine Profiles
-
-### Tiktoken counter — `src/shared/utils/tiktokenCounter.ts` (62 LOC + lib)
- Library: **js-tiktoken** v1.0.21 — pure JS port of tiktoken, no WASM.
- Mechanism: pre-computed BPE rank tables shipped as base64 binary blobs (~6 MB across 6 rank files); byte-pair merge on `TextEncoder` UTF-8 byte arrays.
- Encodings used: `cl100k_base` (default), `o200k_base` (Codex).
- Node deps: `TextEncoder`/`TextDecoder` (built-ins), `base64-js`.
- **Verdict**: pure deterministic BPE → ideal Rust port (`tiktoken-rs` supports cl100k + o200k natively).
-
-### RTK — `open-sse/services/compression/engines/rtk/` (20 files, ~4000 LOC)
- `index.ts` 706, `commandDetector.ts` 482, `filterLoader.ts` 332, `tomlCompatibility.ts` 334, `learn.ts` 290, `lineFilter.ts` + more.
- Deterministic rule engine: regex-based line classification, keep-patterns for code blocks/JSON, folding/merging rules, tool-call-aware filtering (bash vs non-shell tools).
- Called per-request on the whole messages array; **sync** (no awaits in the core).
- **Verdict**: port last. High effort, high risk — but regex crate gives linear-time matching (no backtracking blowups) and output equivalence is testable via golden tests.
-
-### Headroom — `engines/headroom/` (~550 LOC)
- "Tabular compaction": replaces array-of-objects message content with compact columnar blocks (```gcf-generic ... ```). Lossless, conservative (only when strictly smaller), never touches system messages.
- **Verdict**: pure deterministic, port after ionizer.
-
-### Ionizer — `engines/ionizer/` (124 + 205 LOC)
- Lossy statistical sampling of oversized homogeneous JSON arrays: keeps schema + error rows + first/last rows + seeded uniform middle sample.
- Deterministic: FNV-1a hash + mulberry32 PRNG (no Math.random).
- **Verdict**: trivial port, nearly zero risk — do second.
-
-### Caveman — `engines/cavemanAdapter.ts` + `caveman.ts` (~250 LOC)
- Regex rule-based compaction for `standard` mode.
- **Verdict**: port after headroom.
-
-### Other engines (not first-wave)
- `relevance/` — keyword scoring (no embeddings/network).
- `session-dedup/` — dedupe via hash, per-request.
- `llm/`, `llmlingua/` — **LLM-dependent (network)**, opt-in, NOT portable to pure CPU core.
- `ccr/` (Content-Addressable Recovery) — stores full original for reconstruction.
-
-## 4. Orchestrator
-
- Entry points: `applyCompression` (sync) / `applyCompressionAsync` (async) — `open-sse/services/compression/strategySelector.ts:255` / `:459`.
- Exported via `open-sse/services/compression/index.ts:86-91`.
- Mode dispatch: `off | rtk | codex-responses | omniglyph | lite | stacked | standard | aggressive | ultra`.
- `stacked` mode runs engines sequentially by `stackPriority` (rtk=10, ionizer=13, headroom=15, ...).
- Called from `chatCore.ts` via dynamic `import()`; sync CPU-bound → blocks event loop.
- **Verdict**: single pure function `(messages, budget, config) → (messages, metrics)` — clean extraction surface for a Rust core.
-
-## 5. Architecture Decision (revised per ChatGPT review)
-
-### Recommendation: Rust library + thin N-API binding (NOT sidecar-first)
-
-```
-crates/
-  compression-core/     ← pure algorithms, no I/O, no OmniRoute knowledge
-    src/
-      tiktoken/         (cl100k_base, o200k_base)
-      ionizer/
-      headroom/
-      caveman/
-      rtk/              (last)
-  napi/                 ← N-API binding (primary integration path, in-process)
-  sidecar/              ← optional HTTP/Unix-socket server over the same core
-  cli/                  ← CLI harness (bench, golden tests)
-```
-
-**Why N-API first (vs Unix-socket sidecar):**
- Every sidecar call pays serialize→socket→deserialize→compute→serialize→deserialize.
