Use this checklist before publishing to GitHub or tagging a release.
- README explains what the project is and is not.
- Each skill has valid
SKILL.mdfrontmatter. - Detailed references are linked from each
SKILL.md. - Examples are synthetic, public, or explicitly approved.
- License is present.
- Contribution, conduct, security, and privacy notes are present.
- No personal names or local usernames.
- No local absolute paths.
- No private model filenames.
- No thesis drafts, unpublished reports, or manuscript text.
- No raw manufacturer maps or private calibration data.
- No binary research assets are accidentally included.
- The README says the skills are workflow aids, not certified engineering tools.
- Examples are labeled synthetic/reduced where appropriate.
- Validation limits are stated.
- No performance or efficiency result is presented as a real machine claim unless source-backed.
Run:
python scripts/validate_release.pyExpected result:
Repository release checks passed.
These checks cover file structure, privacy, and packaging only;
they do not validate engineering models.
- Run privacy review prompt.
- Run skill-usability review prompt.
- Run engineering-safety review prompt.
- Address or explicitly defer findings.