How does this skill reduce false positives?

It uses a two-stage process where initial issues found by agents are cross-validated by a separate set of Opus subagents to ensure only high-confidence, high-signal feedback is posted.

Can it provide actual code fixes?

Yes, for small, self-contained issues, the skill generates committable suggestion blocks that can be applied directly within the GitHub UI.

How does it handle project-specific architecture rules?

It searches for and reads CLAUDE.md files in the root and relevant subdirectories to ensure the code changes comply with your specific project guidelines.

Does it check for code style and formatting?

No, it intentionally ignores style nits, linting issues, and subjective improvements to focus on critical bugs and project-specific rules defined in CLAUDE.md.

What tools are required to run this skill?

You must have the GitHub CLI (gh) installed and authenticated, as the skill uses it to fetch PR data, view diffs, and post comments.

Automated Code Review

Name: Automated Code Review
Author: Gorillaa-ai-org

byGorillaa-ai-org

0•

コラボレーションツール

Performs multi-agent automated code reviews on GitHub pull requests to identify high-signal bugs and project-specific rule violations.

This skill automates the pull request review process by orchestrating a pipeline of specialized AI agents—Haiku, Sonnet, and Opus—to provide high-quality, actionable feedback. It specifically focuses on identifying critical logic errors, security vulnerabilities, and adherence to project-specific guidelines defined in CLAUDE.md files. By employing a multi-step validation process, the skill filters out pedantic nitpicks and false positives, ensuring that developers receive feedback only on significant issues that require attention before merging, all delivered via inline GitHub comments.

主な機能

010 GitHub stars

02High-signal filtering to eliminate noise, style nits, and issues catchable by standard linters.

03Deep logic and security vulnerability scanning using parallel Opus agents.

04Multi-agent orchestration using specialized models for PR summary, logic analysis, and validation.

05Automated adherence checking for project-specific CLAUDE.md guidelines across directories.

06Direct GitHub integration using the gh CLI for inline comments and committable suggestions.

ユースケース

01Ensuring large-scale pull requests meet internal documentation and coding standards automatically.

02Automating routine compliance checks for open-source contributions in high-velocity repositories.

03Scanning for critical logic errors and security flaws in complex diffs before human review.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add gorillaa-ai-org/pipecat code-review

For use in Claude.ai and ChatGPT

主な機能

010 GitHub stars

02High-signal filtering to eliminate noise, style nits, and issues catchable by standard linters.

03Deep logic and security vulnerability scanning using parallel Opus agents.

04Multi-agent orchestration using specialized models for PR summary, logic analysis, and validation.

05Automated adherence checking for project-specific CLAUDE.md guidelines across directories.

06Direct GitHub integration using the gh CLI for inline comments and committable suggestions.

ユースケース

01Ensuring large-scale pull requests meet internal documentation and coding standards automatically.

02Automating routine compliance checks for open-source contributions in high-velocity repositories.

03Scanning for critical logic errors and security flaws in complex diffs before human review.