- For a 2 ms token count, IPC overhead becomes a large fraction of the call.
- N-API is in-process: zero serialization on the hot path, no process management.
- Keep sidecar only if process isolation / multi-language integration is actually needed.
-
-**Wire format if sidecar is later added:** `bincode` / `postcard` / MessagePack — NOT JSON. Messages are large; JSON round-trip is wasted work.
-
-### Independence from OmniRoute
-
-Make it a **standalone OSS project**: `ai-compression-engine` or `context-engine`.
-
-```rust
-// core API surface
-pub fn compress(messages: &[Message], config: &CompressionConfig) -> CompressionResult;
-pub fn count_tokens(encoding: Encoding, text: &str) -> u64;
-```
-
-No OmniRoute imports anywhere in `compression-core`. Consumers: OmniRoute, OpenCode, Cline, Roo, any AI proxy.
-
-## 6. Phased Roadmap (revised)
-
-| Phase | Work | Effort | Risk |
-|---|---|---|---|
-| 0 | Baseline benchmark (latency, event-loop blocking, compression %) | 1 day | — |
-| 1 | Extract `compression-core` crate + port **tiktoken**, golden tests | 2-3 days | Minimal |
-| 2 | Port **ionizer** | 1-2 days | Nearly zero |
-| 3 | Port **headroom** | 2-3 days | Low |
-| 4 | Port **caveman** | 2 days | Low |
-| 5 | Port **RTK** (biggest, do when harness proven) | 5-7 days | High |
-| 6 | N-API binding as primary path; feature-flag integration with fallback to JS | 2-3 days | Low |
-| 7 | Optional sidecar (isolation/multi-lang) | 2 days | Low |
-| 8 | **Second wave**: SSE parser + OpenAI/Claude/Gemini translators (per-chunk hot path) | TBD | High |
-
-**Total to full compression replacement: ~2-3 weeks.** First measurable win (tiktoken): 2-3 days.
-
-## 6a. Definition of Done (added per review)
-
-| Phase | DoD |
-|---|---|
-| Tokenizer (tiktoken) | Full match with JS on golden tests (0% divergence, byte-in-byte); benchmark < 5 ms per 57K tokens |
-| Ionizer | Output matches JS; no perf regression (or ≥ 10x speedup) |
-| Headroom | Identical compression result (same % savings, same columnar blocks) |
-| RTK | Byte-in-byte equivalence on full dialogue set (500 fixtures); unit tests on rule patterns |
-| Integration (N-API) | JS↔Rust switch via one setting/env; JS fallback when unavailable; zero risk to current deployment |
-| Translation primitives (wave 2) | Byte-in-byte SSE chunk equivalence before/after; no TTFB increase |
-
-Progress is measurable per phase and each phase is independently verifiable.
-
-## 7. Golden Testing (mandatory)
-
-```
-fixtures/
-  conversation1.json
-  ...
-  conversation500.json
-
-JS  run → output_a.json
-Rust run → output_b.json
-assert_eq(output_a, output_b)   // byte-in-byte, 100% required
-```
-
- Until 100% match, **do not** switch the runtime to Rust.
- This makes even the RTK rewrite safe.
- Also validates tiktoken rank tables (JS vs `tiktoken-rs`) on 100+ varied texts.
-
-## 8. Fallback Strategy (for OmniRoute integration)
-
- New env: `OMNIROUTE_COMPRESSION_SIDECAR` (optional, off by default) or N-API availability check.
- Node code stays untouched; a client wrapper (`src/lib/compression-rust/`) tries Rust → falls back to existing JS path (`applyRtkCompression`, `applyCompression`).
- Zero risk to the current deployment.
