Does Critic Loop automatically commit code changes?

No, it performs changes in your working tree and uses 'git add -N' for visibility, allowing you to review and manually commit the final result once the loop finishes.

What is the Critic Loop skill?

Critic Loop is a specialized capability for Claude Code that automates a cycle of implementing code, performing deep AI analysis, and fixing issues until a development goal is fully realized.

Can I customize the review focus?

Yes, you can provide custom instructions for each loop, such as 'focus on security' or 'ensure this follows our specific repository patterns' to guide the AI reviewer.

How does it handle large codebase context?

It uses 'File-Mediated Analysis,' which writes detailed reports and logs to a local directory rather than flooding the main chat context, keeping the session fast and efficient.

Critic Loop

Name: Critic Loop
Author: in-the-loop-labs

byin-the-loop-labs

•

Herramientas de Colaboración

Automates an iterative implement-review-fix development cycle to ensure high-quality code through multi-level AI analysis.

The Critic Loop skill transforms Claude into an autonomous development partner that doesn't just write code but critically evaluates it. It orchestrates a structured workflow where an initial implementation is passed through layered AI analysis—checking for bugs, architectural consistency, and security—before automatically applying fixes in successive iterations. By maintaining a detailed history of changes and analysis in a local directory, it prevents context bloat while ensuring every development objective is met with production-grade quality and clean implementation.

Características Principales

01Multi-level AI analysis covering syntax, logic, and codebase-wide context

02File-mediated architecture to maintain high performance and low context usage

03State persistence and context recovery across session restarts or completions

04Automated iterative fix cycles that address both review findings and incomplete work

05Configurable analysis tiers ranging from fast (Haiku) to thorough (Opus)

0616 GitHub stars

Casos de Uso

01Automated security and logic auditing of new code before manual review

02Developing complex features that require iterative refinement and quality checks

03Refactoring existing codebases while maintaining strict architectural patterns

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add in-the-loop-labs/pair-review loop

For use in Claude.ai and ChatGPT

Download Skill