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

AGENTS.md agentsA

AGENTS.md is agent-read markdown (agents) from hermes-labs-ai/rule-audit.

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

# Owner: Hermes Labs - https://hermes-labs.ai

# AGENTS.md — rule-audit

Guidance for AI agents (Claude Code, Cursor, Aider, Continue, etc.) editing this repo.

Human reviewers read this too. Short, imperative, load-bearing.

## What this repo is

`rule-audit` — a **static analyzer for AI system prompts**. Input: raw prompt text. Output: structured report of contradictions, coverage gaps, priority ambiguities, meta-paradoxes, absoluteness issues, and concrete edge-case attack scenarios.

Pure Python. Zero LLM dependency. Zero network calls. Runs in milliseconds.

Think: `bandit` / `semgrep` — but for system prompts.

Part of the Hermes Labs reliability stack:
- `rule-audit` — static analysis of prompt logical content (this repo)
- `lintlang` — static linter for agent configs and tool descriptions (structure, not logical content) — complementary
- `hermes-jailbench` — dynamic jailbreak regression suite for live endpoints — complementary
- `colony-probe` — prompt-confidentiality audit (system-prompt reconstruction via multi-turn probing)

## Start here

1. Read `.claude/CLAUDE.md` — project conventions, adding detectors, adding clusters.
…

Read the whole file at its exact version.

How to install

Latest version
mdr add hermes-labs-ai/rule-audit/AGENTS.md@git:20260913.d6aea49
Exact content
mdr add hermes-labs-ai/rule-audit/AGENTS.md@sha256:bbf9e37054eec75b

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_wl5yeq42ivbzqjfc.svg)](https://markdownregistry.com/a/art_wl5yeq42ivbzqjfc)

1 badge views in 30 days

Versions

versioncommittedcommitsizeaudit
git:20260913.d6aea49 latest2026-09-13 d6aea49 6,329 BA view · diff
git:20260907.eebbd822026-09-07 eebbd82 6,264 BA view · diff
git:20260606.72770882026-06-06 7277088 6,166 BA view · diff
git:20260419.55a8d152026-04-19 55a8d15 5,932 BA view · diff
git:20260418.8df961f2026-04-18 8df961f 5,930 BA view

Audit of the latest version

A  14 of 14 checks passed. Deterministic, no model, same answer every run.
  • pass: Size between 200 bytes and 200 KB (6329 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_wl5yeq42ivbzqjfc
GET https://markdownregistry.com/api/v1/resolve?ref=hermes-labs-ai/rule-audit/AGENTS.md
GET https://markdownregistry.com/api/v1/blob/bbf9e37054eec75b47cdee5579e2ae5fb3e1fddfda973f97fdcea64d667bae80

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
llms.txt llmstxt
hermes-labs-ai/rule-audit · llms.txt
git:20260923.3cc5946 · audit A · 3 stars
rule-audit skill
hermes-labs-ai/rule-audit · skills/rule-audit/SKILL.md · Audit a named AI system-prompt file for contradictions, coverage gaps, priority ambiguities, meta-paradoxes and…
git:20260920.015a25f · audit A · 3 stars

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

Other files named AGENTS.md

AGENTS.md agents
obra/superpowers · AGENTS.md
git:20260919.5bf4e78 · audit A · 291,204 stars
AGENTS.md agents
affaan-m/ecc · AGENTS.md
git:20260919.91ba9b4 · audit A · 266,832 stars
AGENTS.md agents
firecrawl/firecrawl · AGENTS.md
git:20260613.d9a35c0 · audit A · 184,144 stars
AGENTS.md agents
dietrichgebert/ponytail · AGENTS.md
git:20260710.b6c0448 · audit A · 144,862 stars
AGENTS.md agents
garrytan/gstack · AGENTS.md
git:20260914.4a3c6a8 · audit A · 134,082 stars
AGENTS.md agents
graphify-labs/graphify · AGENTS.md
git:20260414.7ec92ec · audit A · 121,202 stars
AGENTS.md agents
juliusbrussee/caveman · AGENTS.md
git:20260902.d9faebc · audit A · 107,708 stars
AGENTS.md agents
addyosmani/agent-skills · AGENTS.md
git:20260709.b665c5c · audit A · 98,554 stars

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