clawd-docs-v2

What This Skill Does

The clawd-docs-v2 skill provides a highly optimized, multi-tiered documentation access layer for the ClawdBot ecosystem. It is designed to minimize token usage while ensuring that users receive the most current, relevant technical information. By implementing a hierarchical retrieval strategy—checking local golden snippets first, then a local search index, followed by a cached full-page database, and finally on-demand fetching—this skill ensures that the agent remains efficient and cost-effective. It bridges the gap between raw web documentation and immediate agent-readiness by providing structured, markdown-ready data through local paths and smart cache management.

Installation

To install this skill, run the following command in your terminal or via the command interface:

clawhub install openclaw/skills/skills/aranej/clawd-docs-v2

Ensure that your local environment has read permissions for the ~/clawd/data/ directory, as the skill heavily relies on local storage for the index and snippet caching mechanisms.

Use Cases

This skill is indispensable for developers and operators interacting with ClawdBot. Use it when:

• You need quick syntax checks for configuration files or telegram setup protocols.
• You want to troubleshoot OAuth token expiration or credential issues without scouring the web.
• You are performing system updates or restarts and need the official procedure guidelines.
• You require a memory/vector search context for long-running agent tasks.
• You need to cross-reference configuration providers (e.g., Discord or Telegram) during runtime.

Example Prompts

1. "How do I configure the access control for my Telegram bot using the allowFrom setting?"
2. "Check the current update procedure for ClawdBot and see if there are any pending configuration changes."
3. "I am getting an OAuth error with my credentials; can you search the documentation for troubleshooting steps?"

Tips & Limitations

To maintain peak efficiency, always prioritize the golden snippets. They are curated for the most frequent support queries and incur zero token costs. If a snippet is not available, utilize the local search index rather than performing a fresh web fetch immediately. Remember that the web_fetch tool should only be used as a final fallback; it is extremely powerful but should not be abused to keep the system performant. Always check the TTL (time-to-live) status for cached pages to ensure the information provided is not stale, especially if you are working with rapidly evolving platform features.