This repository was archived by the owner on Oct 9, 2026. It is now read-only.
Repository navigation
This repository was archived by the owner on Oct 9, 2026. It is now read-only.
Audit codex-skills inside Codex Lab dogfood #83
Description
Current Status
State: Complete. PR #496 merged into the upstream candidate as b599ca2c747754bed84d8986b9ff5fb352230b41 on July 29, 2026.
Final candidate evidence:
- Binding mandatory, delegated, and independent skill-routing guidance is model-independent and guarded under
AGENT-1. - The restored
gpt-5.6-solskill guidance, cache continuity, project-doc deduplication, and multi-turn scenarios pass in the exact 11-scenario harness. - The external-agent lifecycle scenario recovers both end markers from a large file-backed context handoff without inlining or truncation.
- The exact clean candidate and Every Code
0.6.116ran the samegpt-5.5/high task, repository snapshot, prompt, tests, and rubric. Both used all three implicit skills, changed no tests, and passed 2/2 tests. - Independent Opus review returned PASS/PASS with no blocking finding, dissent, or material quality difference.
- Strict convergence validation passes with 367 guarded paths, zero violations, zero stale waivers, and all four historical snapshots reproducible.
Machine-readable provenance and paired-result evidence is recorded in the July 29 evidence comment. No confirmed residual gap requires a new follow-up issue.
Next action: none for #83; #428 now owns the remaining candidate freeze and pending-restore decisions.
Blocked by: nothing.
Last verified: July 29, 2026.
Context
We want to audit whether codex-skills guidance works naturally inside Codex Lab once Codex Lab is usable as the active Codex-based coding CLI. The goal is not to copy Every Code behavior back into Codex Lab; it is to find where skills, Codex Lab product affordances, or both need to change so agents have low-friction, correct paths.
Audit Scope After Unblock
- GitHub helper-first behavior, comments-first issue orientation, PR checks, merge hygiene, and PR babysitting.
- Local LLM and local-agent flows, including exact prompt control where Codex Lab should avoid unintended harness/system-prompt pollution.
- Auto Review current-target handling, stale detached findings, bounded detail retrieval, and review-result surfacing.
- Browser/tool workflows and real-world task recovery.
- Skill routing, progressive disclosure, and whether skill instructions remain strong enough inside Codex Lab.
- Whether Codex Lab needs product-side affordances so skills feel like the shortest safe path, not a manual workaround layer.
Acceptance
- Codex Lab MVP dogfood plan #28 has reached a dogfoodable state.
- Run real Codex Lab tasks end to end using
codex-skills. - Capture concrete friction with reproduction notes and decide whether each fix belongs in Codex Lab,
codex-skills, or both. - Update or split follow-up issues for any confirmed gaps.
Relationships
- Native parent: Cut over to converged Codex Lab baseline and primary harness #343 primary coding-agent harness cutover.
- Blocked by: Sync TUI, Guardian, and CI substrate #308 final convergence.
- Blocks: Cut over to converged Codex Lab baseline and primary harness #343.
- Completed prerequisite: Make dogfood build provenance visible #311 immutable provenance and exact-candidate evidence.
- Historical foundation: closed Align Codex skill routing with Every Code guidance and harness tests #23 skill guidance/exec harness, Implement skills-cache-continuity exec-harness scenario #33 cache continuity, and Run Codex Lab dogfood readiness sweep #52 readiness sweep.
- Related: Add exact/minimal prompt profiles for non-code agents #93 prompt profiles and Prove current-binary third-party-agent dogfood #312 current-binary external-agent proof.
- related: Evaluate workflow-specific provider diversity after harness routing codex-skills#432 - Evaluate workflow-specific provider diversity after harness routing codex-skills#432
Acceptance Criteria
- The exact immutable Codex Lab candidate and Every Code baseline use the same model, reasoning level, repository snapshot, prompts, and acceptance rubrics.
- Mandatory, delegated, and independent multi-skill routing succeeds without explicit skill names where the contract requires implicit use.
- GitHub issue planning uses the correct durable issue, native relationship graph, multiline helpers, and status recovery conventions.
- Multi-turn resume preserves project/skill context and cache stability without duplicate instruction bloat.
- File-backed bounded context works for large external-agent handoffs without silently truncating required instructions.
- At least one representative code task produces a passing focused validation result and independently clean review in both harnesses.
- Codex Lab has no skipped mandatory instructions, lower task pass rate, unrelated edits, or blocking quality findings on the paired subset.
- Machine-readable artifacts identify the exact binaries, commits, task inputs, outputs, validation, and reviewer disposition.
