Skip to content
Get packs

GitHub Deep Review

Investigates GitHub bugs and PRs using gh and local code to find root causes and best fixes

Source Download skill
Repository
steipete/agent-scripts
Stars
7,433
Licence
MIT
Files
2 files

What it does

When installed, the agent reviews GitHub issues and pull requests in depth rather than giving a generic summary. It uses the GitHub CLI (gh) to pull issue and PR details and diffs, reads the surrounding code and call paths in a local repository, checks git history for where a bug came from, and answers a fixed set of questions: root cause, whether the proposed fix is the best one, whether a larger refactor would help, and what proof exists. It is told to say 'not proven' when evidence is weak.

Great for

  • Maintainers or engineers triaging bug reports and reviewing pull requests on a codebase they have locally
  • Teams that want evidence-based review notes with file and symbol references and an explicit list of unverified risks
  • Developers who want a structured provenance check of when a regression was introduced

Skip it if

  • Non-developers or quick summaries; the skill is written around a specific maintainer's review style and a deep code-reading workflow that is slow and token-heavy
  • Repositories other than the author's ecosystem without adapting it: it references paths like ~/Projects/agent-scripts and a companion author-context skill that is not included in these files
  • Taking action on GitHub: the skill explicitly forbids approving, commenting, closing, merging or pushing unless the user asks, but a misread of its output could still lead to a wrong decision if not checked by a human

Try asking

  • “Review PR 456 in this repo: explain the bug, check the root cause, and say whether the fix is the best shape”
  • “Is issue 123 already fixed on main? Reproduce it if feasible and give me the canonical commit”
  • “Trace where this regression was introduced using git history, and tell me what is verified versus unknown”
  • “Give me a concise Ref/Bug/Cause/Fix/Proof/Risk review of this pull request”
Step-by-step guide: how to use GitHub Deep ReviewInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 2 original files, the MIT licence and an attribution note.

    github-deep-review.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip github-deep-review.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip github-deep-review.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Point the agent at a local clone of the repository first so it can read code and run git history commands, not just the web view
  • Ask for the output template format if you want a short, consistent decision summary
  • Review any commands the agent proposes to run (gh, git fetch, rg) and check the authentication scope of your gh login before use

What's inside

  • agents/openai.yaml
  • SKILL.md

Redistributed unmodified from steipete/agent-scripts under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.