Octo-o-o-oOcto-o-o-o/openclaw-multi-agent-kit/AGENTS.md

openclaw-multi-agent-kit

AGENTS.md0 starsChanged 4 months agoMITView on GitHub

audience: any AI agent (Claude Code / Codex / Cursor / ...) asked to read, review, or extend this kit. A governance layer for multi-agent systems built on OpenClaw: role templates, task contracts, an independent acceptance gate, tiered recovery, and deployment tooling. It is a reference implementation extracted from a production single-operator system. There is no application code here — the product is the orchestration mechanics.

Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,439 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository2 other files it gives its agents.
- decision-claritySkill
- gogSkill

The file

We split this AGENTS.md into the rules it gives an agent

2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Very specific:
Test fixtures under scripts/fixtures/ contain intentionally fake credentials for scanner tests.
line 23
Very specific:
When citing behavior, cite file:line.
line 24

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

AGENTS.md

AGENTS.md — orientation for AI agents working in this repository

audience: any AI agent (Claude Code / Codex / Cursor / ...) asked to read, review, or extend this kit.

What this repository is

A governance layer for multi-agent systems built on OpenClaw: role templates, task contracts, an independent acceptance gate, tiered recovery, and deployment tooling. It is a reference implementation extracted from a production single-operator system. There is no application code here — the product is the orchestration mechanics.

Read in this order

README.md — what exists and why; quickstart tiers.
protocol/PROTOCOL.md — the inter-agent message contract. Everything else implements this.
agent-templates/coordinator/AGENTS.md — how dispatch, verification, and recovery are orchestrated.
ARCHITECTURE-SUMMARY.md (English) or ARCHITECTURE-IMPL.zh.md (Chinese, full detail) — layer map.
deploy.sh header comments — deployment model, shadow mode, drift detection.

Operating rules for agents

Do not run deploy.sh without --shadow unless the human explicitly asks: live mode writes crontab entries and modifies ~/.openclaw.
Do not edit config/profile.json (user-local, gitignored). Template changes go to config/profile.example.json.
Shell must pass shellcheck; JSON/JSON5 must parse. CI enforces both, plus a sensitive-path audit (scripts/check-sensitive-paths.sh).
Placeholders like /Users/USERNAME/..., YOUR_TELEGRAM_CHAT_ID, your-org are intentional — do not "fix" them to concrete values in commits.
Test fixtures under scripts/fixtures/ contain intentionally fake credentials for scanner tests. Do not report them as leaks (see .gitleaks.toml), and do not make them real.
When citing behavior, cite file:line. This repo's own acceptance culture is evidence-first; match it.

Repo-specific vocabulary

Term	Meaning
Done Definition	Completion criteria written at task intake, before dispatch; the acceptance gate verifies against it
Gate / acceptance	Independent role that re-runs tests instead of trusting worker self-reports
Tiered recovery	resume → re-dispatch → human escalation, narrowest tier first
Shadow mode	full deployment into an isolated root, zero live side effects
Intake	quarantined entry pipeline for external material/tasks, with provenance

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free