pr-body
Skill2 starsChanged 5 months agoMITView on GitHub
Write or refresh a pull request title and body for this harness, including approved spec references, complete acceptance evidence, actual validation results, and remaining risks. Use before opening a PR or updating its scope or proof.
- Commits and pushes
3 other projects use a version of this.
Adopted by
Most recently taken up by
- orrrrli/ai-engineering-stacktheir own version, first seen Nov 1, 2026
- orrrrli/clawd-harnesstheir own version, first seen Nov 1, 2026
- andrew-blake/melcloudhometheir own version, first seen Oct 8, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- andrew-blake/melcloudhome★ 143same name
- orrrrli/ai-engineering-stack★ 2same name
- orrrrli/clawd-harness★ 2same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository60 other files it gives its agents.
- .claude/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- analyzeSkill
- animation-disciplineSkill
- api-versioningSkill
- autopilotSkill
- browser-e2eSkill
- characterizeSkill
- clarifySkill
- codemodSkill
- constitution-compactSkill
- constitutionSkill
- context-budgetSkill
- creative-webSkill
- db-migrationSkill
- debugSkill
- detect-stackSkill
- dispatch-criteriaSkill
- dreamSkill
- experimentSkill
- fastapiSkill
- flag-cleanupSkill
- flag-rolloutSkill
- flask-realtimeSkill
- framework-detectorSkill
- front-end-designSkill
- grill-meSkill
- handoffSkill
- image-generationSkill
- implementSkill
- initiativeSkill
- instinctSkill
- loopSkill
- nextjsSkill
- observabilitySkill
- okr-alignSkill
- parallel-swarmSkill
- pivotSkill
- planSkill
- prime-disciplineSkill
- prod-observabilitySkill
- react-nodeSkill
- react-viteSkill
- release-trainSkill
- researchSkill
- reviewSkill
- security-guardSkill
- self-healSkill
- shipSkill
- skill-creatorSkill
- specifySkill
- tailwind-disciplineSkill
- tasksSkill
- tdd-loopSkill
- token-budgetSkill
- typographySkill
- verify-loopSkill
- verifySkill
- wip-checkpointSkill
- <kebab-case-skill-name>Skill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 2 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 9
Read the final diff, approved specs and plans, and actual check output before writing.
- Very specific: line 14
Run bash .claude/scripts/collect-evidence.sh <spec-id> for each applicable spec.
- Very specific: line 15
Commit the lightweight proof files required by the shipping workflow.
- Distinctive: line 17
Use the existing no-AC policy only when the entire diff qualifies.
- Very specific: line 24
Read back gh pr view <number> --json title,body,headRefOid and confirm that the published body includes every applicable bundle and …
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
PR body
Read the final diff, approved specs and plans, and actual check output before writing. Use the repository's PR template as the outline. Lead with the concrete problem and resulting behavior; rewrite the title and summary if scope changed.
Evidence contract
- Identify every spec touched or implemented by the final diff, with its plan and task IDs. Do not choose only the newest evidence directory when multiple specs apply.
- Run
bash .claude/scripts/collect-evidence.sh <spec-id>for each applicable spec. Include each resultingpr-body.mdevidence block in full, alongside the summary. A failed collector means verification is incomplete; report it and keep the candidate a draft rather than inventing a PASS. - Commit the lightweight proof files required by the shipping workflow. Run
python3 .claude/scripts/check-local-evidence.py --changed-files <diff-file>against the final changed-file list before publishing. The list should come fromgit diff --name-only <base>...HEAD, using the PR's actual base. - State which commands actually ran, their results, relevant artifact paths or CI links, and what remains untested. Distinguish local implementation proof from authenticated independent verifier/reviewer receipts. Pending, skipped, and historical checks are not current-head successes.
- Use the existing no-AC policy only when the entire diff qualifies. Describe its scope and actual validation; never fabricate acceptance counts to fill the template.
- Describe material risks and a rollback approach. For merged factory candidates, rollback targets the recorded merge SHA through an approved rollback PR.
Publish and maintain
Save the complete body to a file, preserving actual newlines and literal command text. Publish using gh pr create --body-file <file> or gh pr edit <number> --body-file <file>. --fill alone and a one-line verification summary do not satisfy the evidence contract. This applies to swarm and scripted PR creation too; supply the same complete body to those paths.
Read back gh pr view <number> --json title,body,headRefOid and confirm that the published body includes every applicable bundle and describes the current candidate. Refresh claims after source or scope changes; link final-head CI once available. In Codex, attach any newly created PR with the app's attachment tool.
Writing a body does not authorize merging or deploying. Request factory evaluation through autonomous-ship.sh; actual merge status comes from GitHub and its receipt. The required evidence gate remains the enforcement layer if skill invocation is missed.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

