A curated pack of 5 markdown files to design, architect, and deploy autonomous agents.
What's inside
| # | File | Purpose |
|---|---|---|
| 1 | SOUL.md |
Define your agent's identity, values, decision loop, and anti-patterns |
| 2 | AUTONOMOUS.md |
System architecture: OODA loop, components, MCP/A2A protocols |
| 3 | PROMPTS.md |
5 production-ready system prompts you can paste into any LLM |
| 4 | CHECKLIST.md |
48-step "Zero to Autonomous" deployment checklist |
| 5 | README.md |
This file — how to use the pack |
How to use this pack
- Read
SOUL.mdfirst — it sets the philosophical foundation. Decide what your agent is for and what it must never do. - Open
AUTONOMOUS.md— map the architecture. Identify your sensors, brain, and actuators. - Copy prompts from
PROMPTS.md— paste them into Claude, GPT-4, Kimi, or any LLM with a system-prompt slot. - Follow
CHECKLIST.md— tick boxes until your agent is live.
Who this is for
- Solo developers building their first autonomous tool
- SaaS founders adding an AI layer to an existing product
- Engineers who want a coherent, LLM-readable spec before writing code
Format
Every file starts with YAML frontmatter so LLMs can parse metadata automatically.
All sections use Markdown headers (##, ###) for easy chunking by retrieval systems.
License
This pack is sold as a digital product. Do not redistribute.