A powerful AI-enhanced system for converting unstructured agent logs into structured, searchable Chinese knowledge bases.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install zh-knowledge-manager
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 zh-knowledge-manager 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 zh-knowledge-manager?
The zh-knowledge-manager is a sophisticated skill designed to transform ephemeral agent activity into long-term organizational memory. By integrating this skill into Openclaw Skills, developers can automate the transition from raw logs to structured knowledge using a combination of deterministic classification and AI-driven insights. It is particularly optimized for Chinese language processing, ensuring that project progress, technical issues, and research findings are preserved with high accuracy.
This tool leverages deterministic scripts for core functions like hash-based deduplication while offering optional AI enhancements for semantic analysis. Whether you are managing complex infrastructure changes or tracking business logic discussed in chat, this component of Openclaw Skills provides a reliable framework for personal and team knowledge management.
zh-knowledge-manager Use Cases
- Synchronizing daily agent logs into a structured long-term memory store.
- Extracting actionable knowledge and business rules from lengthy dialogue dumps.
- Automating the tagging of technical notes using native Chinese segmentation.
- Maintaining a clean, deduplicated infrastructure and configuration changelog.
- Generating reviewable drafts of knowledge base entries from raw session data.
How zh-knowledge-manager Works
- The agent records activity in daily markdown files using a standardized prefix-based format.
- Users or automated heartbeats trigger the sync command to process recent logs.
- The system performs hash-based deduplication to ensure no redundant entries are added to the knowledge base.
- If enabled, AI features analyze the content for semantic similarity and suggest intelligent tags using jieba segmentation.
- Knowledge extracted from dialogue is placed in a draft folder for human verification before final import.
zh-knowledge-manager Setup
To get started with this skill in your Openclaw Skills environment, follow these steps:
# Initialize the configuration and kb directory
node km.js init
# Install required dependencies for auto-tagging
npm install @node-rs/jieba
# Set up your environment variable for AI features
export SILICONFLOW_API_KEY='your_api_key_here'
zh-knowledge-manager Data Schema & Taxonomy
The skill organizes data using a strict prefix system and markdown structure to maintain clarity within Openclaw Skills.
| Data Type | Location | Format/Prefix |
|---|---|---|
| Daily Logs | memory/YYYY-MM-DD.md |
### [PREFIX:Name] Title |
| Knowledge Drafts | output/kb-draft-*.md |
Markdown with tags |
| Supported Prefixes | PROJECT, ISSUE, INFRA, CONFIG, RESEARCH, KB, BOSS | Categorized headers |
| Deduplication | Internal Index | Hash-based & Semantic (bge-m3) |
zh-knowledge-manager Advanced Features
- Semantic deduplication using bge-m3 embeddings to identify similar concepts beyond keyword matching.
- Automated Chinese keyword extraction and tagging powered by the jieba segmentation engine.
- LLM-powered knowledge extraction capable of distilling complex conversations into concise summaries.
- Dry-run capabilities to preview synchronization results without modifying the permanent knowledge base.
- Heartbeat integration for fully automated, daily background maintenance within Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Draw.io Diagram Generation
Automatically generate and export professional-grade architectural, database, and process diagrams directly from your AI agent.
Daily Tang Poem
A comprehensive AI agent skill for daily Tang poetry learning, featuring interactive recitation verification and Ebbinghaus-based review cycles.
Book of the Day
A daily book oracle skill that delivers uplifting book recommendations and poetic fortunes to inspire your day.
SOTA Tracker
An automated database tracking the latest state-of-the-art AI models across LLM, image, video, and audio categories with daily updates.
TikTok Scraping with yt-dlp
A comprehensive command-line framework for automated TikTok video downloading, profile crawling, and metadata extraction.
Poseidon OTC Skill
A trustless P2P trading protocol for Solana that enables AI agents and humans to execute atomic token swaps through secure on-chain escrow.
METADATA
Featured*