-
-## 9. Deferred / Stay-in-JS
-
- Dashboard (Next.js App Router, ~200 pages) — never ported.
- Skills, memory, MCP, guardrail management, quotas, combo config — not latency-critical.
- `llm`/`llmlingua` compression engines — network/LLM-dependent, stay in JS.
- SQLite usage persistence — DB-coupled, stays.
-
-## 10. Risks
-
-| Risk | Mitigation |
-|---|---|
-| JS↔Rust behavioral divergence | Golden tests (500 fixtures, byte-in-byte) |
-| RTK `tomlCompatibility.ts` / `learn.ts` / `filterLoader.ts` | Port whole submodules (deterministic); rules format 1:1 |
-| Rust toolchain on LXC 106 (aarch64?) | Check `rustup`/`cargo`; multi-stage Docker build or cross-compile on Mac |
-| N-API ABI mismatch (Node 26) | Use `napi-rs` (prebuilt binaries, Node-version tolerant) |
-| Sidecar JSON overhead | bincode/postcard if sidecar is adopted |
-
-## 11. Open Questions
-
-1. Cargo/rustup present on LXC 106, or cross-compile from Mac?
-2. Confirm N-API as primary integration path (vs sidecar)?
-3. Want the baseline benchmark included in the roadmap before porting?
-4. Repo home for `compression-core`: new repo (`ai-compression-engine`) or `crates/` inside OmniRoute fork first?
--- a/docs/superpowers/plans/2026-07-27-devin-claude-bridge.md
+++ b/docs/superpowers/plans/2026-07-27-devin-claude-bridge.md
@@ -1,252 +0,0 @@
-# Devin Claude Bridge Implementation Plan
-
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
-
-**Goal:** Build a fail-closed `devin-cli-agentic` provider that serves local Anthropic Messages requests through Devin CLI ACP stdio while preserving Claude Code tool-use semantics.
-
-**Architecture:** Add a separate Claude-format provider and executor instead of changing the existing OpenAI-format `devin-cli` summarizer. Keep parsing, prompt serialization, Anthropic response rendering, and ACP process handling in focused files under `open-sse/executors/devin-agentic/`, then wire them into the existing provider and executor registries.
-
-**Tech Stack:** TypeScript ES modules, Node child process stdio, Anthropic Messages JSON/SSE, JSON-RPC 2.0 ACP, Node test runner.
-
---
-
-### Task 1: Agentic Bridge Core
-
-**Files:**
- Create: `open-sse/executors/devin-agentic/types.ts`
- Create: `open-sse/executors/devin-agentic/serializer.ts`
- Create: `open-sse/executors/devin-agentic/toolParser.ts`
- Create: `open-sse/executors/devin-agentic/anthropicResponse.ts`
- Test: `tests/unit/executor-devin-cli-agentic-core.test.ts`
-
- [ ] **Implement and prove serialization, parsing, validation, and Anthropic rendering**
-
-Interfaces:
-
-```ts
-export function serializeAnthropicForDevin(body: unknown): DevinPrompt;
-export function parseDevinToolRequest(text: string, tools: AnthropicTool[]): ParsedToolRequest | null;
-export function buildClaudeTextResponse(args: ClaudeResponseArgs): Record<string, unknown>;
-export function buildClaudeToolUseResponse(args: ClaudeToolUseArgs): Record<string, unknown>;
-export function buildClaudeSseFrames(message: Record<string, unknown>): string;
-```
-
-Invariants:
-
- Preserve `text`, `tool_use`, `tool_result`, `thinking`, and `redacted_thinking`.
- Reject `image` with a clear error.
- Reject unknown content block types.
- Allow only one tool request per model turn.
- Validate tool arguments against object JSON Schema with `required`, `type`, `properties`, `additionalProperties`, `enum`, `items`, and scalar types.
- Generate deterministic ids from tool name and canonicalized arguments.
