Skip to main content
The Code Review Plugin automates pull request review by launching multiple agents in parallel to independently audit changes from different perspectives. It uses confidence scoring to filter out false positives, ensuring only high-quality, actionable feedback is posted.

​
Overview

This plugin runs comprehensive, multi-agent PR reviews that focus on:
  • CLAUDE.md guideline compliance
  • Bug detection in changes
  • Historical context analysis
  • Confidence-based issue filtering (threshold: 80)
Key Feature: Only issues with 80+ confidence scores are reported, eliminating noise from false positives.

​
Command: /code-review

Performs automated code review on a pull request using multiple specialized agents.

​
What It Does

1

Pre-flight Checks

Skips review if PR is:
  • Closed or draft
  • Trivial or automated
  • Already reviewed
2

Context Gathering

  • Collects relevant CLAUDE.md guideline files
  • Summarizes pull request changes
3

Parallel Agent Review

Launches 4 independent agents:
  • Agents #1 & #2: Audit for CLAUDE.md compliance
  • Agent #3: Scan for obvious bugs in changes
  • Agent #4: Analyze git blame/history for context
4

Score and Filter

  • Each issue scored 0-100 for confidence
  • Issues below 80 confidence filtered out
5

Output Results

  • Terminal output by default
  • Or PR comment with --comment flag

​
Usage

Run locally (outputs to terminal): 
/code-review
Post review as PR comment: 
/code-review --comment

​
Example Workflow

# On a PR branch, run locally:
/code-review

# Review the findings in terminal
# Make any necessary fixes

# Post review as PR comment:
/code-review --comment

​
Features

​
Multi-Agent Architecture

  • Multiple Independent Agents: Each agent reviews from a different perspective
  • Redundancy: Two agents check CLAUDE.md compliance for thoroughness
  • Specialization: Dedicated agents for bugs, guidelines, and context
  • Parallel Execution: All agents run simultaneously for speed

​
Confidence Scoring

Each issue is independently scored:
ScoreMeaning
0Not confident, false positive
25Somewhat confident, might be real
50Moderately confident, real but minor
75Highly confident, real and important
100Absolutely certain, definitely real
Threshold: Only issues ≥ 80 are reported.

​
Filtered False Positives

  • Pre-existing issues not introduced in PR
  • Code that looks like a bug but isn’t
  • Pedantic nitpicks
  • Issues linters will catch
  • General quality issues (unless in CLAUDE.md)
  • Issues with lint ignore comments

​
Review Comment Format

## Code review

Found 3 issues:

1. Missing error handling for OAuth callback (CLAUDE.md says "Always handle OAuth errors")

https://github.com/owner/repo/blob/abc123.../src/auth.ts#L67-L72

2. Memory leak: OAuth state not cleaned up (bug due to missing cleanup in finally block)

https://github.com/owner/repo/blob/abc123.../src/auth.ts#L88-L95

3. Inconsistent naming pattern (src/conventions/CLAUDE.md says "Use camelCase for functions")

https://github.com/owner/repo/blob/abc123.../src/utils.ts#L23-L28

​
Best Practices

​
Using /code-review

Maintain CLAUDE.md

Clear CLAUDE.md files lead to better compliance checking

Trust the Threshold

The 80+ confidence threshold filters false positives effectively

Run on All PRs

Use for all non-trivial pull requests

Starting Point

Use agent findings as a starting point for human review

​
When to Use

  • All pull requests with meaningful changes
  • PRs touching critical code paths
  • PRs from multiple contributors
  • PRs where guideline compliance matters

​
When Not to Use

  • Closed or draft PRs (automatically skipped anyway)
  • Trivial automated PRs (automatically skipped)
  • Urgent hotfixes requiring immediate merge
  • PRs already reviewed (automatically skipped)

​
Workflow Integration

​
Standard PR Review

1

Create PR

Push your changes and create the pull request
2

Run Local Review

/code-review
Review feedback in terminal
3

Fix Issues

Address any high-confidence issues found
4

Post as Comment

/code-review --comment
Share findings with team

​
CI/CD Integration

# .github/workflows/pr-review.yml
name: Automated PR Review

on:
  pull_request:
    types: [opened, synchronize]

jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - name: Run Code Review
        run: |
          claude /code-review --comment

​
Requirements

  • Git repository with GitHub integration
  • GitHub CLI (gh) installed and authenticated
  • CLAUDE.md files (optional but recommended for guideline checking)

​
Configuration

​
Adjusting Confidence Threshold

The default threshold is 80. To adjust, modify commands/code-review.md: 
Filter out any issues with a score less than 80.
Change 80 to your preferred threshold (0-100).

​
Customizing Review Focus

Edit commands/code-review.md to add or modify agent tasks:
  • Add security-focused agents
  • Add performance analysis agents
  • Add accessibility checking agents
  • Add documentation quality checks

​
Troubleshooting

​
Review Takes Too Long

Issue: Agents are slow on large PRs Solution:
  • Normal for large changes - agents run in parallel
  • 4 independent agents ensure thoroughness
  • Consider splitting large PRs into smaller ones

​
Too Many False Positives

Issue: Review flags issues that aren’t real Solution:
  • Default threshold is 80 (already filters most false positives)
  • Make CLAUDE.md more specific about what matters
  • Consider if the flagged issue is actually valid

​
No Review Comment Posted

Issue: /code-review runs but no comment appears Check if:
  • PR is closed (reviews skipped)
  • PR is draft (reviews skipped)
  • PR is trivial/automated (reviews skipped)
  • PR already has review (reviews skipped)
  • No issues scored ≥ 80 (no comment needed)

​
GitHub CLI Not Working

Issue: gh commands fail Solution:
  • Install GitHub CLI: brew install gh (macOS) or GitHub CLI installation
  • Authenticate: gh auth login
  • Verify repository has GitHub remote

​
Tips

Write specific CLAUDE.md files: Clear guidelines = better reviews
Include context in PRs: Helps agents understand intent
Use confidence scores: Issues ≥ 80 are usually correct
Iterate on guidelines: Update CLAUDE.md based on review patterns

​
Plugin Information

Author: Boris Cherny ([email protected])
Version: 1.0.0

Build docs developers (and LLMs) love

Get started for freeTalk to us