Description

Search and manage knowledge bases using LightRAG API. Supports multiple servers, context-aware writing, and direct information retrieval. Use when the user wants to query a LightRAG-powered knowledge base or use it as context for tasks.

README (SKILL.md)

LightRAG Skill

Name: LightRAG Search Skill
Author: ruslanlanket

This skill allows you to interact with one or more LightRAG API servers. You can perform queries in various modes (local, global, hybrid, mix, naive) and use the retrieved context for further processing.

Configuration

The skill uses a configuration file at ~/.lightrag_config.json to store server details. Format:

{
  "servers": {
    "alias1": {
      "url": "http://server1:9621",
      "api_key": "optional_key"
    },
    "alias2": {
      "url": "http://server2:9621",
      "api_key": "optional_key"
    }
  },
  "default_server": "alias1"
}

Workflows

1. Direct Search

To find information, use scripts/query_lightrag.py. Modes: local, global, hybrid, mix, naive.

2. Using Context for Writing

To use a knowledge base as context (e.g., for a test or article):

Run query_lightrag.py with the --only-context flag.
Pass the resulting context to your writing task/model.

Reference

See API_DOCS.md for endpoint details.

Usage Guidance

This skill appears to do what it says (configure and query LightRAG servers) but has two practical security issues you should consider before installing or adding credentials: (1) The script disables SSL certificate verification when making requests. That means a malicious or compromised network or server could intercept or tamper with requests and responses. Prefer only adding trusted servers, or modify the script to validate certificates (remove the ssl.CERT_NONE override) or add an explicit flag to opt-out only when needed. (2) API keys (if used) are stored unencrypted in ~/.lightrag_config.json. If you must store keys, restrict the file permissions (e.g., chmod 600) or use a secure credential store / OS keyring instead. Additional suggestions: review the script before use, avoid sending sensitive queries to untrusted servers (a configured server could log or exfiltrate queries/responses), and consider running network-restricted/isolated instances if you will point the skill at untrusted endpoints. If you want higher assurance, ask the author to remove the SSL bypass and to add guidance for secure storage of API keys.

Capability Analysis

Type: OpenClaw Skill Name: lightrag Version: 1.0.1 The skill is classified as suspicious primarily due to the `scripts/query_lightrag.py` file explicitly disabling SSL certificate verification (`ctx.check_hostname = False`, `ctx.verify_mode = ssl.CERT_NONE`). This critical security vulnerability makes all network communication with the configured LightRAG servers susceptible to Man-in-the-Middle (MITM) attacks, potentially exposing API keys and query data to interception or modification. While the skill's stated purpose and other functionalities appear benign, this significant security flaw introduces a high risk.

Capability Assessment

✓ Purpose & Capability

Name/description match the provided code and instructions: the script configures servers and queries a LightRAG /query endpoint and supports the documented modes.

⚠ Instruction Scope

SKILL.md and the script limit actions to reading/writing ~/.lightrag_config.json and making HTTP POSTs to configured servers, which aligns with the purpose. However, the code unconditionally disables SSL certificate verification (ssl.CERT_NONE) when contacting servers — a risky behavior not mentioned in SKILL.md and which can enable MitM attacks or accidental data exposure.

✓ Install Mechanism

No install spec; this is an instruction-only skill with one included script. Nothing is downloaded or installed by the skill itself.

ℹ Credentials

No environment variables or unrelated credentials are requested. The script stores optional API keys in ~/.lightrag_config.json in plaintext — reasonable for functionality but provides no encryption or guidance on file permissions.

✓ Persistence & Privilege

The skill does not request persistent/global privileges (always=false). It only reads/writes its own config file and does not modify other skills or system-wide settings.

Version History

v1.0.1

Slightly improved work with the API. - No user-visible changes in this version. - No file or documentation updates were detected.

v1.0.0

A professional skill for integrating with LightRAG (Simple and Fast Retrieval-Augmented Generation) API servers. This skill enables your agent to perform deep searches across knowledge bases built on Knowledge Graphs and Vector Retrieval. Key Features: • Multi-server Support: Connect to multiple LightRAG instances. Assign a unique alias to each server for quick switching. • Advanced Query Modes: • local: Focuses on context-dependent information. • global: Utilizes global knowledge across the entire database. • hybrid: Combines both local and global retrieval methods. • mix: Integrates Knowledge Graph and vector retrieval for balanced results. • naive: Performs standard full-text/vector search. • Context Extraction: A specialized mode to retrieve "raw" context, perfect for providing background data for writing tasks, test generation, or summarization. • Security First: All sensitive data (URLs and API keys) are stored locally on your server and are never shared when the skill is published or distributed. Usage: 1. Configure a server: lightrag config --alias "my_base" --url "https://your-lightrag-url.com" --key "your_api_key" --mode local 2. Search for information: lightrag query "Your question here" --alias "my_base" 3. Retrieve context for tasks: lightrag query "Topic" --alias "my_base" --only-context Requirements: • A running LightRAG server. • Network access to the server's API.

Metadata

Slug lightrag

Version 1.0.1

License —

All-time Installs 4

Active Installs 4

Total Versions 2

Frequently Asked Questions

What is LightRAG Search Skill?

Search and manage knowledge bases using LightRAG API. Supports multiple servers, context-aware writing, and direct information retrieval. Use when the user wants to query a LightRAG-powered knowledge base or use it as context for tasks. It is an AI Agent Skill for Claude Code / OpenClaw, with 2404 downloads so far.

How do I install LightRAG Search Skill?

Run "/install lightrag" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is LightRAG Search Skill free?

Yes, LightRAG Search Skill is completely free (open-source). You can download, install and use it at no cost.

Which platforms does LightRAG Search Skill support?

LightRAG Search Skill is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created LightRAG Search Skill?

It is built and maintained by RuslanLanket (@ruslanlanket); the current version is v1.0.1.

More Skills

LightRAG Search Skill