implement-audit

Audit repository implementation completeness and output structured reports.

Updated Feb 27, 2025
One-click install
npx skills add https://github.com/tuanemuy/dotfiles --skill implement-audit
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: implement-audit
Source: https://github.com/tuanemuy/dotfiles/tree/main/config/claude/skills/implement-audit
Command: npx skills add https://github.com/tuanemuy/dotfiles --skill implement-audit

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

実装の完全性を監査するための専門スキル。TODO・FIXME・仮実装・省略箇所・未完成コードを網羅的に検出し、spec/ との照合で実装漏れを特定して構造化レポートとして出力する。

Core Features & Use Cases

  • パターンマッチング+コード意味分析+仕様照合の三段構えで未完成箇所を検出
  • .audit/completeness/ に構造化レポートを出力
  • プロジェクト全体を対象にスキャン。特定ディレクトリのみを対象にも設定可能

Quick Start

プロジェクト全体を対象に実装完全性監査を実行し、レポートを .audit/completeness/report.md に出力してください。

Frequently Asked Questions about implement-audit

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

FAQPage Schema
How do I audit implementation completeness across my entire codebase?

You can audit implementation completeness by scanning source directories for TODOs, FIXMEs, and placeholder code, then outputting a machine-readable report to .audit/completeness/.

What is the best way to detect unimplemented functions and placeholder code in a repository?

Detecting unimplemented functions requires a three-phase approach: pattern matching for markers, semantic analysis for unimplemented signals, and spec verification for missing coverage.

How do I verify if my source code matches the specifications in spec directory?

Spec verification cross-references the spec/ directory against source code to identify missing coverage, then outputs gaps in a structured completeness report under .audit/completeness/.

Can I run a code audit on a specific target directory instead of the whole project?

Yes, you can target a specific directory for the code audit; the Skill scans all source directories by default but accepts a specific target, using CLAUDE.md to guide scope and rules.

What limitations should I expect when scanning for TODO and FIXME markers?

Scanning for TODO and FIXME markers is limited by pattern matching alone, which misses semantic gaps; this Skill overcomes the limitation by adding semantic analysis and spec verification.