ClawExpert is a persistent learning system that enables AI agents to research complex topics and build indexed, citable local knowledge bases.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install clawexpert
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 clawexpert 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 ClawExpert?
ClawExpert is a sophisticated domain expert learning system designed to provide AI agents with long-running autonomous research capabilities. Unlike standard LLMs that rely solely on parametric memory, this skill allows an agent to systematically explore topics for hours—searching, downloading, and organizing authoritative materials into a structured, citable local repository. It bridges the gap between static training data and the need for up-to-date, specialized expertise.
By leveraging Openclaw Skills, ClawExpert creates a persistent knowledge layer stored directly on your machine. It features a multi-tiered source priority system, ensuring that conclusions are drawn from high-quality, peer-reviewed, or institutional sources. The system is designed for reliability, featuring interrupted session resumption, parallel subagent execution, and automated deduplication of findings.
ClawExpert Use Cases
- Conducting deep-dive research into emerging scientific or technical fields.
- Building a persistent local library of market analysis and economic frameworks.
- Generating citable reports based on institutional and primary documents.
- Automating the monitoring of specific domains through proactive learning queues.
- Importing and indexing large PDF collections for structured retrieval.
How ClawExpert Works
- The user initiates a learning task via the /clawexpert learn command, specifying a topic and time budget.
- ClawExpert decomposes the main topic into distinct subtopics and assigns search angles prioritized by source quality.
- Parallel subagents are spawned to conduct multi-round web searches and retrieve full-text content from web pages and PDFs.
- A quality filter applies a four-tier priority system to discard low-value content and SEO-heavy material.
- Collected materials are processed into markdown nodes, complete with verbatim excerpts and source metadata.
- The system merges subagent findings, deduplicates information, and builds a hierarchical knowledge tree.
- A central root index and category-specific indices are updated to provide the main agent with meta-awareness of its learned expertise.
ClawExpert Setup
To initialize the system and prepare your local environment for Openclaw Skills, run the following command:
/clawexpert init
This command creates the necessary knowledge base directory structure at ~/.openclaw/workspace/knowledge/ and initializes the state configuration and default categories.
ClawExpert Data Schema & Taxonomy
ClawExpert organizes information into a strict hierarchy to ensure efficient retrieval and traceability:
| Level | Component | Description |
|---|---|---|
| Root | _index/ |
Contains _root_index.md for global meta-awareness of learned topics. |
| Topic | slug/ |
A unique directory for each topic (e.g., rl-reasoning). |
| Raw | raw/ |
Faithful reproductions of source materials (web/PDF) named by URL hash. |
| Nodes | nodes/ |
Structured markdown files containing summaries, conclusions, and evidence links. |
| Meta | meta.json |
Stores topic status, keywords, subtopic tracking, and session history. |
ClawExpert Advanced Features
- Proactive Mode: Automatically tracks knowledge gaps during conversations and suggests or auto-starts new learning sessions.
- Superlearn: An end-to-end multi-round research workflow that continues until specific depth or time budgets are met.
- PDF Processing Engine: Intelligent routing for small, medium, and large PDFs with automated text extraction and visual element description.
- Deepen Rounds: Specialized gap analysis to expand existing knowledge trees through breadth expansion or depth questioning.
- Knowledge Radar: Manual or scheduled scans across L1 categories to identify and fill information voids.
SKILL.md
Loading
Related Openclaw Skills
SSL Monitor
An automated utility for checking and monitoring domain SSL certificate expiration status and remaining validity days.
Price Drop Tracker
An intelligent automated agent that monitors Fliggy travel product prices and executes notifications or auto-bookings when target prices are reached.
PA Status
A comprehensive health dashboard and reporting tool for monitoring the status of personal assistant networks.
Seedance Prompt Library Skill
A sophisticated AI prompt engineering system designed to search, rewrite, and synthesize high-performance cinematic video prompts for the Seedance 2.0 model.
Travel Safety & Etiquette Companion
An automated tool that generates professional, card-based HTML safety and etiquette dashboards for global travel destinations.
Marine Weather (NOAA)
A specialized MCP server providing real-time NOAA marine weather conditions, multi-day forecasts, and active maritime alerts based on GPS coordinates.
Featured*
