MetaClaw is a specialized framework for managing long-term memory, hybrid information retrieval, and automated skill quality assurance.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install metaclaw
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 metaclaw 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 MetaClaw?
MetaClaw serves as a robust foundation for memory systems within the ecosystem of Openclaw Skills. It addresses the challenge of information retention and retrieval by combining the strengths of semantic vector search with traditional full-text keyword indexing. This hybrid approach ensures that agents can access the most relevant context regardless of whether they are searching for abstract concepts or specific technical terms.
Beyond simple storage, MetaClaw implements a sophisticated skill evolution pipeline. It integrates PluginEval to monitor and improve the quality of other skills, aiming for high-performance badges. By automating the extraction of preferences and project states from active sessions, MetaClaw keeps an agent's internal knowledge base synchronized with real-world progress, making it an essential tool for developers building complex, long-running AI workflows.
MetaClaw Use Cases
- Implementing a persistent memory layer that evolves based on real-time project interactions.
- Performing high-accuracy documentation and session searches using a blend of semantic and keyword logic.
- Automating the maintenance of skill quality through periodic evaluations and auto-fix routines.
- Triggering scheduled data consolidation to prevent information bloat in long-term memory storage.
How MetaClaw Works
- The session counter tracks agent activity and increments on every heartbeat to monitor usage density.
- When a search is performed, the hybrid retrieval algorithm calculates a combined score from semantic vectors (via LanceDB) and full-text keyword matches.
- The memory extraction pipeline parses session logs to identify and store updates for preferences and project states.
- PluginEval runs layer-based quality checks against the skill directory to detect anti-patterns and suggest improvements.
- Upon reaching a defined session threshold, the system triggers the Auto-Dream consolidation process to summarize and compress memory files.
MetaClaw Setup
To integrate this component into your suite of Openclaw Skills, navigate to the skill directory and initialize the virtual environment:
cd ~/.openclaw/workspace/skills/metaclaw
python3 -m venv .venv-metaclaw
source .venv-metaclaw/bin/activate
pip install lancedb sentence-transformers
Initialize the vector database if necessary:
python3 skills/metaclaw/scripts/memory_vector_init.py
MetaClaw Data Schema & Taxonomy
MetaClaw organizes its data through a configuration-driven approach and structured JSON outputs for search results:
| File Path | Description | Data Type |
|---|---|---|
memory/config.yaml |
Global settings for retrieval weights and consolidation intervals | YAML |
scripts/memory-vector-wrapper.py |
The primary interface for querying memory with metadata | Python/JSON |
scripts/session-counter.sh |
State file tracking the number of active sessions | Bash/Integer |
scripts/skill-eval.py |
Quality metrics and badge status (Gold/Silver/Bronze) | Python/CLI |
MetaClaw Advanced Features
- Hybrid Scoring Algorithm: Dynamically weights semantic relevance (0.6) against keyword/FTS matches (0.4) for optimized search results.
- Auto-Fix Capabilities: Automatically refactors code and documentation to meet quality badge standards via PluginEval.
- Session-Based Consolidation Logic: Prevents memory fragmentation by triggering digests only after a specific number of sessions or time intervals.
- Heartbeat Integration: Hooks into the system heartbeat to maintain an accurate count of agent interactions for lifecycle management.
SKILL.md
Loading
Related Openclaw Skills
Save Up Then Travel
A travel-focused incentive system that transforms daily high-energy behaviors into tangible travel points and rewards.
E-Commerce Video Marketing
A comprehensive toolkit for creating and optimizing high-conversion video marketing strategies across major e-commerce platforms.
eBay Advertising
An expert AI skill designed to master eBay Promoted Listings, bidding strategies, and cross-channel ad optimization for global sellers.
Dynamic Pricing for E-Commerce
Automate e-commerce pricing strategies using demand-based, competitor-responsive, and algorithmic models to maximize profit margins and market share.
Shopify Conversion Optimization
Boost your Shopify store revenue through data-driven conversion rate optimization and funnel analysis.
Feishu Bitable Data Import
An enterprise-grade data synchronization tool that automates the batch import of CSV, Excel, and JSON files into Feishu Bitable with intelligent field type inference.
Featured*
