ponytail-audit

Audits an entire codebase for over-engineering and ranks deletable code, dependencies, and abstractions.

Updated May 24, 2026
One-click install
npx skills add https://github.com/MWest2020/skill-forge --skill ponytail-audit-mwest2020
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ponytail-audit
Source: https://github.com/MWest2020/skill-forge/tree/main/skills/ponytail-audit
Command: npx skills add https://github.com/MWest2020/skill-forge --skill ponytail-audit-mwest2020

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Codebases accumulate dead code, redundant dependencies, and speculative abstractions over time, and manually hunting for this bloat across an entire repository is slow and inconsistent. This Skill performs a whole-repo audit that surfaces exactly what to delete, simplify, or replace with standard library or native platform equivalents. ## Core Features & Use Cases - Whole-repo over-engineering scan: Unlike a diff-based review, it walks the entire codebase and ranks findings biggest cut first. - Tagged findings: Each finding is labeled as delete, stdlib, native, yagni, or shrink, with a concrete replacement named for each. - Ranked one-shot report: Outputs one line per finding with file paths, ending with a net estimate of removable lines and dependencies. - Use Case: Before a major refactor, run the audit on a legacy service to discover that a hand-rolled retry helper duplicates a stdlib function, three config flags are never set, and a factory class has only one product, then delete them all. ## Quick Start Ask the agent to audit this codebase for over-engineering and list what can be deleted or replaced with stdlib equivalents.

Frequently Asked Questions about ponytail-audit

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

FAQPage Schema
How do I audit a codebase for over-engineering?

Run a whole-repo audit that scans every file for dead code, single-implementation abstractions, and hand-rolled utilities the standard library already provides. Findings are ranked biggest cut first with a concrete replacement named for each.

How to find unused dependencies and dead code in a repository?

Scan the full project tree for dead flags, unset config, wrappers that only delegate, and dependencies duplicating platform features. Each finding is tagged by type and reported with its file path and a net line and dependency reduction estimate.

What is the difference between ponytail-audit and ponytail-review?

ponytail-audit scans the entire repository instead of a single diff, using the same tagging system as ponytail-review. It produces a one-shot ranked report of what to cut across the whole codebase.

Does the over-engineering audit apply fixes automatically?

No, the audit only lists findings and applies nothing. It is a one-shot report covering complexity only; correctness bugs, security holes, and performance issues belong to a separate normal review pass.

When should I not use a whole-repo over-engineering audit?

Avoid it when you need correctness, security, or performance analysis, since the audit covers complexity only. It is also unsuitable for reviewing a single pull request, where a diff-scoped review is the better fit.