review-pr-lite — Claude Code skill for PR review and setup
Review pull requests for code defects and security risks in a single skill, without delegating to subagents. It also checks GitHub Actions after the review.
npx skills add dyoshikawa/rulesync --skill review-pr-lite -a claude-codeWhat review-pr-lite does
review-pr-lite reads pull request metadata and changes through gh pr view and gh pr diff, then examines the patch itself. It looks for behavioral bugs, edge cases, inadequate tests, and maintainability problems. Its security checks cover injection, authorization mistakes, exposed secrets, and unsafe handling of files, network access, shell commands, or deserialization.
The report includes the PR number and numbered findings ranked by severity: low, mid, high, or critical. More severe issues appear first. Once the code review is complete, the skill checks GitHub Actions and reports the state of every workflow run, including runs still in progress. For failures, it investigates likely causes and suggests possible fixes.
The review runs without subagents and leaves the local branch untouched. When no PR is named, the skill targets the pull request associated with the current branch. The repository must be accessible through git and gh; the skill reads remote changes instead of checking out the pull request locally.
What it helps with
- Catch bugs and regressions in pull request changes
- Inspect patches for injection risks and exposed secrets
- Spot missing tests and overlooked edge cases
- Investigate failed GitHub Actions runs
- Rank review findings by severity in one report
Example prompts for the agent
- Review the pull request for the current branch and flag code defects or security risks.
- Review PR 123, rank the findings by severity, and check its GitHub Actions runs.
- Inspect PR 123 without changing my local branch and suggest fixes for failed workflow checks.
How to install review-pr-lite
1Skills CLI (Vercel)
npx skills add dyoshikawa/rulesync --skill review-pr-lite -a claude-code2Manually: put the files in place
~/.claude/skills/review-pr-lite/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/review-pr-lite/
1Skills CLI (Vercel)
npx skills add dyoshikawa/rulesync --skill review-pr-lite -a codex2Manually: put the files in place
.agents/skills/review-pr-lite/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add dyoshikawa/rulesync --skill review-pr-lite -a cursor2Manually: put the files in place
.cursor/skills/review-pr-lite/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install dyoshikawa/rulesync review-pr-lite2Manually: put the files in place
.github/skills/review-pr-lite/SKILL.md1Skills CLI (Vercel)
npx skills add dyoshikawa/rulesync --skill review-pr-lite -a gemini-cli2Platform CLI
gemini skills install https://github.com/dyoshikawa/rulesync --path .rulesync/skills/review-pr-liteManually: put the files in place
.devin/skills/review-pr-lite/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add dyoshikawa/rulesync --skill review-pr-lite -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about review-pr-lite
How do I install review-pr-lite in Claude Code?
Run npx skills add dyoshikawa/rulesync --skill review-pr-lite -a claude-code. The installed skill is located at ~/.claude/skills/review-pr-lite/SKILL.md.
What does the review-pr-lite skill do?
It reviews pull request changes for code quality and security issues, then checks GitHub Actions. The output ranks findings by severity and includes possible next steps for failed workflows.
Does review-pr-lite check out the pull request locally?
No. It reads PR metadata and diffs through gh, using read-only Git operations when needed. Your current local branch stays unchanged.
Source text
SKILL.md · MIT
--- name: review-pr-lite description: >- Review a pull request for code quality and security issues without using subagents. Use when the user wants a lighter-weight PR review in a single skill. targets: - "*" --- target_pr = the user's request If target_pr is not provided, use the PR of the current branch. ## Important: Do Not Switch the Local Branch - Do NOT check out, switch, or otherwise move the local branch (e.g., `git checkout`, `git switch`, `gh pr checkout`). - To inspect the PR's changes, use `git` and `gh` commands that read remote state without moving the working tree. For example: - `gh pr view $target_pr` to read PR metadata and description. - `gh pr diff $target_pr` to read the diff. - `gh pr view $target_pr --json files` or `gh api` to list changed files. - `git fetch origin pull/<PR_NUMBER>/head:refs/remotes/origin/pr-<PR_NUMBER>` and `git diff origin/main...origin/pr-<PR_NUMBER>` if a local read-only ref is needed. ## Step 1: Gather PR Context Run the following in parallel: - Get the PR description and metadata via `gh pr view $target_pr` (do not check out the PR locally). - Get the PR diff via `gh pr diff $target_pr`. - If needed, inspect changed files individually for deeper context using `gh` / `git` commands (e.g., `gh api`, `git show`) — without switching the local branch. ## Step 2: Review the Changes Review the PR directly in this skill without calling any subagents. Focus on both of the following: 1. **Code Review** - Bugs or behavioral regressions - Incorrect assumptions or edge cases - Missing or weak tests - Maintainability issues that could cause near-term problems 2. **Security Review** - Injection risks - Auth/authz mistakes - Secrets exposure - Unsafe file, network, shell, or deserialization behavior - Dependency or configuration changes that could weaken security ## Step 3: Report Findings Integrate all findings into one review result. Please output the PR number in the result so that the user can easily find the PR. ## Reporting Rules - Assign a severity level to each finding: low, mid, high, or critical. - Assign a sequential number to each finding (e.g., #1, #2, #3, ...). - Present findings first, ordered by severity. - If no findings are discovered, explicitly state that no findings were found. - Keep the summary brief and focused on risk and testing gaps. ## Step 4: Check GitHub Actions Workflows After completing the content review, check the status of the GitHub Actions workflows for $target_pr (for example, using `gh pr checks` or `gh run list`). - Report the status of every workflow run to the user, including runs that are still in progress. - For each failing workflow, also report: - The likely cause of the failure (based on the relevant logs or job output). - Candidate solutions or next steps to resolve it.
© dyoshikawa · MIT · dyoshikawa/rulesync