The good stuff written elsewhere. We do not repeat what others have done well — we link, with a one-line note about why each source is worth your time. If you think we are missing something important, tell us.
Official
- OpenClaw documentation Source of truth for every concept, schema, and CLI command. Some pages dense; we plain-English the gnarly ones in §1.
- OpenClaw GitHub repo README is the canonical install + supported-channels list. Skim every release note.
- OpenClaw documentation index (llms.txt) Machine-readable doc index. Useful for navigating the docs programmatically.
- docs.openclaw.ai install/docker Canonical Docker recipe — verify against this before using §2.7.
Community
- OpenClaw Discord Where the maintainers and serious users hang out. Worth joining if you are running OpenClaw.
- GitHub issues Triage your bug-or-feature questions here. Good closed-issue archive.
- Awesome OpenClaw (community curated) Watch for community-curated lists of skills, recipes, deployments. Use the trust signals from §6.2 before installing anything.
MCP / agent-runtime context
- Model Context Protocol The protocol OpenClaw mcp tool speaks. Useful background for §3.3 Tools.
- Reference MCP servers 200+ MCP servers you can wire into OpenClaw. Filesystem, GitHub, Slack, Postgres, browser, etc.
- Agentic Planet (sister site) Cockpit-style field guide to the broader agentic AI space. MCP scorecards, vendor takes, recipes.
Reading on agent design
- Anthropic on building effective agents The reference essay on when agents earn their keep. Pairs with §6.4 What NOT to build.
- Simon Willison on LLM tooling Practical, opinionated, technically grounded.