Review workflow
Harness Kit is a reusable Codex workflow layer for repositories that want a clean operating system around implementation work.
npx -y skills add HyunKN/harness-kit --skill review-workflowAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Use when a change should receive a structured review before merge, handoff, or closure and the findings should be captured in `docs/reviews` or the issue doc. Trigger for non-trivial code changes, architecture changes, risky bug fixes, or external-tool evaluations. This skill defines review content and findings, not whether sub-agents should be used to perform the review.
SKILL.md
1.3 KB, 202 tokens by cl100k_base, as published. Nobody here has run it
Review Workflow
Run meaningful changes through a repeatable review instead of relying on scattered comments.
Default flow
- Confirm what changed and why it was done.
- Read validation evidence before commenting on style.
- Check:
- correctness
- readability
- boundary fit
- operational risk
- follow-up work
- Separate blockers from suggestions.
- Record the result where the next contributor will see it:
- issue doc for small reviews
docs/reviews/for broader analysis or reusable findings
Good fit
- multi-file implementation work
- high-risk fixes
- architecture or tooling changes
- external package or service evaluation
Avoid for
- obvious typo fixes
- tiny changes with no real behavioral risk
Resources
- See
references/review-lenses.mdfor a short review checklist. - Use
subagent-workflowonly if the review should be delegated or split across roles.
What ships with it: 2 files
395 B alongside SKILL.md
agents/
- openai.yaml136 B
references/
- review-lenses.md259 B