graft-pr-review

Automate GitHub pull request review and classify AI findings.

6|Updated May 11, 2026
One-click install
npx skills add https://github.com/GeWuYou/Graft --skill graft-pr-review
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: graft-pr-review
Source: https://github.com/GeWuYou/Graft/tree/main/.agents/skills/graft-pr-review
Command: npx skills add https://github.com/GeWuYou/Graft --skill graft-pr-review

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires requests, github3.py, yaml, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the GitHub PR review workflow for the Graft repository, streamlining the process of inspecting PRs, extracting AI review findings, and verifying which issues should be addressed in the local codebase.

Core Features & Use Cases

  • GitHub PR Review: Inspects GitHub pull requests and extracts AI review findings from CodeRabbit, greptile-apps, gemini-code-assist, and github-advanced-security.
  • Review Inventory: Generates an exhaustive inventory of findings, including failed checks, MegaLinter warnings, and failed test signals.
  • Classification & Resolution: Classifies each finding as fixed, delegated, blocked, stale, or noise, and facilitates resolution within the local codebase.
  • Inventory Management: Maintains a PR issue comment ledger with append-only entries for tracking review outcomes over time.

Quick Start

Run the graft-pr-review skill to review the current branch PR's findings and resolve them accordingly.

Frequently Asked Questions about graft-pr-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate GitHub PR review for AI findings in my repository?

Automating GitHub PR review involves inspecting pull requests to extract AI review findings from tools like CodeRabbit and greptile-apps, then generating an inventory of issues to classify and resolve them within your local codebase.

What is the best way to track AI code review findings across multiple GitHub PR checks?

Tracking AI code review findings is done by maintaining an append-only PR issue comment ledger, which records classification outcomes like fixed, delegated, blocked, stale, or noise for review signals and failed checks over time.

How do I consolidate MegaLinter warnings and failed test signals from a GitHub pull request?

To consolidate MegaLinter warnings and failed test signals, you can run an automated script that inspects the PR, extracts these specific findings, and generates an exhaustive inventory for local codebase resolution.

Do I need GitHub API access to extract CodeRabbit and gemini-code-assist review comments?

Yes, you need GitHub API access to programmatically extract review comments from CodeRabbit, gemini-code-assist, and github-advanced-security, utilizing Python libraries like github3.py and requests for the automation scripts.

Can I classify AI code review findings as noise or stale directly in my local codebase?

Yes, you can classify findings as noise, stale, fixed, delegated, or blocked directly in your local codebase, as the review workflow provides classification options for each extracted AI finding to facilitate resolution.

Why are my automated PR review scripts failing to classify github-advanced-security findings?

Automated PR review scripts may fail to classify github-advanced-security findings if the GitHub API access credentials are improperly configured or if the yaml dependency for scripting and automation is missing from the environment.