A specialized tool for local Chinese language knowledge base retrieval and evidence extraction via the LibRAG API.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install librag-knowledge-recall
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 librag-knowledge-recall 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 LibRAG Chinese Knowledge Recall?
The LibRAG Chinese Knowledge Recall skill is designed to empower AI agents with the ability to retrieve information from local knowledge bases using the LibRAG framework. This skill is particularly optimized for Chinese language contexts, providing a bridge for agents to access private documentation, perform semantic search, and extract specific evidence fragments. By leveraging Openclaw Skills, developers can integrate sophisticated Retrieval-Augmented Generation (RAG) capabilities without needing to manually handle complex API interactions.
This skill supports various retrieval modes including vector search, full-text search, and hybrid models. It is ideal for scenarios where data privacy and localized knowledge are paramount, allowing agents to find answers, locate original sources, and provide grounded responses based on verified internal data repositories.
LibRAG Chinese Knowledge Recall Use Cases
- Retrieving specific evidence or original paragraphs from local documentation repositories.
- Powering document-based question answering systems for technical or legal queries.
- Extracting citations and source text to ensure AI responses are verifiable.
- Performing hybrid data recall to combine the precision of keyword search with the depth of vector search.
- Grounding AI reasoning tasks in specific, user-defined knowledge base contexts.
How LibRAG Chinese Knowledge Recall Works
- The AI agent identifies a user request that requires searching or recalling information from a specific knowledge base.
- The system verifies the availability of the recall.py script and the required configuration file containing API credentials.
- The agent triggers a Python CLI command, passing the user's question and optional parameters like knowledge base IDs or recall modes.
- The script interfaces with the local LibRAG API endpoint (/api/v1/librag/knowbase/recall) to execute the search.
- The skill processes the results, including scores and source tree structures, and returns a structured JSON response to the agent.
- The agent utilizes the recalled data to provide an accurate, document-supported answer to the user.
LibRAG Chinese Knowledge Recall Setup
To utilize this skill within Openclaw Skills, ensure you have Python installed and the LibRAG service running locally. You must provide a valid config.json in the skill directory.
# Example config.json structure
{
"base_url": "http://localhost:8080",
"api_key": "your_secret_key",
"kb_id": "1",
"recall_mode": "hybrid"
}
Run a retrieval task using the following command:
python scripts/recall.py --config ./config.json --question "您的查询问题"
LibRAG Chinese Knowledge Recall Data Schema & Taxonomy
The skill manages data through a configuration file and outputs detailed retrieval metadata.
| Attribute | Type | Description |
|---|---|---|
question |
String | The query or condition used for the search. |
kb_id |
Integer | The identifier of the target knowledge base. |
recall_mode |
Enum | The retrieval method: vector, hybrid, or reasoning. |
response.data |
Object | The actual segments or tree structure retrieved from LibRAG. |
score_threshold |
Float | The minimum score required for a result to be returned. |
has_source_text |
Boolean | Indicates if the original source text is included in the output. |
LibRAG Chinese Knowledge Recall Advanced Features
- Reasoning Enhance Mode: Toggle advanced reasoning to improve the accuracy of complex query understanding.
- Tree Structure Returns: Maintain document hierarchy by returning results in a structured tree format rather than flat lists.
- Dynamic Parameter Overrides: Override default configurations directly via CLI arguments for specific search tasks.
- Multi-K Retrieval: Independently configure top-k parameters for both vector and full-text search components.
- Automatic Error Handling: Built-in recognition for 401, 403, and 404 status codes to provide clear feedback on authentication or resource issues.
SKILL.md
Loading
Related Openclaw Skills
AI 3D Generation
Generate, convert, and scale high-fidelity 3D models from text or images using the Neural4D API engine.
Content Production
An automated execution engine that transforms raw topics into publish-ready, SEO-optimized articles and technical guides.
Campaign Analytics
A production-grade suite for analyzing marketing campaign performance through attribution modeling, funnel conversion, and ROI calculation.
Agent Control
A command-driven interface for orchestrating isolated OpenClaw agents directly through chat shorthand.
Telegram Cross-Instance Communication
A communication framework for Openclaw instances to interact, mention, and message each other via Telegram bots.
Attorney AI Agent Skill
A professional AI guidance skill for navigating legal professional engagements and optimizing attorney-client relationships.
Featured*
