ccreview

Coordinates two-round cross-model code reviews with GPT-5.4 and Claude via OpenREST API key.

1|Updated Mar 20, 2026
One-click install
npx skills add https://github.com/jamie950315/ccreview --skill ccreview
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ccreview
Source: https://github.com/jamie950315/ccreview/tree/main
Command: npx skills add https://github.com/jamie950315/ccreview --skill ccreview

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Cross-model code review automates the process of reviewing code changes by sending diffs to GPT-5.4 for feedback, triage, and verification, improving accuracy and consistency across reviews.

Core Features & Use Cases

  • Two-round cross-model review: initial review, triage, fixes, and a follow-up verification round.
  • Cross-model perspective: GPT-5.4 reviews while Claude triages and applies fixes.
  • Sensitive file protection and secret redaction: excludes sensitive files and redacts keys before sending data.
  • Rejection history: maintains per-project history to avoid repeating rejected suggestions.
  • Use Case: when reviewing a PR, this skill ensures suggestions are validated and only safe changes are merged.

Quick Start

Install the skill into your Claude Code skills directory and trigger it with /ccreview or say "code review" to start the two-round workflow.

Frequently Asked Questions about ccreview

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

FAQPage Schema
How does cross-model code review with Claude and GPT-5.4 work?

Cross-model code review uses GPT-5.4 to analyze code diffs and provide feedback, while Claude triages suggestions and applies automated fixes. This two-round workflow ensures validated, safe changes are merged by leveraging multiple model perspectives.

How do I automate code review for a PR using Claude Code?

Automate PR code review by installing the ccreview skill into your Claude Code directory. Trigger it with /ccreview or say "code review" to start the two-round cross-model workflow for automated triage and fixes.

Do I need an OpenRouter API key for automated diff review?

Yes, automated diff review requires an OpenRouter API key to send code changes to GPT-5.4. The cross-model review workflow also requires Python 3.8+ and Claude Code to coordinate triage and apply fixes.

How are sensitive files and API keys handled during automated code review?

Automated code review includes sensitive file protection and secret redaction, excluding sensitive files and redacting keys before sending diffs to GPT-5.4. This prevents secret exposure during the cross-model review process.

How can I prevent automated code review from repeating rejected suggestions?

Automated code review maintains a per-project local rejection history to avoid repeating rejected suggestions. This prevents GPT-5.4 and Claude from re-surfacing previously dismissed feedback during subsequent cross-model diff reviews.