code-improving

Apply automated fixes for critical, high, and medium code-review issues.

2|Updated Jan 28, 2026
One-click install
npx skills add https://github.com/chriswch/praxis --skill code-improving
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-improving
Source: https://github.com/chriswch/praxis/tree/main/skills/code-improving
Command: npx skills add https://github.com/chriswch/praxis --skill code-improving

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Fix code quality issues identified by code-reviewing. Auto-fixes critical, high, and medium severity issues while preserving all existing tests. Leaves low-severity items for user decision. Cannot modify test files. Use after code-reviewing produces a review report.

Core Features & Use Cases

  • Automates applying fixes to code quality issues identified in reviews.
  • Ensures tests remain intact while fixing critical, high, and medium issues; leaves low-severity concerns for user review.
  • Use Case: after code-reviewing produces a review report, the tool automatically applies fixes and records results.

Quick Start

Run code-improving on the latest review to auto-fix qualifying issues and update the results.

Frequently Asked Questions about code-improving

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

FAQPage Schema
How do I auto-fix code quality issues reported in a code review?

To auto-fix code quality issues from a review, run this tool on the review report to apply automated, safe changes. It resolves critical, high, and medium issues while preserving all tests and leaving low-severity items for your decision.

Will automated refactoring modify or break my existing tests?

Automated refactoring will not modify or break existing tests because the tool explicitly preserves all test files. It applies fixes only to source code issues while keeping the test suite completely intact during the process.

What is the best way to handle low-severity linting issues after an automated code review?

The best way to handle low-severity linting issues after an automated code review is manual evaluation. The tool intentionally leaves these low-severity items unresolved for user decision, fixing only critical, high, and medium problems automatically.

Can I use this auto-fix workflow without a prior code review report?

You cannot use this auto-fix workflow without a prior code review report because it uses the review output as context. The tool requires an existing review to identify and apply fixes for the reported code-quality issues.

Where can I find the summary of automated code improvements after running the fix?

You can find the summary of automated code improvements at {artifact-dir}/improvement.md. The tool also records detailed automated fix results in a JSON artifact located at {artifact-dir}/results/code-improving.json.