Does this skill run my unit tests or CI pipeline?

No, it is designed for static analysis and logic review. It assumes that standard build steps, linting, and type-checking are handled separately by your CI/CD provider.

Will it catch bugs that require historical context?

Yes, one of the specialized agents is dedicated to reading git blame and the history of modified code to identify regressions or bugs in light of past changes.

How does this skill prevent low-quality review comments?

It uses a multi-stage validation process where a secondary agent scores every identified issue on a scale of 0-100. Only issues with a confidence score of 80 or higher are posted to the PR.

Does it support custom project rules?

Yes. It specifically looks for CLAUDE.md files in the repository root and modified directories to ensure all code changes adhere to your project-specific guidelines.

What permissions are required?

The skill requires access to the 'gh' (GitHub CLI) tool and appropriate permissions to view issues, pull requests, and post comments on your behalf.

GitHub PR Code Reviewer

Name: GitHub PR Code Reviewer
Author: Euricom-IO

byEuricom-IO

0•

协作工具

Conducts multi-agent, high-confidence code reviews for GitHub pull requests using historical context and project-specific guidelines.

This skill provides a rigorous, automated code review process for GitHub pull requests by orchestrating specialized AI agents. It begins by filtering out ineligible PRs (like drafts or automated updates) and gathering context from CLAUDE.md files and git history. Five parallel agents then audit the changes for bugs, stylistic adherence, and historical patterns. To ensure high-quality feedback, a secondary validation layer scores every finding, only reporting issues with a confidence score of 80% or higher. This eliminates pedantic noise and false positives, delivering concise, actionable feedback directly to GitHub via the CLI.

主要功能

01High-confidence filtering to remove pedantic nitpicks and false positives

02Seamless GitHub integration for automated commenting with cited code links

03Project-specific guidance through automated CLAUDE.md adherence checks

04Deep analysis of git blame and previous pull request discussion history

050 GitHub stars

06Multi-agent parallel auditing for bugs, style, and historical git context

使用场景

01Automating the initial code review pass to catch significant bugs before human intervention

02Enforcing consistent coding standards and CLAUDE.md instructions across a distributed team

03Reviewing complex changes by analyzing historical context and previous file modifications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add euricom-io/euri-ai-marketplace pr-code-review

For use in Claude.ai and ChatGPT

Download Skill