from claude-skill-registry326
Automated, structured PR reviewer: gathers metadata, diffs, CI results, dependency changes and provides a concise verdict with testing and documentation recomme
The Review Pull Request skill automates a thorough code review workflow for GitHub pull requests. Given a PR identifier, it fetches metadata and diffs, creates an isolated worktree, checks CI and lint results, inspects dependency changes (pyproject.toml), analyzes related issues, evaluates test coverage, and produces a structured review verdict. The skill emphasizes actionable findings and reviewer-focused recommendations rather than verbatim code dumps.
Use this skill when you need a comprehensive, repeatable PR review — for maintainers handling incoming contributions, release engineering, or when triaging complex PRs that touch dependencies, tests, or public APIs. It is especially useful when you want parallel data collection (diffs, checks, issues) and an audit-style final report.
gh and local helper scripts (e.g., dependency checker, worktree creation). No bundled scripts are included in the SKILL.md file itself.Agents with access to a local git workspace, the GitHub CLI (gh), and the ability to run shell scripts (maintainer tooling agents, dev-focused assistants) are ideal. The skill assumes the ability to create and operate inside an isolated worktree and to run repository-local tooling.
PR review skill for GitHub repos. Well-structured SKILL.md with 9 detailed steps covering worktree setup, diff analysis, dependency checking, CI lint status, issue context, code analysis, test coverage review, docs coverage, and structured verdict. No bundled scripts — all work via gh CLI and shell commands. Specific to a particular repo's conventions (dlt-hub) with references to CLAUDE.md and tools/check_dependency_changes.py.
Thorough PR review workflow. No security concerns — uses standard gh CLI commands, no remote code execution, no credential handling. Minor deduction for shell injection risk in unquoted PR_ID interpolation in gh commands, but PR_ID comes from gh pr view --json parsing so practical risk is low. Architecture is solid with clear step-by-step structure, but it's somewhat monolithic as a single large SKILL.md without scripts/ or references/ separation. Usefulness is good for teams doing structured PR reviews but repo-specific conventions reduce portability.