-
-Run: `node --import tsx/esm --test tests/unit/executor-devin-cli-agentic-core.test.ts`
-Expected: core tests pass after dependencies are installed.
-
-### Task 2: ACP Executor And Provider Wiring
-
-**Files:**
- Create: `open-sse/executors/devin-cli-agentic.ts`
- Modify: `open-sse/executors/index.ts`
- Create: `open-sse/config/providers/registry/devin-cli-agentic/index.ts`
- Modify: `open-sse/config/providers/index.ts`
- Test: `tests/unit/executor-devin-cli-agentic-acp.test.ts`
-
- [ ] **Implement and prove fail-closed ACP execution**
-
-Behavior:
-
- `buildUrl()` returns `devin://acp/stdio`.
- `buildHeaders()` returns `{}`.
- `execute()` spawns only `devin acp` by default or the explicit `CLI_DEVIN_AGENTIC_BIN`/`CLI_DEVIN_BIN` override.
- The child environment removes Anthropic and Claude routing credentials before spawn.
- The executor sends `initialize`, `session/new`, and `session/prompt`.
- The executor collects `agent_message_chunk` text and `session/prompt` final result.
- Non-streaming Claude clients receive native Anthropic JSON.
- Streaming Claude clients receive native Anthropic SSE lifecycle frames.
- Spawn failure, ACP error, timeout, and early exit produce non-2xx responses with sanitized messages.
-
-Run: `node --import tsx/esm --test tests/unit/executor-devin-cli-agentic-acp.test.ts`
-Expected: ACP mock tests pass after dependencies are installed.
-
-### Task 3: Isolation Scripts And Documentation
-
-**Files:**
- Create: `scripts/devin-bridge/verify-anthropic-isolation`
- Create: `scripts/devin-bridge/test-unit`
- Create: `scripts/devin-bridge/launch`
- Create: `docs/DEVIN_CLAUDE_BRIDGE.md`
- Modify: `.gitignore`
-
- [ ] **Implement offline guardrails and operator docs**
-
-Behavior:
-
- `verify-anthropic-isolation` fails if `CLAUDE_CONFIG_DIR` is missing, points outside an isolated path, or if Anthropic routing env vars are present.
- `test-unit` runs the focused unit tests.
- `launch` refuses to start unless `ENABLE_LIVE_DEVIN_TESTS=1` for live Devin or `DEVIN_BRIDGE_OFFLINE=1` for offline mock mode.
- Documentation distinguishes tested offline behavior from live Devin opt-in behavior.
-
-Run: `./scripts/devin-bridge/verify-anthropic-isolation` with explicit isolated env.
-Expected: exits 0 with isolated env and non-zero without it.
-
-### Task 4: Verification
-
-**Files:**
- No additional source files.
-
- [ ] **Run proportional checks and capture real output**
-
-Commands:
-
-```bash
-./scripts/devin-bridge/test-unit
-npm test
-```
-
-Expected in this workspace before installing dependencies: both commands fail with `ERR_MODULE_NOT_FOUND` for `tsx`. Expected after `npm install`: focused tests pass; `npm test` outcome must be reported from real output.
-
-### Task 5: Close Core Security And Protocol Gaps
-
-**Files:**
- Modify: `open-sse/executors/devin-cli-agentic.ts`
- Modify: `open-sse/executors/devin-agentic/*.ts`
- Modify: `tests/unit/executor-devin-cli-agentic-*.test.ts`
-
- [ ] **Prove environment allowlisting, response-id correlation, strict standalone tool envelopes, unique ids, bounded repair, size limits, cancellation cleanup, sanitized errors, and explicit `devin://acp/stdio` validation**
-
-Run with `HOME`, `DATA_DIR`, and `SQLITE_FILE` under `.sandbox`; expected: all focused tests pass and an outside-path test fails closed.
