explain-pr
Skill8 starsChanged 6 months agoMITView on GitHub
Explain a pull request exhaustively and post the explanation as a GitHub PR comment. Use when the user says "/explain-pr", "explain PR in the comments", asks for a file-by-file PR walkthrough, or wants every PR to document what changed, why, verification, and its contribution to a secure, scalable, runnable agent or product.
8 other projects use a version of this.
Adopted by
Most recently taken up by
- PeteChu/skillstheir own version, first seen Nov 8, 2026
- developzone5/sanadtheir own version, first seen Nov 8, 2026
- mryll/skillstheir own version, first seen Nov 1, 2026
- EternallLight/tgc-skillstheir own version, first seen Nov 1, 2026
- kameyama/claude_skillstheir own version, first seen Oct 25, 2026
- timothyjoh/skillstheir own version, first seen Oct 13, 2026
Other versions8 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 8
- luainkernel/lunatik★ 611same name
- EternallLight/tgc-skills★ 11same name
- mryll/skills★ 8same name
- developzone5/sanad★ 3same name
- kimgoetzke/coding-agent-configs★ 2same name
- PeteChu/skills★ 1same name
- timothyjoh/skills★ 1same name
- kameyama/claude_skills★ 0same 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 repository6 other files it gives its agents.
- babysit-prSkill
- drive-prSkill
- goated-reviewSkill
- junior-modeSkill
- understand-prSkill
- windows-docker-readySkill
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 7 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 8
Create an evidence-backed walkthrough of the PR and post it as one durable PR comment.
- Distinctive: line 13
Read the PR metadata, linked issue, commits, full diff, and repository guidance.
- Distinctive: line 15
Read enough surrounding code and relevant history to explain behavior and intent.
- Very specific: line 18
Use the connected GitHub tools when available.
- Distinctive: line 22
Use this structure, omitting only sections that truly do not apply:
- Distinctive: line 63
Check existing PR comments for an earlier walkthrough from this skill.
- Distinctive: line 64
Update the existing walkthrough when possible.
- Distinctive: line 68
Do not approve, merge, modify code, or resolve review threads as part of this skill.
“Common” is measured: kept by at least 6 in 10 of the 7 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
Explain PR
Create an evidence-backed walkthrough of the PR and post it as one durable PR comment. An explicit invocation authorizes posting the comment; otherwise prepare the comment and ask before posting.
Inspect the actual PR
- Resolve the repository, PR number, head branch, and base branch. For a stacked PR, compare against its declared base, not
main. - Read the PR metadata, linked issue, commits, full diff, and repository guidance.
- Enumerate changed files from Git. Account for every file, including generated files, lockfiles, docs, and tests.
- Read enough surrounding code and relevant history to explain behavior and intent. Do not infer motivation from filenames alone.
- Collect verification evidence from checks and commands that actually ran. Never claim an unobserved result.
Use the connected GitHub tools when available. Use gh for details the connector cannot supply. Keep the operation read-only until posting the final comment.
Write the walkthrough
Use this structure, omitting only sections that truly do not apply:
## PR walkthrough
### What this PR does
<Plain-English outcome and the user-visible or developer-visible behavior.>
### Why this change exists
<Problem, regression, goal, or operational need. Distinguish stated facts from inference.>
### How it works
<End-to-end flow through the affected components. Explain important decisions and boundaries.>
### Every changed file
| File | What changed | Why it changed |
|---|---|---|
| `path` | ... | ... |
### Value toward the product goal
- **Secure:** <validation, isolation, auth, secret handling, regression protection, or “no direct change”.>
- **Scalable:** <architecture, performance, maintainability, test scaling, or “no direct change”.>
- **Runnable:** <local setup, deterministic execution, CI, failure diagnostics, or “no direct change”.>
- **Agent quality:** <behavior, prompts, tools, scenarios, evaluation feedback, or “no direct change”.>
### Verification
| Check | Result |
|---|---|
| `<exact command or CI check>` | Pass/fail/not run, with concise evidence |
### Risks, limits, and follow-ups
<Behavior intentionally not covered, rollout concerns, known failures, and stacked-PR dependency. Say “None identified” only after checking.>
### Reviewer path
<The shortest sensible order for reviewing the files and validating the change.>Keep the opening accessible, then become technically precise. Explain every touched file individually; do not hide files behind “miscellaneous.” Treat lockfiles and generated artifacts briefly but explicitly. Connect value claims to concrete changes. If a dimension receives no value, say so instead of inventing a benefit.
Post and verify
- Check existing PR comments for an earlier walkthrough from this skill.
- Update the existing walkthrough when possible; otherwise post one new comment. Avoid duplicates.
- Re-read the posted comment and confirm formatting, file coverage, base branch, and links are correct.
- Return the PR URL and a concise note identifying any unverified claims or known failing checks.
Do not approve, merge, modify code, or resolve review threads as part of this skill.
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.

