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/lichas/maxclaw/agents-mdgit clone --depth 1 https://github.com/Lichas/maxclawWhat 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 | $0.00663 | $0.00663 |
| Opus 5 | $0.00331 | $0.00331 |
| Sonnet 5 | $0.00133 | $0.00133 |
| Haiku 4.5 | $0.00066 | $0.00066 |
Grade A, and why
maxclaw AGENTS.md 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.
How it starts
The opening of the file, as written. The whole thing — 45 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
This file provides guidance to coding agents (Codex, Claude, and similar) when working in this repository.
Required Workflow
- Complete the user request end-to-end: implement, verify, then report.
- Run relevant validation (tests/build/smoke) yourself; do not ask the user to verify in your place.
- If a request changes repository files, run the relevant E2E/regression test script(s) before finishing. If any E2E test fails, investigate the failure and either fix it or clearly report the blocker; do not ignore the failure.
- After successfully completing a request that changes repository files, run
make buildand then create agit commitfor that request. - If a user request results in a repository change (feature, bug fix, behavior change, config change, or docs change), you must append an entry to
CHANGELOG.mdunder## [Unreleased]. - Changelog entries should be concise and include:
- what changed
- where (key files)
- how it was verified (test/build command)
- If no repository files were changed, explicitly state that no changelog update is needed.
- When changing stateful UI or session-management logic, think through the full interaction loop before editing:
- identify every effect, auto-redirect, default selection, and persistence path that can overwrite the new state
- verify the intended empty/initial state, transition state, and restored state separately
- avoid "local fix, global regression" changes where one handler is updated but linked sync logic still forces old behavior back
- Core flows must gain regression coverage when they change:
- for session creation, chat routing, interrupt/resume, spawn/callback, and other user-visible control flow, add or extend an automated E2E/regression test in
e2e_test/ - do not treat manual clicking as sufficient verification for these paths
- if full UI automation is not available yet, still add the closest end-to-end regression that exercises the real persistence/API boundary and document the remaining gap
- for session creation, chat routing, interrupt/resume, spawn/callback, and other user-visible control flow, add or extend an automated E2E/regression test in
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 · 45 lines · 663 tokens per session scan A e3b175ba855a
maxclaw AGENTS.md is an instructions file published in the GitHub repository Lichas/maxclaw (231 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 663 tokens to every session, about $0.0033 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
hotplex AGENTS.md
Instructions for hrygo/hotplex, covering hotplex 项目知识库, 约定与规范, 必须遵守, 反模式（禁止） and 代码编辑规则.
openclaw-kapso-whatsapp CLAUDE.md
Instructions for Enriquefft/openclaw-kapso-whatsapp, covering claude.md, project, commands, structure and conventions.
hotplex CLAUDE.md
Instructions for hrygo/hotplex, a project described as: HotPlex — Unified access layer for AI Coding Agent.
azure-sdk-for-go go-code.instructions.md
Instructions for Azure/azure-sdk-for-go: All code should follow the guidelines from the Azure Go SDK Guidelines. This document is a summary of the most important guidelines to follow when contributing to the Azure Go SDK.
librefang AGENTS.md
Instructions for librefang/librefang, covering stack, layout, build, architecture and kernelhandle trait.
autonomous-os AGENTS.md
Instructions for autonomous-ai/autonomous-os, covering agents.md, multi-ide rules, working style, parallel work / subagents and device access rules.