[Plamyth]

Summary

• Normalize a requested native OpenAI reasoning effort against the selected model's published ladder before sending a Responses request. For gpt-6.1-sol, unsupported none and minimal become low.
• Apply the same final-route check to direct requests, compaction model switches, and combo failover. Keep routed provider behavior and the existing ultra wire conversion.
• Add a failing-before regression commit and outbound-body checks; update the configuration and transport contracts.

Verification

• Before the fix, bun test tests/codex-integration/codex-catalog-ladders.test.ts --test-name-pattern 'native model rejects none' failed: expected low, received null.
• After the fix, bun test tests/codex-integration/codex-catalog-ladders.test.ts tests/responses/responses-compaction-override.test.ts: 88 passed, 0 failed. The outbound-body checks cover direct, compaction-switch, and combo-failover sends.
• bunx tsc --noEmit --pretty false, bun scripts/privacy-scan.ts, and bun scripts/structure-ssot.ts passed. An independent read-only review found no concrete defect.
• bun scripts/test.ts --changed=dev selected 1,253 of 2,615 test files and reached its built-in 900-second limit (exit 124). The full selected suite is not claimed as passing; CI must complete the remaining coverage.
• A standalone candidate built from this head and reported opencodex 2.93.0. The shared proxy was not replaced while other requests were active.

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.

Review readiness checklist

This PR stays in draft until every box below is ticked. Tick all four boxes once the requirements are met:

• Required local validation passed; commands, results, and any full-suite exception are documented.

• I pushed my PR to a recent dev commit (at most 10 behind; a maintainer may still ask for the exact tip before merge).

• I resolved all correct Codex and CodeRabbit findings.

• My PR is ready for review.

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: lidge-jun/opencodex
• Branch: dev
• File: .coderabbit.yaml
• Commit: 8ebe46b
• Layers: Repository YAML
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: 68be5abb-a9ff-4f72-9e5d-0adcdbc2d67e

Commits

Reviewing files that changed from the base of the PR and between 8ebe46b and 283f2d0.

Files selected for processing: 6

• docs-site/src/content/docs/reference/configuration/agents.md
• src/codex/catalog/effort.ts
• structure/catalog.md
• structure/transports/responses-failover.md
• tests/codex-integration/codex-catalog-ladders.test.ts
• tests/responses/responses-compaction-override.test.ts

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 9 remain after this review.

Walkthrough

The native effort clamp now maps unsupported reasoning effort values to supported levels on the selected model’s ladder. Tests and documentation cover direct requests, compaction routing, and combo failover.

Changes: Native effort normalization

Layer / File(s)	Summary
Clamp unsupported effort values src/codex/catalog/effort.ts:56-69, 80-84, tests/codex-integration/codex-catalog-ladders.test.ts:218-221	The clamp recognizes none and minimal, converts ultra to max, and selects the highest supported effort at or below the request. If no supported effort is at or below the request, it selects the lowest supported effort. The test checks that none maps to low for gpt-6.1-sol.
Cover final-route normalization tests/responses/responses-compaction-override.test.ts:296-347, docs-site/src/content/docs/reference/configuration/agents.md:387-389, structure/catalog.md:484, structure/transports/responses-failover.md:287-290	Tests cover direct and compaction-routed requests, plus combo failover to gpt-6.1-sol. The documentation describes applying the selected native model’s effort ladder before sending the request.

Priority: Normal

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Bug fix

Possibly related issues

• Issue 7590 reports invalid native reasoning effort values for gpt-6.1-sol, including ultra. This PR changes the same native effort normalization path, while preserving the existing ultra conversion.

Possibly related PRs

• lidge-jun/opencodex#7600: Both changes modify native reasoning-effort normalization in src/codex/catalog/effort.ts. This PR extends the clamp behavior while preserving the ultra to max conversion.
• lidge-jun/opencodex#7605: Both describe the native-versus-routed wire contract. This PR adds effort clamping against the selected native model’s ladder.

Merge Risk: Minimal · up to 283f2

No actionable issue is established for the effort-normalization change. Complete the remaining CI checks before merging.

Pre-merge checks: 4 · 1

Failed checks: 1 warning

Check name	Status	Explanation	Resolution
Docstring Coverage		Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 2 functions across 3 files. (3 skipped: 3 …	Write docstrings for the functions missing them to satisfy the coverage threshold.

Passed checks: 4 passed

Check name	Status	Explanation
Description Check		Check skipped - CodeRabbit’s high-level summary is enabled.
Title check		The title clearly and concisely describes the primary change: clamping native reasoning efforts against the selected model's ladder.
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Docstring Coverage

Explanation

Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 2 functions across 3 files. (3 skipped: 3 unsupported.)

Resolution

Write docstrings for the functions missing them to satisfy the coverage threshold.

Finishing Touches

🧪 Generate unit tests (beta)

• 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.

[github-actions]

✅ READY

• all PR quality gates passed; the review readiness checklist is complete.

Review readiness checklist

• ✅ Required local validation passed; commands, results, and any full-suite exception are documented.
• ✅ I pushed my PR to a recent dev commit (at most 10 behind; a maintainer may still ask for the exact tip before merge).
• ✅ I resolved all correct Codex and CodeRabbit findings.
• ✅ My PR is ready for review.

✅ 4/4 boxes ticked.

This pull request has been marked Ready for Review.
The review-ready label marks this PR as ready; review automation runs independently.
Maintainers notified: @lidge-jun @Ingwannu

[lidge-jun]

Maintainer dev integration record (release 2.93 dev-merge loop), per the MAINTAINERS.md maintainer-integration exception for dev. This is maintainer integration, not an approving review.

• Exact head: 283f2d0; no change requests; 0 review threads.
• Required CI on the exact head (fork workflows approved by the maintainer after review): CI run 40232423717 (ci, gates, test 1/4–4/4, structure gate, docker smoke, docs site build: success), Catalog async contracts 40232423846: success, enforce-target 40233721475: success, hygiene: success. The changes job did not flag this PR as Windows-sensitive.
• Independent technical review (separate agent): 0 blocking code findings; direct, compaction-switch and combo-failover outbound bodies covered; no security-boundary change; nothing unsafe to run in CI.
• Maintainer recheck on the head merged onto origin/dev 88c17bc (after fix(responses): notice a Codex WebSocket that dies after the response started #7725): bun test --isolate codex-catalog-ladders, responses-compaction-override, responses-native-ultra-effort-wire, claude-model-info, file-size ratchet and both layout guards: 152 pass / 0 fail (pinned Bun 1.4.4); typecheck exit 0; structure:check pass.
• Sequencing: lands before fix(reasoning): separate support, thinking toggles and effort controls #7531/feat(codex): expose opt-in Flex tier for eligible routed models #7534, which should rebase onto it.