Repository navigation
feat: expand v0.16 goldens and close remaining operator debt - #37
Conversation
Add surface golden fixtures (design expand/summarize/qa, script rewrite/compress/scene, image generate, video blocking/timeline), stage CI hardening under docs/ci pending workflow OAuth scope, verify OpenClaw parity, and refresh Hermes submission packaging docs/PR body. Packaging script gains a tar fallback when rsync is unavailable.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 6b7fb5c803
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".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "design", "expand", input_text=design_md, evidence=BRIEF, project_id="golden-authority" | ||
| ) | ||
| summarized = run_production("design", "summarize", input_text=design_md, project_id="golden-authority") | ||
| design_qa = run_production("design", "qa", input_text=design_md, project_id="golden-authority") |
There was a problem hiding this comment.
In the supported stdlib-only environment where PyYAML is absent, _extract_brief_fields falls back to parsing only scalar keys from this YAML brief, so the generated design lacks geometry and image/video rules; this QA call consequently returns NOT_COMPUTABLE, while the checked-in design-qa.json expects PROPOSED. Running python3 scripts/test_golden_v016.py in that environment consistently fails at the stable-view assertion for design-qa.json, despite the release documentation reporting this test as passing. Use an input format the fallback fully understands, extend the fallback, or make the golden explicitly account for the degraded path.
Useful? React with 👍 / 👎.
1 participant
Summary
evals/golden/v016/from 6 → 15 fixtures (design expand/summarize/qa, script rewrite/compress/scene, image generate, video blocking/timeline)docs/ci/(token still lacksworkflowscope)origin/openclaw@db695bc(smoke/doctor/phased/golden PASS)PR_BODY.mdfor v0.16; packaging script tar fallbackTest plan
python3 scripts/test_golden_v016.pypython3 scripts/test_v016_phased_acceptance.pypython3 scripts/audit_release_version.py --strictpython3 scripts/check_hub_sync.py