-
-### Task 6: Build Reproducible Containers And Network Guard
-
-**Files:**
- Create: `docker/devin-bridge/Dockerfile`
- Create: `docker/devin-bridge/compose.yml`
- Create: `docker/devin-bridge/network-guard/*`
- Create: `docker/devin-bridge/mock-devin/*`
- Create: `.env.devin-bridge.example`
-
- [ ] **Pin Claude Code 2.1.220 and Devin CLI 3000.2.17, create non-root offline/live profiles, separate auth/config volumes, explicit env allowlist, no host credential mounts, and denied-domain telemetry**
-
-Run: `docker compose -f docker/devin-bridge/compose.yml --profile offline config`; expected: no forbidden mounts/env inheritance and only internal runtime networks.
-
-### Task 7: Deliver Isolation And Operator Scripts
-
-**Files:**
- Create/modify: `scripts/devin-bridge/{build,test-unit,test-contract,test-e2e-mock,verify-anthropic-isolation,login-devin,test-live-devin,launch,clean}`
-
- [ ] **Make every command idempotent, sandbox-scoped, fail-closed, and secret-safe**
-
-Run: `./scripts/devin-bridge/verify-anthropic-isolation`; expected: positive offline proof passes and each deliberately removed guard returns non-zero.
-
-### Task 8: Real Claude Code Offline E2E
-
-**Files:**
- Create: `tests/fixtures/devin-bridge/e2e-workspace/*`
- Create: `tests/e2e/devin-claude-bridge.e2e.*`
-
- [ ] **Run pinned Claude Code in the offline container through local `/v1/messages` and mock ACP, proving CLAUDE.md, skill, command, hook, Read/Edit/Bash, tests, multi-turn continuation, and no Anthropic traffic**
-
-Run: `./scripts/devin-bridge/test-e2e-mock`; expected: workspace diff and tests prove Claude Code executed tools while mock Devin only requested them.
-
-### Task 9: Regression, Documentation, Live Gate, And Delivery
-
-**Files:**
- Modify: `docs/DEVIN_CLAUDE_BRIDGE.md`
- Create: `docs/DEVIN_CLAUDE_BRIDGE_PROGRESS.md`
-
- [ ] **Run focused suites, typecheck, lint, build, docs checks, offline E2E, and isolation proof with fresh output; then run live only after official in-container Devin login**
-
-If login is unavailable, record live as not tested and expose exactly `./scripts/devin-bridge/login-devin` followed by `./scripts/devin-bridge/test-live-devin`. Commit each reversible unit; do not merge or publish until all offline critical checks are green.
-
-### Task 10: Close The Authenticated Live Runtime
-
-**Files:**
- Modify: `open-sse/executors/devin-cli-agentic.ts`
- Modify: `docker/devin-bridge/compose.yml`
- Create: `docker/devin-bridge/network-guard/policy.mjs`
- Modify: `docker/devin-bridge/network-guard/proxy.mjs`
- Modify: `scripts/devin-bridge/select-live-model.mjs`
- Modify: `scripts/devin-bridge/common`
- Modify: `scripts/devin-bridge/login-devin`
- Modify: `scripts/devin-bridge/test-live-devin`
- Modify: `scripts/devin-bridge/verify-anthropic-isolation`
- Modify: `tests/unit/executor-devin-cli-agentic-acp.test.ts`
- Create: `tests/unit/devin-bridge-live-runtime.test.ts`
-
- [ ] **Implement and prove the authenticated network, auth, and catalog boundaries with block-level TDD**
-
-Invariants:
-
- The ACP child receives proxy variables only when `DEVIN_BRIDGE_PROXY_URL` is exactly
-  `http://network-guard:8080`; arbitrary inherited proxy and credential variables stay absent.
- The guard permits suffixes `.devin.ai` and `.cognition.ai`, exact hosts
-  `server.codeium.com` and `unleash.codeium.com`, and nothing else.
- Claude services cannot mount `devin-auth`; non-Claude services cannot mount the Claude config.
