audit

Audit code into four issue layers and generate a pass/fail report.

Updated Feb 25, 2026

Search Tags:#devops #quality assurance #code review #security #bug fixing #code audit

Ready to run

npx skills add https://github.com/perfectuser21/cecelia --skill audit-perfectuser21

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: audit
Source: https://github.com/perfectuser21/cecelia/tree/main/packages/quality/skills/audit
Command: npx skills add https://github.com/perfectuser21/cecelia --skill audit-perfectuser21

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill provides a structured and bounded approach to code auditing and fixing, preventing endless deep dives and ensuring critical issues are addressed efficiently.

Core Features & Use Cases

Layered Auditing: Classifies issues into four layers: L1 (blocking), L2 (functional), L3 (best practices), and L4 (over-optimization).
Bounded Scope: Defaults to L2, aiming to clear L1 and L2 issues to declare completion, avoiding scope creep.
Use Case: When a new feature is developed, use this Skill to audit the code for critical bugs (L1) and functional errors (L2) before merging, ensuring stability.

Quick Start

Initiate a code audit for the current branch, focusing on functional issues.

Frequently Asked Questions about audit

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

FAQPage Schema

How do I perform a scoped code audit to catch critical bugs before merging?▼

Perform a scoped code audit by classifying issues into four layers: blocking, functional, best practices, and over-optimization. This approach resolves critical bugs and functional errors to ensure stability before merging.

What is the best way to prevent scope creep during a security and quality assurance code review?▼

Prevent scope creep during code review by bounding your audit to resolve blocking and functional issues. Defaulting to this layer avoids infinite deep dives into over-optimization and declares completion once stability is ensured.

How does layer categorization work when fixing code quality issues?▼

Layer categorization works by classifying code issues into L1 blocking, L2 functional, L3 best practices, and L4 over-optimization. This structured audit generates a report with a pass or fail decision based on issue severity.

Can I use bounded code auditing for functional bug fixing on a new feature branch?▼

Yes, you can use bounded code auditing for new feature branches. It targets L1 critical bugs and L2 functional errors to verify stability and prevent endless deep dives during the bug fixing process.

What are the limitations of preventing infinite deep dives during a code audit?▼

The limitation of preventing infinite deep dives is that the audit defaults to clearing L1 and L2 issues. It may leave L3 best practices and L4 over-optimization issues unresolved to maintain a bounded scope and ensure project stability.