Does it automatically post comments to my Pull Request?

No. By default, it generates a local markdown report in the .code-reviews/ directory to prevent noise. It will only post to the PR if you explicitly give it instructions to do so.

Can I use this for specific files or just whole PRs?

You can target specific files or directories by providing a path as an argument, or let it automatically detect uncommitted local changes.

How does it handle conflicting feedback from different agents?

The skill uses a consolidation process that deduplicates findings by file and line number, merges descriptions, and uses a majority vote to resolve severity conflicts.

What platforms are supported for Pull Request reviews?

This skill natively supports GitHub using the GitHub CLI (gh) and Azure DevOps via MCP tools to fetch metadata and diffs.

Advanced Parallel Code Review

Name: Advanced Parallel Code Review
Author: alphaleonis

byalphaleonis

0•

Collaboration Tools

Orchestrates parallel AI agents to conduct comprehensive, multi-perspective code reviews with consolidated findings.

The decaf-review skill enhances Claude's code review capabilities by running multiple specialized agents in parallel to analyze changes for logic, security, design patterns, and test quality. It supports varying intensity levels—from a 'quick' generalist sweep to a 'max' coverage analysis—and can integrate directly with GitHub and Azure DevOps pull requests. Findings are automatically deduplicated and consolidated into a unified, timestamped report that includes prioritized severities, agent reasoning, and a spec-compliance audit if a plan document is provided.

Key Features

01Support for reviewing local uncommitted changes or remote GitHub/Azure DevOps PRs

02Specification compliance checking against provided documentation or plans

03Parallel multi-agent execution for rapid, multi-perspective code analysis

04Smart triage that automatically selects specialist agents for security, design, and tests

050 GitHub stars

06Consolidated reporting with deduplicated findings and cross-agent severity voting

Use Cases

01Conducting high-signal pre-commit reviews to catch logic errors before pushing code

02Auditing legacy codebases against new design specifications or security requirements

03Automating the first-pass review of complex Pull Requests on GitHub or Azure DevOps

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add alphaleonis/decaf-claude-config code-review

For use in Claude.ai and ChatGPT

Key Features

01Support for reviewing local uncommitted changes or remote GitHub/Azure DevOps PRs

02Specification compliance checking against provided documentation or plans

03Parallel multi-agent execution for rapid, multi-perspective code analysis

04Smart triage that automatically selects specialist agents for security, design, and tests

050 GitHub stars

06Consolidated reporting with deduplicated findings and cross-agent severity voting

Use Cases

01Conducting high-signal pre-commit reviews to catch logic errors before pushing code

02Auditing legacy codebases against new design specifications or security requirements

03Automating the first-pass review of complex Pull Requests on GitHub or Azure DevOps

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add alphaleonis/decaf-claude-config code-review

For use in Claude.ai and ChatGPT