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/poco-ai/poco-claw/agents-mdgit clone --depth 1 https://github.com/poco-ai/poco-clawWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/poco-ai/poco-claw/agents-md)<a href="https://agentmods.dev/instructions/poco-ai/poco-claw/agents-md"><img src="https://agentmods.dev/badge/instructions/poco-ai/poco-claw/agents-md.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.02866 | $0.02866 |
| Opus 5 | $0.01433 | $0.01433 |
| Sonnet 5 | $0.00573 | $0.00573 |
| Haiku 4.5 | $0.00287 | $0.00287 |
Grade A, and why
poco-claw 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 4d 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 — 315 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
This file provides guidance for AI coding agents working with code in this repository.
Project Overview
Poco is a multi-service AI agent execution platform that orchestrates Claude AI agents to perform coding tasks. The system consists of four main components:
- Frontend (Next.js 16) - Web UI for task management and monitoring
- Backend (FastAPI) - API server, database management, and session orchestration
- Executor (FastAPI + Claude Agent SDK) - Agent execution engine with hook-based extensibility
- Executor Manager (FastAPI + APScheduler) - Task scheduling and dispatch service
Architecture Flow
- User creates task via Frontend
- Executor Manager receives task, creates session via Backend
- Executor Manager schedules task with APScheduler
- Task Dispatcher sends task to Executor with callback URL
- Executor runs Claude Agent SDK with configured hooks
- Hooks send progress callbacks to Executor Manager during execution
- Executor Manager forwards callbacks to Backend for persistence
- Frontend polls Backend for session status updates
Development Commands
Frontend (Next.js)
cd frontend
pnpm install # Install dependencies
pnpm dev # Development server
pnpm build # Build for production
pnpm start # Start production server
pnpm lint # ESLint
pnpm format # Prettier
Python Services (Backend, Executor, Executor Manager)
Each Python service has its own directory with a pyproject.toml. Run commands from within the service directory:
cd <service> # backend/, executor/, or executor_manager/
uv sync # Install dependencies
uv run python -m app.main # Run development server
# Or directly with uvicorn:
uvicorn app.main:app --reload
Database Migrations (Backend)
cd backend
uv run -m alembic revision --autogenerate -m "description" # Autogenerate migration (review and adjust)
uv run -m alembic upgrade head # Apply migrations
uv run -m alembic downgrade -1 # Rollback one migration
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.
- 4d ago First seen · 315 lines · 2,866 tokens per session scan A 4a8b71f11f2c
poco-claw AGENTS.md is an instructions file published in the GitHub repository poco-ai/poco-claw (1,353 stars, last pushed 29d ago), licensed MIT. It adds 2,866 tokens to every session, about $0.0143 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
lobu AGENTS.md
Instructions for lobu-ai/lobu, covering repo map, unrecoverable — never do these, facts you cannot derive, ship a change and how to work.
lobu CLAUDE.md
Instructions for lobu-ai/lobu, a project described as: Open-source control plane and runtime for organisational agents: shared company context, isolated execution, approvals and MCP.
milady AGENTS.md
Instructions for milady-ai/milady, covering milady — project codex guidance, cloud frontend visual review (required for any ui change), how to run, hard rules and elizaos naming.
zhin zhin-plugin.instructions.md
Instructions for zhinjs/zhin, covering zhin plugin runtime authoring, package contract, convention directories, imports and native typescript and command routes.
skill-based-architecture AGENTS.md
Instructions for WoJiSama/skill-based-architecture, covering agents.md, quick routing (survives context truncation), auto-triggers and red flags — stop.
octo-agent CLAUDE.md
Claude Code instructions for open-octo/octo-agent, covering claude.md, project, commands, architecture and adding capability.