[lidge-jun]

Summary

• Translate the English Mistral retry/cache and OpenGateway Responses behavior into all seven provider guides (fr, ja, ko, ru, tr, zh-cn, zh-tw). Preserve the reviewed distinction between same-key retry limits and recovery using a healthy replacement key.
• Carry closed, unmerged docs(i18n): translate OpenGateway and Mistral provider-guide sections #7432 by MannXo and add its omitted configuration reference coverage: Mistral's default retry policy, shared recovery waits and opt-outs, and promptCacheKey forwarding. Include the current English OpenGateway custom-tool conversion paragraph in every locale.
• Closes [Docs]: translate the #7395 OpenGateway and Mistral provider-guide sections #7405.

Decisions: keep the change within the fourteen translated Markdown pages. The regression check lives in ignored scratch space to avoid touching shared test registrations. Korean OpenGateway already contained the earlier behavior, so only its missing custom-tool paragraph was added. #7432's latest Cross-platform CI run failed with no jobs and no retrievable log; it does not establish a documentation failure. Its actionable replacement-key review finding is preserved. The two commits currently exposed by #7432 use a personal address rather than a noreply address; attribution uses MannXo's verified GitHub user ID and canonical noreply address (17333793+MannXo@users.noreply.github.com).

Verification

• python3 .tmp/l7/provider_consistency.py: 28 passed, 0 failed; the same check against original tracked pages failed all 28 cases (exit 1) before editing. Checks cover both changed pages in each locale.
• Root and docs-site/: bun install --frozen-lockfile — exit 0; no manifest or lockfile changes.
• cd docs-site && bun run build — exit 0; 593 pages built, 84,861 internal links checked. Includes the built-in internal-link gate; the docs package has no separate check script.
• bun test tests/ci-workflows/docs-provider-discovery-limits.test.ts tests/ci-workflows/docs-provider-preset-counts.test.ts tests/ci-workflows/docs-provider-billing-claims.test.ts tests/ci-workflows/docs-429-failover-claims.test.ts tests/ci-workflows/docs-readme-translation-parity.test.ts tests/ci-workflows/docs-developer-role-policy.test.ts tests/ci-workflows/docs-link-targets.test.ts — exit 0; 212 passed, 0 failed across seven files.
• bun run test:changed — exit 1: 0 tests selected for 14 Markdown changes. The runner explicitly rejects an empty import-graph selection; the read-as-data documentation contracts above were run directly instead. This is not recorded as a passing check.
• bun run typecheck, bun run privacy:scan, bun run structure:check, git diff --check — each exit 0.
• git fetch origin dev and git merge --no-edit origin/dev — already up to date before publication.
• NOT RUN: full bun run test; disproportionately broad for a C0 prose-only update on a host shared with another delivery lane. Focused documentation/locale contracts and the docs build cover this change; general runtime suite coverage remains with CI. Windows dispatch is not applicable to this docs-only change.
• Independent review: fresh gpt-6.1-sol with high reasoning returned PASS at cd01e2de4ffda008fc788a2290cf707ebba4c3ab, with no actionable findings. Independently checked the fourteen-page scope, MannXo attribution, English semantic parity, 94 tables, 90 fenced blocks, and 14 added links with zero Markdown errors. Idiomatic polish has not received a native-speaker review; no material semantic uncertainty was identified. Main also checked that all fourteen added/updated reference rows render correctly in built HTML.
• Exact-head PR CI: Cross-platform CI run 40052009130, event pull_request, head cd01e2de4ffda008fc788a2290cf707ebba4c3ab, attempt 1; queued at last inspection, not passing evidence. Draft retained pending hosted checks; merge belongs to the manager.

Checklist

• Scope stays focused and avoids unrelated cleanup.
• Docs or release notes were updated when needed.
• Security-sensitive changes were reviewed for secrets, auth, and unsafe defaults. Documentation only; no runtime credential or auth behavior changed.

[coderabbitai]

Caution

Review failed

The pull request is closed.

