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/anshulixyz/multi-agent-loop-kit/agents-mdgit clone --depth 1 https://github.com/anshulixyz/multi-agent-loop-kitWhat 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 | $0.01558 | $0.01558 |
| Opus 5 | $0.00779 | $0.00779 |
| Sonnet 5 | $0.00312 | $0.00312 |
| Haiku 4.5 | $0.00156 | $0.00156 |
Grade A, and why
multi-agent-loop-kit 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 — 244 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — Multi-Agent Ownership Registry
Setting this kit up in this repo? If you are an AI agent helping configure the kit, read
prompts/setup-interview.mdfirst — it tells you how to interview the user and fill this file in for their project. The contents below are a generic template until that interview is done.
This file is the source of truth for who can edit what.
Every coding agent must obey path ownership, journal updates, proposal rules, approval gates, and PR discipline.
Two meanings of “agent”
| Term | Meaning |
|---|---|
| Coding agent | A Cursor / Claude / Codex session with a codename, worktree, owned paths, branch prefix, and kickoff prompt |
| Runtime agent | An app-level process in your product, e.g. an agent that subscribes to events, calls APIs, or executes user workflows |
This file is about coding agents.
Core rules
- One coding agent per worktree.
- One worktree per Cursor / Claude / Codex window.
- Every path has exactly one owner.
- No coding agent commits outside owned paths.
- Cross-boundary changes require a proposal or approval request.
- Shared contracts are frozen and operator-owned.
- Agents coordinate through journals, task briefs, approvals, and proposals.
- Agents never coordinate through private chat if the information affects build state.
- All implementation work happens on
<codename>/<slug>branches. - All meaningful work ends with tests, journal update, and PR.
Project Lead / Operator
| Field | Value |
|---|---|
| Codename | operator |
| Human | Project Lead / repo owner |
| Owns | Product decisions, contracts, specs, runbook, rules, final approval |
| Branch | main unless intentionally editing |
| Commit prefix | operator: |
Project Lead / operator owns:
AGENTS.mdOPERATING_GUIDE.mdPROTOCOL.mdRUNBOOK.mdLOOP_RADAR.mdfinal objectiveLOOP_MEMORY.mdfinal decisions- shared contracts/types/schemas
.cursor/rules/**agents-status/proposals/**decisionsagents-status/approvals/**decisions
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 · 244 lines · 1,558 tokens per session scan A 6498fc2a580c
multi-agent-loop-kit AGENTS.md is an instructions file published in the GitHub repository anshulixyz/multi-agent-loop-kit (5 stars, last pushed 2mo ago), licensed MIT. It adds 1,558 tokens to every session, about $0.0078 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 2026-08-31.
Other instructions, from other repositories
nobrainer-tech-skills AGENTS.md
Instructions for nobrainer-tech/nobrainer-tech-skills, covering working in nobrainer-tech-skills, repository map, skill format, nobrainer delivery contract and minimum sufficient change.
nobrainer-tech-skills copilot-instructions.md
Instructions for nobrainer-tech/nobrainer-tech-skills: The fifteen canonical Agent Skills are under skills/; retired predecessors exist only in Git history. Read the relevant skills/ /SKILL.md before changing a skill.
nobrainer-tech-skills GEMINI.md
Instructions for nobrainer-tech/nobrainer-tech-skills, a project described as: Portable agentic workflows for guarded autonomy, visible sessions, SDD, decisions, RCA, evaluation, and durable knowledge.
nobrainer-tech-skills CLAUDE.md
Instructions for nobrainer-tech/nobrainer-tech-skills, covering working in nobrainer-tech-skills, repository map, skill format, nobrainer delivery contract and minimum sufficient change.
meta-llm-charter CLAUDE.md
Instructions for entropyvortex/meta-llm-charter, covering meta v3.1 core charter, bias, meta-0, r1 decompose and r2 ask gate.
open-tag CLAUDE.md
Instructions for fancyboi999/open-tag, a project described as: Open-source, self-hostable alternative to Claude Tag — a Slack-style workspace where your team and its AI agents (Claude Code, Codex, GitHub Copilot, and more) work as teammates in channels, threads, DMs, and shared tasks. Your data stays on your machines.