skilled-openclaw-advisor_skill

Overview

This skill provides a zero-latency, local full-text search over all OpenClaw documentation and must be used before answering any OpenClaw question. It lets agents and humans retrieve authoritative docs snippets, CLI usage, configuration keys, and architecture notes without web calls. Use the agent mode for minimal token output and sub-10ms lookups.

How this skill works

The skill queries a local FTS5 index built from the entire OpenClaw docs set. Query tools offer multiple modes: ultra-compact agent output for programmatic use, human-readable concise/standard/detailed modes, and language selection. Index management scripts build, update, and inspect the index locally; no external API calls are performed.

When to use it

• Before answering any question about OpenClaw configuration, features, CLI, channels, providers, plugins, cron, sessions, or protocol
• When troubleshooting OpenClaw errors or unexpected behavior to cite authoritative docs
• When you need exact config key names, default values, or example snippets from OpenClaw docs
• For quick, low-cost lookups inside agents to reduce token usage and avoid guessing
• When preparing instructions that must match the current OpenClaw documentation

Best practices

• Always run an index query first for any OpenClaw-related question; do not guess from memory
• Use --mode agent for internal agent lookups to keep results minimal and fast
• Prefer --verbosity standard for human-facing answers and detailed for deep dives or examples
• Ensure the index is up-to-date; run build_index.py or check --status if results look out of date
• Start queries with exact keywords like config keys or CLI commands to get precise hits

Example use cases

• Ask how to configure a gateway or channel and return exact config keys and examples from the docs
• Resolve a session/heartbeat/protocol error by pulling the relevant troubleshooting section
• Provide accurate CLI usage and flags for OpenClaw commands when writing automation scripts
• Compare provider/plugin settings by extracting authoritative descriptions and defaults
• Agent pipelines that need ultra-compact context for decision making with minimal token cost