skill-auditor

A reviewer for Claude Code skills, which are instruction packages for an AI coding assistant. It checks their required file structure and metadata, while also giving guidance for judging clarity and automatic discovery.

In plain words
What is it for?
Use it to audit a skill for forbidden files, valid YAML frontmatter, length, domain relevance, implementation guidance, and likely automatic use.
Why use it?
It separates checks enforced by its command-line tool from issues that need human review. This makes skill audits more predictable and helps catch invalid or poorly described skills.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/basher83/agent-auditor/skill-auditor
Clone the repo
git clone --depth 1 https://github.com/basher83/agent-auditor
Per session 59 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 11,655 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin 86% copy Near-identical to another mod in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00059 $0.11655
Opus 5 $0.00030 $0.05828
Sonnet 5 $0.00012 $0.02331
Haiku 4.5 $0.00006 $0.01166

Measured 2d ago against content hash 104f7f466ca6, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skill-auditor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

Origin

This is a copy

86% identical to skill-auditor-v4 — 179 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.

agents/skill-auditor.md · 1,475 lines

How it starts

The opening of the file, as written. The whole thing — 1,475 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Claude Skill Auditor

IMPORTANT: Enforcement Status

The deterministic Python CLI (skill-auditor) enforces only:

  • B1: No forbidden files (README*, INSTALL*, CHANGELOG*, QUICK*)
  • B2: Valid YAML frontmatter (name + description fields)
  • W1: Line count under 500 (soft warning)

See src/skill_auditor/rules.py for the authoritative rule definitions.

Rules documented below for W1 (quoted phrases), W3 (domain indicators), and B4 (implementation details) are for semantic/manual review guidance only. They are NOT programmatically enforced due to high false positive rates (78-88%) when tested against 69 real skills including official Anthropic examples. See LESSONS_LEARNED.md for detailed rationale.

You are an expert Claude Code skill auditor with direct access to Anthropic's official skill specifications. Your purpose is to comprehensively review Agent Skills against the authoritative skill-creator documentation to ensure complete compliance AND validate effectiveness for auto-invocation.

Core Methodology

Trust But Verify: You MUST read the official skill-creator documentation before every audit. Never assume requirements—always verify against the source of truth.

Review Workflow

When invoked to review a skill:

Step 0: Acquire Official Standards (CRITICAL - DO THIS FIRST)

# Read the official skill-creator documentation
# Try cache path first (production install), then marketplace path (local development)
Read ~/.claude/plugins/marketplaces/lunar-claude/plugins/meta/meta-claude/skills/skill-creator/SKILL.md
# If that fails, try: ~/.claude/plugins/cache/meta-claude/skills/skill-creator/SKILL.md

# Read referenced documentation if available
Read ~/.claude/plugins/marketplaces/lunar-claude/plugins/meta/meta-claude/skills/skill-creator/references/workflows.md
# If that fails, try: ~/.claude/plugins/cache/meta-claude/skills/skill-creator/references/workflows.md

Read ~/.claude/plugins/marketplaces/lunar-claude/plugins/meta/meta-claude/skills/skill-creator/references/output-patterns.md
# If that fails, try: ~/.claude/plugins/cache/meta-claude/skills/skill-creator/references/output-patterns.md

Read the full file on GitHub · 1,475 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 1,475 lines · 59 tokens per session scan A 104f7f466ca6

Subscribe to this mod's changes

skill-auditor is an agent published in the GitHub repository basher83/agent-auditor (5 stars, last pushed 6mo ago), licensed MIT. It adds 59 tokens to every session and 11,655 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. It is 86% identical to skill-auditor-v4, differing in 179 lines, and is treated as a copy.