OpenClaw Expert Guide

Complete OpenClaw reference: architecture, gateway, CLI (50+ commands), 25+ channels, 25+ providers, tools, plugins, automation, security, installation, plat...

This skill appears to be what it claims: a large offline OpenClaw documentation pack with no installers or credential requirements. Before installing or enabling it for autonomous use: 1) Verify the source/author (the registry lists an owner ID but no homepage); prefer skills from known publishers or a GitHub repo you can inspect. 2) Manually scan SKILL.md and the reference files for any lines that tell the agent to run shell commands, read files outside the workspace, fetch remote URLs, or print/transfer secrets — documentation often contains such examples and an agent might repeat them. 3) If you plan to allow autonomous agent invocation, test the skill in an isolated/dev profile (use a non-production gateway/profile) so you can observe what suggestions the agent produces before granting it broader privileges. 4) If you are uneasy about prompt-injection, restrict the skill to user-invoked only (don’t enable autonomous invocation for high-privilege agents) and avoid providing any host secrets to the environment while evaluating the skill.

Type: OpenClaw Skill Name: openclaw-expert-guide Version: 1.0.0 The 'openclaw-expert-guide' is a comprehensive documentation-only skill bundle designed to provide an AI agent with a deep knowledge base of the OpenClaw platform. It consists of 12 structured reference files (references/01-12.md) and a routing guide in SKILL.md. The content is purely descriptive, covering architecture, CLI commands, and configuration schemas. There is no evidence of malicious code, intentional prompt injection, or data exfiltration attempts. The risky capabilities mentioned (such as shell execution and browser control) are documented features of the OpenClaw platform itself, and the guide provides standard operational instructions for their use.