Activity
[-]Blocked: audit codex-skills inside Codex Lab dogfood[/-][+]Audit codex-skills inside Codex Lab dogfood[/+]shiny-code-bot commentedon Jul 17, 2026
Planning note — July 17, 2026
The paired context/skills proof should include a generated Odoo workspace scenario now tracked by
cbusillo/codex-lab#360, with producer-side hardening incbusillo/odoo-devkit#82under parentcbusillo/odoo-devkit#81.The scenario should prove that the exact Codex Lab candidate starts from a generated non-Git root, loads the canonical workspace
AGENTS.md, preserves ownership guidance for child/delegated agents, surfaces stale generated context, and grants only typed editable source roots. This does not change #83's current #308 blocker or its exact-candidate sequencing; it records the Odoo workspace as one representative context-transport fixture once that gate opens.cbusillo commentedon Jul 24, 2026
Odoo generated-workspace consumer evidence landed in PR #467 (merge
fa23beb87fb0a4da197c74e42f6bc1a382df55aa). The checked-ingenerated-odoo-workspace-consumptionscenario now records exact candidate/source provenance, canonical guide loading, non-Git workspace state, exact writable roots, and denial for generated-root, managed, and outside paths; it passes on macOS and the Linux exec-harness runner. Native child-session root inheritance is covered separately.This supplies a representative scenario for the later paired quality/context proof, but it does not unblock or close this issue: #83 remains blocked by #308 and still needs the same-candidate Every Code comparison when that dependency clears.
shiny-code-bot commentedon Jul 29, 2026
Final #83 evidence for PR #496 (July 29, 2026).
{ "schemaVersion": 1, "issue": 83, "recordedAt": "2026-07-29", "candidateBranchHead": "67d008ca791949d1ea20511648f7012fe44aaeb7", "candidateRuntime": { "sourceCommit": "67d008ca791949d1ea20511648f7012fe44aaeb7", "binarySha256": "6593e1fd1f5144d12e6be1912bfae5eb5746ca0a738b7456f7bb5598e441d960", "buildProfile": "debug", "buildChannel": "dev", "dirtyState": "clean", "provenanceStatus": "current" }, "everyCodeBaseline": { "version": "0.6.116", "binarySha256": "a039bb8bc43df69818849c51388c4059e33a8c74896af3c3a86dfaf1100c9b56" }, "pairedTask": { "repositorySnapshot": "c70673cc96f2f2f5c38264fa2f6accc441742c01", "model": "gpt-5.5", "reasoningEffort": "high", "promptSha256": "d931416efb279e54144f7062734948656bb33974bf568d1a61d97f9991643645", "samePromptAndRubric": true, "candidate": { "returnCode": 0, "mandatoryMarker": "JS_DEFECT_SKILL_READ", "delegatedMarker": "TEST_VALIDATION_SKILL_READ", "independentMarker": "RELEASE_NOTE_SKILL_READ", "testsPassed": 2, "testsFailed": 0, "testsEdited": false, "blockingReviewFindings": 0 }, "everyCode": { "returnCode": 0, "mandatoryMarker": "JS_DEFECT_SKILL_READ", "delegatedMarker": "TEST_VALIDATION_SKILL_READ", "independentMarker": "RELEASE_NOTE_SKILL_READ", "testsPassed": 2, "testsFailed": 0, "testsEdited": false, "blockingReviewFindings": 0 }, "independentReviewer": { "provider": "claude-opus-5", "candidateDisposition": "PASS", "baselineDisposition": "PASS", "materialQualityDifference": false, "dissent": false } }, "deterministicHarness": { "model": "gpt-5.6-sol", "scenarioCount": 11, "passed": 11, "failed": 0, "includes": [ "binding mandatory/delegated/independent guidance", "multi-turn project and skill cache continuity", "project-document and skill deduplication", "large file-backed external-agent handoff with first/last marker recovery" ] }, "convergence": { "reviewBase": "5ae93aee76aca2f498150f9a5dd3cc70f4633612", "guardedPaths": 367, "violations": 0, "staleWaivers": 0, "pendingRestoreWaivers": 24, "historicalSnapshotsReproduced": 4, "passed": true } }Additional focused gates passed: 122 core-skills unit tests, 6 environment-loader tests, 57 skills-extension tests, and the 2 executor + 1 host-refresh app-server skill tests. The earlier broad app-server run passed all 1,068 tests after building required external fixtures.
Gemini/Antigravity completed both requested review batches but returned no written report. Opus supplied the recorded independent PASS/PASS disposition with no blocking finding or material quality difference.
shiny-code-bot commentedon Jul 29, 2026
Completed by merged PR #496 (
b599ca2c747754bed84d8986b9ff5fb352230b41). All acceptance criteria are checked in the issue body. The exact candidate has clean provenance, 11/11 deterministic harness scenarios, strict convergence, file-backed external-agent context recovery, and paired PASS/PASS task quality against Every Code0.6.116with machine-readable evidence in the July 29 comment.