Finishing Touches

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

✅ Deterministic PR hygiene checks passed.

[lidge-jun]

Local CI-equivalent (Linux) at 2220e1f

Overall: PASS under the temporary local-verification rule. This is local Linux evidence, not hosted CI.

• Host: lidge-AI-AI; Ubuntu 24.04.4 LTS; x86_64; kernel 7.0.0-30-generic.
• Bun test runner/gates: 1.4.3 (scratch-installed official binary, first on PATH); runtime pin: 1.4.4 (used for structure/docs roles).
• Exact detached head: 2220e1f4b1ea9e5a040d2cfddb6dd833ca812da0; expected brief: 2220e1f4b1ea9e5a040d2cfddb6dd833ca812da0; moved: False.
• Base comparison: 47316142fe06171a94de02d63f1727dbed4c0382; git merge-base origin/dev HEAD: 47316142fe06171a94de02d63f1727dbed4c0382.
• Workflow authority: origin/dev at 8066099651f8ed356bb73a67d9c99978ea65447c, .github/workflows/ci.yml and .github/actions/setup-project-bun/action.yml.
• Tests ran outside hooks, explicitly without GIT_DIR, GIT_INDEX_FILE, or GIT_WORK_TREE. CI=true; no shard/batch/parallel/timeout overrides. Original check failures were retained; no head-check retries.
• Scratch artifacts: /home/lidgeai/ocx-linux-ci.p2R3NT. Each command has its original recorded output and exit code.

Command (repository root unless noted)	Bun	Exit	Pass / fail / skip
bun install --frozen-lockfile	1.4.3	0	N/A
cd gui && bun install --frozen-lockfile	1.4.3	0	N/A
bun x tsc --noEmit	1.4.3	0	N/A
bun x tsc --noEmit -p tests/tsconfig.doctor-service-memory-contract.json	1.4.3	0	N/A
bun x tsc --ignoreConfig --noEmit --strict --target ESNext --module ESNext --moduleResolution bundler --types bun-types --skipLibCheck scripts/ci/docker-smoke.ts	1.4.3	0	N/A
bun run privacy:scan	1.4.3	0	N/A
bun run structure:check	1.4.4	0	N/A
cd docs-site && bun install --frozen-lockfile	1.4.4	0	N/A
cd docs-site && bun run build	1.4.4	0	N/A

Failure classification (failed commands remain failed):

• None.

Base comparisons (same Bun/environment; once per distinct failing file, reused where recorded):

• None needed.

NOT RUN: General test shards 1–4, GUI lint/tests/build, skill surface, release syntax, CLI help, storage-policy and api-usage: not selected by manager’s docs-only scope; Hosted CI and platforms/jobs outside the requested Linux scope

All original check statuses are retained. No head checks were retried. Diagnostic attribution counts (separate from primary evidence where applicable): {'pass': 0, 'fail': 0, 'skip': 0}.

Head re-fetched before posting: 2220e1f4b1ea9e5a040d2cfddb6dd833ca812da0. Source status: clean (git status --short empty); HEAD unchanged.

Recorded UTC: 2026-10-24T10:12:34.610452+00:00

[lidge-jun]

Maintainer integration under the owner-approved temporary merge rule (GitHub Actions stalled account-wide since ~02:20Z; MAINTAINERS.md maintainer-integration for dev).

Exact head 2220e1f4b1ea9e5a040d2cfddb6dd833ca812da0:

