project-review

Automate architectural-conformance checks for code repositories with mechanical and LLM passes.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python, claude-code, and includes scripts (resource) components.

What problem does it solve?

This Skill automates architectural-conformance checks for code repositories, ensuring design invariants are maintained.

Core Features & Use Cases

  • Conformance Gate: Confirms code adheres to design invariants, preventing design drift.
  • Mechanical Checks: Executes deterministic checks on the entire repository.
  • LLM Judgment Pass: Utilizes LLM for judgment invariants and non-deterministic checks.
  • Use Case: Before a heavy-weight code review, this Skill can quickly identify gross design drift or violations.

Quick Start

Run the project-review skill to check the architectural conformance of the repository.

Frequently Asked Questions about project-review

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

FAQPage Schema
How do I automate architectural conformance checks for my code repository?

Automating architectural conformance checks involves running deterministic mechanical checks across the repository followed by an LLM judgment pass to identify design drift. This ensures code adheres to structural invariants before heavy-weight code reviews.

What is an LLM judgment pass in code review?

An LLM judgment pass in code review uses language models to evaluate non-deterministic judgment invariants. It complements mechanical repository audits by catching subtle architectural conformance issues and design violations that static rules cannot detect.

Do I need Python and Claude Code to run repository design verification?

Yes, running repository design verification requires both Python and Claude Code context. These dependencies are necessary to execute the scripts that perform mechanical checks and the LLM judgment pass for architectural conformance.

What is the best way to prevent design drift in a codebase?

The best way to prevent design drift is implementing a conformance gate that combines deterministic mechanical checks with LLM judgment. This verifies code against design invariants quickly, catching gross violations before manual code review.

Can I use this for repository audits on non-deterministic design rules?

Yes, repository audits can evaluate non-deterministic design rules using the LLM judgment pass. While mechanical checks handle deterministic repository constraints, the LLM judgment assesses complex architectural conformance and design invariants.