Does it support languages other than Python?

The current implementation is optimized for Python environments, utilizing tools like uv, pytest, and ruff for validation and testing.

How does this skill ensure fixes don't break existing functionality?

The skill follows a strict validation workflow: it reads the full context of the file, applies the fix, verifies the file parses correctly, runs linters, and executes relevant unit tests using pytest. If any step fails, the change can be rolled back.

How are the fixes documented?

The skill maintains a cumulative PR_DESCRIPTION.md file that tracks all changes across multiple sessions, including stats on fixed issues, testing results, and remaining work.

What happens if a fix requires significant changes?

If a single fix requires changing more than 100 lines of code, the skill will alert you with a detailed explanation and require explicit approval before proceeding.

Can I limit the number of fixes applied at once?

Yes. Before starting the fix process, the skill asks for your preferences regarding issue types, complexity levels, and a maximum number of issues to address in a single run.

Automated Code Quality Fixer

Name: Automated Code Quality Fixer
Author: UKGovernmentBEIS

byUKGovernmentBEIS

•

380

•

Seguridad y Pruebas

Systematically resolves code quality issues identified in reviews through automated validation, unit testing, and recursive re-evaluation.

The code-quality-fix-all skill is a comprehensive automation tool designed to bridge the gap between identifying code smells and resolving them. It processes results from the code-quality-review-all skill to categorize issues by complexity, apply fixes using predefined patterns, and verify changes through rigorous linting and unit testing. By managing branch creation, recursive re-review cycles, and cumulative PR documentation, it ensures that technical debt is addressed safely and systematically without manual intervention for every line of code.

Características Principales

01Automated issue categorization by complexity (Easy, Medium, Hard)

02Interactive filtering to target specific evaluations or issue types

03Recursive re-review to verify fixes and identify newly introduced issues

04380 GitHub stars

05Multi-phase validation using repository-specific linting and pytest

06Cumulative PR description generation for comprehensive change tracking

Casos de Uso

01Automating the cleanup of technical debt identified during automated security or quality audits

02Systematically refactoring deprecated API patterns across large-scale repositories

03Ensuring code consistency and quality standards in complex AI evaluation frameworks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ukgovernmentbeis/inspect_evals code-quality-fix-all

For use in Claude.ai and ChatGPT

Download Skill