Multi-Agent PR Code Review for Openclaw

A sophisticated multi-agent review system that scans pull requests for bugs, history context, and compliance using confidence-based filtering.

emersonbraun
v1.0.0
Apr 7, 2026
0
915
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install eb-code-review

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install eb-code-review using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Multi-Agent PR Code Review?

This skill provides a high-fidelity, automated code review process designed to act like a senior engineer. By integrating with Openclaw Skills, the system utilizes multiple parallel agents to analyze pull requests from various perspectives, including bug detection, git history, and local documentation compliance. It moves beyond simple linting by evaluating the intent and historical context of changes.

The core strength of this skill lies in its noise-reduction strategy. It uses a secondary scoring phase where agents assign confidence levels to every identified issue. By filtering out any findings below a strict threshold, the skill ensures that developers only receive actionable, high-impact feedback, effectively eliminating the pedantic nitpicks common in automated tools.

Multi-Agent PR Code Review Use Cases

  • Automating high-quality code reviews for GitHub pull requests to reduce human reviewer fatigue.
  • Enforcing project-specific coding standards and architectural rules defined in CLAUDE.md files.
  • Analyzing historical context via git blame to identify potential regressions or recurring issues.
  • Integrating a multi-agent validation layer into existing dev-workflows for consistent code quality.

How Multi-Agent PR Code Review Works

  1. Performs an eligibility check to skip drafts, closed PRs, or already reviewed content.
  2. Scans the repository for relevant CLAUDE.md files to understand local project constraints.
  3. Generates a concise summary of the proposed changes using a fast inference agent.
  4. Launches five parallel specialized agents to audit for bugs, compliance, git history, and previous PR comments.
  5. Executes a confidence scoring phase where each identified issue is evaluated on a scale of 0-100.
  6. Filters the results to keep only high-confidence issues (score 80+) and posts a formatted summary to GitHub.

Multi-Agent PR Code Review Setup

This skill requires the GitHub CLI tool to be installed and authenticated on your system.

# Install the GitHub CLI
brew install gh

# Authenticate your account
gh auth login

# The skill will automatically use the authenticated gh session to interact with PRs

Multi-Agent PR Code Review Data Schema & Taxonomy

The skill organizes review data based on the following taxonomy:

Component Description Data Source
Change Context Summary of the PR and modified files gh pr view
Compliance Rules Project-specific guidelines CLAUDE.md
Audit Logs Findings from the five parallel review agents Multi-agent output
Confidence Map Numerical scores (0-100) assigned to each finding Scoring agents
Final Output Filtered, markdown-formatted comments gh pr comment

Multi-Agent PR Code Review Advanced Features

  • Multi-agent parallel execution using specialized focus areas for deeper analysis.
  • Automated confidence-based filtering to maintain a high signal-to-noise ratio.
  • Historical git context analysis to identify bugs rooted in legacy code patterns.
  • Deep integration with CLAUDE.md for custom, project-aware linting and auditing.
  • Support for advanced review patterns including blast-radius analysis and test coverage gap detection.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*