review-pr

Orchestrates multi-agent PR reviews to identify duplicates, dependency changes, and architecture violations.

4|3|Updated Feb 7, 2026
One-click install
npx skills add https://github.com/steilerDev/cornerstone --skill review-pr-steilerdev
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-pr
Source: https://github.com/steilerDev/cornerstone/tree/main/.claude/skills/review-pr
Command: npx skills add https://github.com/steilerDev/cornerstone --skill review-pr-steilerdev

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill orchestrates a comprehensive PR review workflow using a distributed team of agents to identify duplicates, suspicious changes, dependency changes, architectural violations, and spec non-compliance, then consolidates findings into a final verdict.

Core Features & Use Cases

  • Multi-agent review orchestration: delegates checks across product-owner, security, architecture, and quality agents to ensure broad coverage.
  • Context-aware analysis: analyzes PR metadata and diffs to classify affected areas and guide agent launches.
  • Consolidated reporting: aggregates independent agent findings into a unified review decision and actionable remediation notes.

Quick Start

Provide a PR number to initiate the multi-agent PR review.

Frequently Asked Questions about review-pr

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate multi-agent PR reviews for architectural and security violations?

Multi-agent PR reviews are orchestrated by delegating checks across specialized product-owner, security, architecture, and quality agents, then consolidating their findings into a final verdict. This automation classifies affected areas from PR metadata and diffs.

What is the best way to review external and Dependabot GitHub PR diffs for suspicious changes?

The best way to review GitHub PR diffs is through a deterministic workflow that performs context gathering, redundancy checks, and parallel agent reviews. This process identifies suspicious changes, dependency updates, and spec non-compliance before aggregating a final review decision.

Can I use a multi-agent workflow to re-review GitHub PRs after contributor fixes?

Yes, this multi-agent workflow supports re-reviews after contributor fixes. It consumes updated PR metadata and diffs to launch specialized agents for duplicate checks, architecture conformance, and code quality, producing an independent behavior change summary.

How does a distributed team of agents consolidate findings into a unified PR review decision?

A distributed team of agents consolidates findings by running parallel reviews for security implications and architecture violations, then aggregating those independent results into a unified review decision with actionable remediation notes.

Do I need to provide a PR number to initiate automated architecture conformance checks?

Yes, you need to provide a PR number to initiate the multi-agent PR review. This input triggers the deterministic workflow, which gathers context and launches specialized agents to evaluate architecture conformance and code quality.