A high-efficiency documentation access layer for ClawdBot that prioritizes local snippets and search to minimize token usage and latency.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install clawd-docs-v2
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install clawd-docs-v2 using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Clawd-Docs v2?
Clawd-Docs v2 is a specialized skill designed to streamline how AI agents interact with technical documentation. By implementing a sophisticated three-tier architecture—comprising a local search index, pre-cached Golden Snippets, and an intelligent on-demand fetch system—it ensures that agents have immediate access to the right information while drastically reducing LLM token consumption and API latency.
This skill is essential for developers using Openclaw Skills who need a token-efficient way to handle frequently asked questions, configuration guides, and troubleshooting procedures. By serving common answers from local markdown snippets, it avoids the overhead of reloading entire documentation pages for every query, resulting in faster and more cost-effective agent interactions.
Clawd-Docs v2 Use Cases
- Retrieving instant setup guides for Telegram, Discord, and OAuth integrations.
- Reducing token consumption during long-running agent sessions by using local snippets.
- Performing keyword-based lookups via local search indexes before triggering external web fetches.
- Managing documentation freshness automatically through category-specific TTL tracking.
- Troubleshooting configuration errors using cached snippets for common error codes.
How Clawd-Docs v2 Works
- The agent identifies a documentation request and first scans the local Search Index in docs-index.json for keyword matches and synonyms.
- If a match is found, the agent checks the docs-snippets directory for a Golden Snippet, which provides high-value answers for minimal tokens.
- If no snippet is available, the agent checks the docs-cache for a full-page markdown version of the documentation.
- If the content is missing or expired (based on TTL), the skill utilizes the native web_fetch tool to retrieve the latest version from the web.
- The retrieved content is cached locally, and its metadata is updated to track freshness for future requests.
Clawd-Docs v2 Setup
To integrate this skill into your environment, ensure the native web_fetch tool is active and the data directories are initialized. Use the following commands to verify your local knowledge base:
# Verify the existence of Golden Snippets
ls ~/clawd/data/docs-snippets/
# Inspect the search index for keyword mappings
cat ~/clawd/data/docs-index.json
# Check for full-page cached documents
ls ~/clawd/data/docs-cache/
Clawd-Docs v2 Data Schema & Taxonomy
The skill organizes its knowledge base within the ~/clawd/data/ directory using the following structure:
| Component | File/Path | Description |
|---|---|---|
| Search Index | docs-index.json |
Maps keywords and synonyms to documentation paths. |
| Golden Snippets | docs-snippets/ |
Pre-fetched, low-token markdown snippets for common tasks. |
| Full Cache | docs-cache/ |
Storage for full-length documentation pages in markdown format. |
| Usage Stats | docs-stats.json |
Tracks page freshness and retrieval statistics. |
Clawd-Docs v2 Advanced Features
- Fuzzy matching logic using a comprehensive synonym dictionary for more accurate documentation discovery.
- Category-aware TTL (Time-To-Live) strategy, ranging from 1 day for update procedures to 30 days for stable references.
- Zero-cost web retrieval using the native web_fetch tool, eliminating the need for expensive third-party scraping services.
- Smart token efficiency guide that prioritizes 300-500 token snippets over 12k token full-page fetches.
- Seamless integration with Openclaw Skills workflows for automated multi-agent knowledge sharing.
SKILL.md
Loading
Related Openclaw Skills
Agency Guardian
A mindful skill designed to preserve human autonomy and critical thinking while using AI assistants.
ClawZone AI Game Arena
A game-agnostic arena skill enabling AI agents to compete in real-time matches through automated polling and RESTful API interactions.
A2A Delegation Setup
A dedicated setup and maintenance utility for the OpenClaw agent-to-agent outbound delegation plugin.
VizClaw
A specialized skill to bridge OpenClaw agent runs with VizClaw live rooms for real-time event streaming and visualization.
ArcAgent MCP
A comprehensive MCP toolset designed to automate the entire lifecycle of ArcAgent bounties from discovery to payout.
Arayun 173 Lonsdaleite Executive
An uncompromising executive agent skill designed to enforce high-value financial convergence and system law adherence.
METADATA
Featured*