• Independent review: gpt-6.1-sol PASS, 0 actionable findings (L7 lane; same content, head only adds a dev merge).
• Local CI-equivalent, Linux (lidge-ts, Ubuntu 24.04, Bun 1.4.3 test runner / 1.4.4 runtime): root+gui frozen install, the three CI typecheck commands, privacy:scan, structure:check, docs-site frozen install + build — all exit 0 (docs-only PR: test suite not applicable; see the verifier comment above).
• Combined simulation on current dev 8066099 (lidge-ts): no file overlap with dev since 4731614; merge clean; typecheck / structure:check / privacy:scan exit 0; test:changed correctly selects 0 tests for 14 Markdown files, so it was replaced by every docs source-oracle test (17 files): 365 pass / 0 fail; docs-site build exit 0 (593 pages, 84861 internal links).
• Co-authored-by MannXo (carry of docs(i18n): translate OpenGateway and Mistral provider-guide sections #7432, decision D10).
  Hosted CI for this head is queued/cancelled by the stall; dev push CI will be checked once Actions recovers.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-24T10:48:28.331849Z	2220e1f	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[lidge-jun]

Local CI-equivalent (Linux) coverage addendum at 2220e1f

Final verdict: PASS. Docs-only scope selects no suite shards; this confirms the previously completed gate/docs results without rerunning them: #7646 (comment)

Host lidge-AI-AI, Ubuntu 24.04.4 LTS x86_64, kernel 7.0.0-30-generic. Bun runner/gates 1.4.3; runtime pin 1.4.4. Detached exact head 2220e1f4b1ea9e5a040d2cfddb6dd833ca812da0; brief head 2220e1f4b1ea9e5a040d2cfddb6dd833ca812da0; moved from brief False. Base 47316142fe06171a94de02d63f1727dbed4c0382; observed merge-base with origin/dev 47316142fe06171a94de02d63f1727dbed4c0382.

All new suite/base invocations use documented OPENCODEX_TEST_TRUSTED_FIXTURE_DIR=/home/lidgeai/ocx-linux-ci.p2R3NT/trusted-fixtures (new private directory, mode 0700 under trusted ownership). Original results remain unchanged; no completed head batch was retried. GIT_DIR, GIT_INDEX_FILE, and GIT_WORK_TREE are explicitly absent. CI=true; shard scope general, batch size at most 12, no parallel override, per-test timeout 60000 ms, batch process deadline 120 s, TERM/KILL grace 15 s.

Batch file lists are generated by the exact head’s original planner. Only its script-directory anchors point back to the original location; its test command/deadlines are unchanged. Every previously executed batch file list was checked against the reconstructed manifest. Remaining batches ran serially; any new failure stayed failed and did not suppress later missing batches.

Shard	Original batches executed	Remaining batches executed	Required total	Coverage
docs-only scope	N/A	N/A	N/A	All requested docs/gates completed; no suite batches selected

Remaining primary batches — each command is timeout --signal=TERM --kill-after=15s 120s bun test --isolate --timeout 60000 <exact manifest files>:

Shard / batch	Exit	Pass	Fail	Skip
None required	N/A	N/A	N/A	N/A

Remainder totals: 0 pass / 0 fail / 0 skip. Combined original + remainder command counts: 0 pass / 0 fail / 0 skip. Failed checks remain failed; final disposition uses only the explicit BASE-IDENTICAL exception.

Failure classification across original and remainder evidence:

• None.

Base comparisons (cached once per distinct file/exact failed batch):

• None needed.

Every original non-test gate result is retained in #7646 (comment). No source edits, pushes, approvals or merges were performed. Source status: clean (git status --short empty); HEAD unchanged. Final fetched head 2220e1f4b1ea9e5a040d2cfddb6dd833ca812da0.

NOT RUN: General test shards 1–4, GUI lint/tests/build, skill surface, release syntax, CLI help, storage-policy and api-usage: not selected by manager’s docs-only scope; Hosted CI and platforms/jobs outside the requested Linux scope

Exact manifests and raw logs (retained on the named Linux host):

• Per-batch argv/exits/counts: /home/lidgeai/ocx-linux-ci.p2R3NT/evidence-7646/remainder-results.json; all original commands: /home/lidgeai/ocx-linux-ci.p2R3NT/evidence-7646/results.json; final evidence: /home/lidgeai/ocx-linux-ci.p2R3NT/evidence-7646/coverage-analysis.json.

UTC: 2026-10-24T11:19:22.629638+00:00