dev-cross-review

Orchestrates Claude and two Codex agents to produce a unified GitHub PR review report.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/yosuke7040/dotfiles --skill dev-cross-review
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dev-cross-review
Source: https://github.com/yosuke7040/dotfiles/tree/main/claude/skills/dev-cross-review
Command: npx skills add https://github.com/yosuke7040/dotfiles --skill dev-cross-review

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Orchestrates three AI reviewers (Claude, Codex #1, Codex #2) to analyze a GitHub PR in parallel and produce a consolidated review report for faster, more reliable feedback.

Core Features & Use Cases

  • Parallel PR reviews: three AI agents run concurrently in a managed orchestration layer to inspect diffs and PR metadata.
  • Cross-referencing and consolidation: the orchestrator aggregates findings, resolves conflicts, and rates issues by severity.
  • Output generation: creates a Markdown report and a terminal summary suitable for quick triage and merge decisions.

Quick Start

Copy this skill into your /skills directory and run the orchestrator with a PR URL to trigger parallel AI reviews.

Frequently Asked Questions about dev-cross-review

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

FAQPage Schema
How do I automate parallel AI PR reviews on GitHub?

You can trigger parallel AI PR reviews by copying this skill into your /skills directory and running the orchestrator with a GitHub PR URL. It fetches PR data and diffs, runs Claude and two Codex agents concurrently, and produces a consolidated markdown review report.

What is multi-agent orchestration for code reviews?

Multi-agent orchestration for code reviews coordinates concurrent AI agents to inspect PR diffs and metadata. This skill runs Claude and two Codex agents in parallel, cross-references their findings, resolves conflicts, and rates issues by severity for reliable merge decisions.

Do I need a local repository to run automated PR reviews?

Yes, you need a local repository to run automated PR reviews with this skill. It requires a local repo cloned at ~/src/{repo}/, fetched PR data and diff files, and agent prompts in the agents/ directory to execute the workflow with worktree isolation.

How does cross-referencing work in multi-agent PR review workflows?

Cross-referencing in multi-agent PR review workflows aggregates findings from concurrent agents, resolves conflicts, and rates issues by severity. This orchestrator collects analyses from Claude and Codex, consolidating them into a structured terminal summary and markdown report for quick triage.

What is the best way to consolidate AI code review findings into a single report?

The best way to consolidate AI code review findings is using a managed orchestration layer that runs agents in parallel and aggregates results. This skill produces a unified markdown report and terminal summary by cross-referencing findings from Claude and two Codex agents.