review-fix-loop

Automate iterative code reviews with fix-and-re-review loops until severity thresholds are met.

1|1|Updated May 9, 2026
One-click install
npx skills add https://github.com/butterflyskies/claude-marketplace --skill review-fix-loop-butterflyskies
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-fix-loop
Source: https://github.com/butterflyskies/claude-marketplace/tree/main/plugins/bsky/skills/review-fix-loop
Command: npx skills add https://github.com/butterflyskies/claude-marketplace --skill review-fix-loop-butterflyskies

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires bsky:elbow-grease, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates iterative code reviews to converge on zero findings at or above a specified severity threshold, improving code quality and efficiency.

Core Features & Use Cases

  • Automated Code Review: Executes code reviews and fixes automatically until a certain severity threshold is reached.
  • Customizable Thresholds: Set severity levels to trigger fixes, including P1, P2, and P3.
  • Incremental Reviews: Reviews changes incrementally to avoid re-reviewing unchanged code.
  • Fix Agents: Uses fix agents to apply minimal changes to resolve findings without altering design intent.
  • Comprehensive Reporting: Provides a detailed report of the review process, including the number of rounds, findings, and fixed issues.
  • Use Case: Use this Skill to automatically review and fix code during development, ensuring high-quality code is merged into the main branch.

Quick Start

Run the review-fix-loop skill with the specified arguments, e.g., review-fix-loop --pr <PR_NUMBER>.

Frequently Asked Questions about review-fix-loop

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

FAQPage Schema
How do I automate iterative code reviews to fix issues continuously?

Automated code review uses fix-and-re-review loops to apply corrections and re-evaluate code until a specified severity threshold is met. It supports incremental reviews to avoid re-checking unchanged code.

Can I run automated code reviews directly on GitHub pull requests?

Yes, automated code reviews support GitHub pull requests. You can execute the review process by specifying the pull request number, allowing the fix agents to resolve findings directly within that context.

How do I set severity thresholds to trigger fixes during automated code reviews?

You can set customizable severity thresholds, such as P1, P2, and P3, to trigger fixes. The review loop continues applying corrections until zero findings remain at or above your specified severity level.

Do I need any specific dependencies to run automated code review loops?

Yes, running automated code review loops requires the bsky:elbow-grease dependency and appropriate fix models. These components enable the fix agents to apply minimal changes that resolve findings without altering design intent.

What is the best way to ensure high-quality code is merged into the main branch?

Automated iterative code reviews ensure high-quality merges by using fix agents to apply minimal corrections until severity thresholds are satisfied. This provides comprehensive reporting on review rounds and fixed issues.

Why does an iterative code review loop avoid re-reviewing unchanged code?

Iterative code review loops use incremental reviews to avoid re-evaluating unchanged code. This focuses subsequent review rounds strictly on new changes and applied fixes, ensuring faster quality convergence.