- A zero exit from `devin auth status` is insufficient when output contains a server-fetch failure.
- `family_uid: swe-1.7-lightning` resolves to catalog id `swe-1-7-lightning`; unknown normalized
-  values fail instead of becoming model ids.
- Login uses the official manual-token flow so no container loopback callback is required.
-
-Run:
-
-```bash
-./scripts/devin-bridge/test-unit
-node --import tsx/esm --test tests/unit/devin-bridge-live-runtime.test.ts
-./scripts/devin-bridge/verify-anthropic-isolation --static
-```
-
-Expected: focused tests and static isolation pass; deliberate untrusted proxy, host, mount, auth
-status, and model fixtures fail closed.
-
- [ ] **Commit the reversible live-runtime repair**
-
-```bash
-git add open-sse/executors/devin-cli-agentic.ts docker/devin-bridge \
-  scripts/devin-bridge tests/unit/devin-bridge-live-runtime.test.ts \
-  tests/unit/executor-devin-cli-agentic-acp.test.ts
-git commit -m "fix: close Devin bridge live runtime gaps"
-```
-
-### Task 11: Prove Offline And Live Completion
-
-**Files:**
- Modify: `docker/devin-bridge/run-claude-live-e2e.sh`
- Modify: `docs/DEVIN_CLAUDE_BRIDGE.md`
- Modify: `docs/DEVIN_CLAUDE_BRIDGE_PROGRESS.md`
-
- [ ] **Run the complete deterministic bridge proof before any paid request**
-
-```bash
-./scripts/devin-bridge/test-unit
-./scripts/devin-bridge/test-contract
-./scripts/devin-bridge/test-e2e-mock
-./scripts/devin-bridge/verify-anthropic-isolation
-npm run typecheck:core
-npm run lint
-npm run build
-npm run check:docs-all
-```
-
-Expected: all bridge-specific checks, typecheck, lint, build, and documentation checks pass with
-isolated data paths. Any unrelated full-suite infrastructure hang is recorded separately and is
-not converted into a pass.
-
- [ ] **Run exactly the three authorized live scenarios and the no-fallback failure probe**
-
-```bash
-ENABLE_LIVE_DEVIN_TESTS=1 ./scripts/devin-bridge/test-live-devin
-```
-
-Expected: dynamic discovery selects a returned Devin catalog model; Claude Code reads without
-editing, then edits and runs the fixture test, then executes the fixture command. Evidence shows
-native tool use by Claude Code, only `devin-cli-agentic` routing, no allowed non-Devin egress,
-and an Anthropic-shaped error after the Devin backend is deliberately made unavailable.
-
- [ ] **Update verified documentation and commit the evidence-backed delivery state**
-
-```bash
-git add docker/devin-bridge/run-claude-live-e2e.sh docs/DEVIN_CLAUDE_BRIDGE.md \
-  docs/DEVIN_CLAUDE_BRIDGE_PROGRESS.md
-git commit -m "docs: record verified Devin bridge live delivery"
-```
--- a/Show More
+++ b/Show More
				`@@ -1 +0,0 @@`
				- fix(executors): Vertex AI now routes Claude models through the native Anthropic `rawPredict` endpoint instead of the generic OpenAI-compatible partner endpoint, and synthesizes a real streaming response so Claude-via-Vertex works with `stream: true` ([#8909](https://github.com/diegosouzapw/OmniRoute/pull/8909)) — thanks @wgordon17
				`@@ -1 +0,0 @@`
				`- fix(dashboard): model-level allowed/blocked param edits now persist when the compatibility popover is closed by clicking outside, and a failed save no longer clears the edit or reports success ([#9013](https://github.com/diegosouzapw/OmniRoute/pull/9013))`
				`@@ -0,0 +1 @@`
				`- fix(qoder): include actionable CLI_QODER_BIN hint in connection test when qodercli is not found (#9277)`