skills.rest
HACK-WU
HACK-WUCOMMUNITY
Explore all skills by this author →
GitHub

auto-review

Automatically reviews and fixes AI-generated Markdown and code files after each write operation.

2|Updated Jun 8, 2026
Search Tags:#automation#quality assurance#code review#markdown#auto-fix#ai workflow#yaml frontmatter
One-click install
npx skills add https://github.com/HACK-WU/skills --skill auto-review-hack-wu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: auto-review
Source: https://github.com/HACK-WU/skills/tree/main/skills/auto-review
Command: npx skills add https://github.com/HACK-WU/skills --skill auto-review-hack-wu

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI-generated files often ship with formatting errors, missing frontmatter fields, security issues, or logic flaws that require manual rework. This Skill enforces an automatic review-fix-re-review loop on every file the AI writes, so deliverables meet quality standards before reaching the user. ## Core Features & Use Cases - Automatic review pipeline: Triggers after every write_to_file or replace_in_file call on .md, code, config, and script files, with specialized checks for Skill files (YAML frontmatter, AI description layer) and Rules files (alwaysApply, enabled, updatedAt fields). - Closed-loop fixing with limits: Classifies issues as P0/P1/P2, auto-fixes them, and re-runs full review up to 3 rounds before escalating unresolved problems to the user. - Complex-scenario escalation: Detects complex changes (multi-file edits, core logic, bug fixes, key paths) and automatically invokes the challenger skill for second-pass questioning, then filters the challenge report by code vs. non-code layers. - Use Case: After the AI writes a new Python script with exception handling and data persistence, the Skill runs code-review, fixes missing docstrings, judges the change as complex, invokes challenger for deep questioning, and delivers a filtered report highlighting only the issues that truly need attention. ## Quick Start Ask the AI to write or modify any Markdown or code file and let the automatic review loop validate and fix it before delivery.

Frequently Asked Questions about auto-review

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

FAQPage Schema
How do I automatically review AI-generated code files?▼

This Skill triggers automatically after every write_to_file or replace_in_file call on supported file types. It runs a review, auto-fixes found issues, and re-reviews up to 3 rounds before delivering the file.

What file types does automatic code review support?▼

It covers Markdown documents, Skill and Rules files, Python, C/C++, Java, Go, Rust, Shell, PowerShell, TOML/YAML/JSON configs, and JS/TS/Vue/CSS frontend code. Pure data files like CSV, logs, and binaries are excluded.

How does the Skill validate Skill and Rules file frontmatter?▼

Skill files are checked for name and description fields, name format (max 64 chars, lowercase with hyphens), and an overview section. Rules files are checked for description, alwaysApply, enabled, and updatedAt fields, with a warning when alwaysApply is false.

When does auto-review invoke the challenger skill?▼

Challenger is invoked when changes are judged complex: 3+ files or 50+ lines changed, control flow or concurrency edits, bug fixes, new features, refactoring, key business paths, or cross-module impact. Simple typo or formatting fixes skip it.

Can I skip the automatic review for a file?▼

Yes. Review is skipped when the user explicitly says to skip review, uses a --no-review marker, or makes a single-character change. You can also skip only the challenger second-pass while keeping the base review.

What happens if automatic fixing fails after multiple rounds?▼

The loop stops after 3 review-fix rounds. If issues remain, the Skill stops auto-fixing and reports the unresolved problems to the user with suggestions for manual judgment.

Related Skills

Explore More Software Engineering Skills→
medusajsmedusajs
35.7k

API Reference Documentation Writer

Generate and update MDX API reference documentation for the Medusa platform.

Official
Intermediate
launchapp-devlaunchapp-dev
38

design-review-process

Codify a four-gate design review process with criteria, checklists, and approvals.

Official
Advanced
mcouthonmcouthon
77

critic

Identify weaknesses and edge cases in designs through adversarial questioning.

Community
Advanced
mengfei0053mengfei0053
1

openharmony-error-detection

Orchestrate linting, unit tests, and build checks to detect OpenHarmony errors.

Community
Advanced