Prompts, skills, and workflows for building AI agent systems with OpenClaw.
These are the actual prompts and configurations I use in my day-to-day setup — not theoretical, not hypothetical. If you've seen it in a video, the real version lives here.
| Folder | What | Description |
|---|---|---|
defenseclaw/ |
🛡️ Security Governance | Install DefenseClaw — scan skills, MCP servers, tool calls, and LLM traffic for threats |
agent-setup/ |
🤖 Agent Identity | AGENTS.md, SOUL.md, and USER.md templates for giving your agent memory and personality |
skills/ |
🧰 OpenClaw Skills | Drop-in skill files for specific capabilities |
workflows/ |
🔄 Workflows | Multi-step agent workflows (daily notes, content pipelines, task management) |
Most of these are designed for OpenClaw, but the prompts and patterns work with any AI agent system. Copy what's useful, adapt what isn't.
For the full agent workflow setup (Obsidian + Linear + OpenClaw), see 👉 AI Agent Workflow
Each prompt links back to the video where it was introduced. If you want context before copy-pasting, start there.
Found a bug in a prompt? Have an improvement? Open an issue or PR. These are living documents — they get better with use.
MIT