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.
npx agentmods add instructions/getmaat/maat/agents-mdgit clone --depth 1 https://github.com/getmaat/maatWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01462 | $0.01462 |
| Opus 5 | $0.00731 | $0.00731 |
| Sonnet 5 | $0.00292 | $0.00292 |
| Haiku 4.5 | $0.00146 | $0.00146 |
Grade C, and why
maat AGENTS.md scanned grade C with 1 finding 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.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
<!-- This is the single source of truth for AI coding agents working in this repository. It follows the AGENTS.md open standard (https://agents.md) and is read natively by Codex, Cursor, Gemini CLI, Jules, Factory, Aider How it starts
The opening of the file, as written. The whole thing — 152 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Ma'at
Project overview
Ma'at is a documentation-as-code framework that keeps three things in
lockstep: a repository's docs/ tree, its cross-agent instruction files, and
its source code. It is designed so that documentation can be maintained
interchangeably by AI coding agents, human developers, and CI — and so that
any agent harness (Claude Code, Copilot, Codex, Cursor, Windsurf, Hermes,
opencode, and others) discovers and updates the same docs.
Ma'at ships a small Go CLI (maat) with zero runtime dependencies
(see ADR 0011) —
distributed as a single static binary — with three verbs — init, sync,
check — and a documented convention for how the docs/ tree is structured
and kept current.
This repository dogfoods itself: the docs you are reading were scaffolded
by maat init and are validated by maat check in CI.
Where the documentation lives
All durable knowledge lives in docs/. It is documentation-as-code:
versioned with the source, reviewed in pull requests, and treated as a
first-class part of every change.
- Start here:
docs/llms.txt— a machine-readable index of every document. Read it first to locate what a task needs. - Human entry point:
docs/index.md. - Organized into:
docs/architecture/— how the system is built.docs/decisions/— Architecture Decision Records; the why, append-only.docs/guides/— how to develop, test, deploy, troubleshoot.docs/reference/— factual surface: CLI, config, front-matter schema.docs/meta/— conventions and glossary for the docs.
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.
- 2d ago First seen · 152 lines · 1,462 tokens per session scan C c29cd5e7c52c
maat AGENTS.md is an instructions file published in the GitHub repository getmaat/maat (4 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 1,462 tokens to every session, about $0.0073 per session on Opus 5. A static security scan graded it C with 1 finding (hidden instructions). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
dox AGENTS.md
AGENTS.md instructions for agent0ai/dox, covering dox framework, core contract, read before editing, update after editing and hierarchy.
talisman-skill AGENTS.md
Instructions for jeunecraack/talisman-skill, covering talisman — agent playbook (v1), when to run, workflow (8 phases), key features and repo structure.
dear-agent AGENTS.md
Instructions for bestcow/dear-agent, covering 작업 지침, 문서 규약 (항상 적용), 참고 자산 (선택), 서브에이전트·병렬 작업 (모든 ai 도구 공통) and 세션 절차 (모든 ai 도구 공통).
repo-audit AGENTS.md
Instructions for conorbronsdon/repo-audit, covering agents.md, what this repo is, rules, changing the skill and deliberately not here.
agent-rules-skill AGENTS.md
AGENTS.md instructions for netresearch/agent-rules-skill, covering agents.md, commands (unverified), workflow, file map and heuristics (quick decisions).
agent-rules-skill CLAUDE.md
Claude Code instructions for netresearch/agent-rules-skill, a project described as: Agent Skill for generating AGENTS.md files following the agents.md convention | Claude Code compatible.