Can it handle very large pull requests?

To maintain high accuracy and avoid context window limitations, the skill automatically flags and skips pull requests containing more than 1,000 lines of changes.

What are CLAUDE.md files in this context?

CLAUDE.md files are project-specific guides. This skill reads them to ensure that the code changes align with your team's specific implementation patterns and best practices.

Does this skill run my build or unit tests?

No, it focuses on logic, context, and project guidelines. It is designed to complement your CI/CD pipeline, assuming that linters and test runners will handle syntax and build-related checks.

Does it provide line-specific feedback?

Yes, the skill generates formatted comments that include specific file paths, line numbers, and direct GitHub links with git SHAs for precise context.

How does the skill ensure high-quality feedback?

The skill uses five independent agents to audit the code from different perspectives and employs a scoring system that filters out any issues with a confidence score below 80%.

Professional Code Reviewer

Name: Professional Code Reviewer
Author: TheBushidoCollective

byTheBushidoCollective

•

Collaboration Tools

Automates multi-agent, high-confidence pull request reviews by analyzing code changes against project guidelines, historical context, and logic patterns.

The Code Reviewer skill transforms Claude into a sophisticated pull request auditor using a multi-agent orchestration pattern. It coordinates five parallel Sonnet agents to evaluate PRs for logic bugs, adherence to local CLAUDE.md guidelines, historical git context, previous PR discussions, and inline comment compliance. By utilizing a rigorous deduplication and confidence-scoring system (0-100), the skill filters out noise and nitpicks, ensuring only verified, high-impact issues are reported back to GitHub. This is ideal for teams looking to maintain high code standards and reduce the manual burden on senior engineers during the review process.

Key Features

01Confidence-based scoring system to eliminate false positives and nitpicks

02Automated adherence checking for project-specific CLAUDE.md guidelines

03Proof-of-work verification to ensure thorough agent analysis

04Parallel multi-agent audits for bugs, style, and historical context

0572 GitHub stars

06Integration with GitHub CLI for automated, formatted PR commenting

Use Cases

01Enforcing consistent project standards across distributed development teams

02Auditing complex changes against historical git context and previous PR feedback

03Pre-screening pull requests to catch logic errors before human review

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add thebushidocollective/han code-review

For use in Claude.ai and ChatGPT

Download Skill