This is a comprehensive maintenance reference for OpenClaw, a self-hosted AI agent gateway that routes conversations across WhatsApp, Telegram, Discord, Slack, iMessage, and a dozen other channels simultaneously. It covers the full operational surface: channel setup and troubleshooting, model provider configuration, multi-agent routing, security hardening, the exec tool approval system, browser automation, memory management, and the entire CLI toolkit. The documentation is dense and practical, walking through real failure modes, config paths, and recovery commands. If you're running OpenClaw locally and need to debug a channel connection, tune agent behavior, or harden your deployment, this gives you the full map without having to hunt through scattered docs.
npx -y skills add win4r/openclaw-skill --skill openclaw --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills