Home / hermes-labs-ai / rule-audit · skills/rule-audit/SKILL.md · GitHub

rule-audit skillA

rule-audit is agent-read markdown (skill) from hermes-labs-ai/rule-audit: Audit a named AI system-prompt file for contradictions, coverage gaps, priority ambiguities, meta-paradoxes and absoluteness issues, using the local rule-audit static analyzer (offline, deterministic, no model call). Use when the user asks to audit, lint, review or check a system prompt, agent instruction file or persona file for internal conflicts, and names the file. Do not use for general code review, for prose documentation, or when no specific prompt file has been named..

Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.

What the file says

# Audit a system prompt with rule-audit

`rule-audit` is a static analyzer for AI system prompts — sentence splitting,
modal-verb regexes and hand-curated keyword clusters. It runs locally, makes no
network calls, calls no model, and is deterministic. Think `bandit` or `semgrep`,
but for prompts (https://github.com/hermes-labs-ai/rule-audit).

## What to do

1. **Resolve one target file.**
   - Use the file the user named.
   - If the user named no file, ask which one. Do not guess, and do not audit
     every candidate you can find.
   - `AGENTS.md`, `CLAUDE.md` and similar repository guides are *not* system
     prompts, and rule-audit is not calibrated for them. If the user points at
     one, audit it as asked, then say plainly that the result is unreliable for
     that kind of document — see "Scope" below.

2. **Run the bundled adapter, once.** It lives next to this file, in
   `scripts/`. Take the directory this `SKILL.md` was loaded from — Codex shows
   it as the `<path>` given with this skill, Claude Code as the skill's base
   directory, Gemini CLI as the skill's location — and run:

   ```bash
   python3 "<that directory>/scripts/codex_audit.py" "<the file to audit>"
…

Read the whole file at its exact version.

How to install

Latest version
mdr add hermes-labs-ai/rule-audit/rule-audit@git:20260920.015a25f
Exact content
mdr add hermes-labs-ai/rule-audit/rule-audit@sha256:4fda05cbddf252ed

Pin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_7wxrcgqbublgmhvh.svg)](https://markdownregistry.com/a/art_7wxrcgqbublgmhvh)

1 badge views in 30 days

Versions

versioncommittedcommitsizeaudit
git:20260920.015a25f latest2026-09-20 015a25f 6,522 BA view · diff
git:20260913.d6aea492026-09-13 d6aea49 6,522 BA view

Audit of the latest version

A  17 of 17 checks passed. Deterministic, no model, same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (6522 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

hermes-labs-ai/rule-audit · 3 stars · license MIT · pushed 2026-09-23 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_7wxrcgqbublgmhvh
GET https://markdownregistry.com/api/v1/resolve?ref=hermes-labs-ai/rule-audit/rule-audit
GET https://markdownregistry.com/api/v1/blob/4fda05cbddf252ed0a6e270a9b41c5817b29b46d79255e5ef6290829d512a68f

Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

More from hermes-labs-ai/rule-audit

CLAUDE.md@.claude claude
hermes-labs-ai/rule-audit · .claude/CLAUDE.md
git:20260913.d6aea49 · audit A · 3 stars
AGENTS.md agents
hermes-labs-ai/rule-audit · AGENTS.md
git:20260913.d6aea49 · audit A · 3 stars
llms.txt llmstxt
hermes-labs-ai/rule-audit · llms.txt
git:20260923.3cc5946 · audit A · 3 stars

Every file in hermes-labs-ai/rule-audit

Browse by kind, by grade A, or by owner.