dify-kb-search

What This Skill Does

The dify-kb-search skill integrates OpenClaw AI agents with Dify knowledge base datasets. It allows your agent to perform high-quality Retrieval-Augmented Generation (RAG) by fetching precise context chunks from your uploaded documents. Instead of relying solely on the LLM's pre-trained knowledge, the agent can search internal documentation, company wikis, or technical manuals stored in Dify. It supports sophisticated search strategies including hybrid, semantic, and keyword matching, ensuring that the retrieved information is highly relevant to the user's specific query. By utilizing the dify_list and dify_search tools, agents can dynamically adapt to various information sources, making it a critical component for building intelligent support bots or documentation assistants.

Installation

To begin using this skill, ensure you have the necessary dependencies by running pip3 install requests. Next, add the skill to your project using the command: clawhub install openclaw/skills/skills/xiaowenzhou/dify-kb-search. Once installed, configure your openclaw.json file by adding your DIFY_API_KEY and the DIFY_BASE_URL within the environment variables block. This configuration enables secure communication between your agent and your Dify instance. Without these variables, the tools will be unable to authenticate or locate your hosted knowledge bases, resulting in connection errors during execution.

Use Cases

This skill is ideal for technical customer support scenarios where agents need to reference specific API guides or troubleshooting manuals. It is also highly effective for internal company portals where employees query organizational policies or HR documents. Furthermore, developers can leverage this to create documentation chatbots that allow users to search through large repositories of technical specifications effortlessly. It is particularly powerful when dealing with domain-specific knowledge that is too niche or frequently changing to be included in an LLM's base training.

Example Prompts

1. "Search the knowledge base for steps on how to reset the authentication token in our system."
2. "Look up the most recent hardware specifications for the latest product release in the API reference dataset."
3. "Based on our internal product documentation, what is the recommended procedure for handling critical security vulnerabilities?"

Tips & Limitations

For optimal results, ensure your datasets in Dify are well-indexed with proper metadata. While 'hybrid_search' is the default and recommended method, you can experiment with 'semantic_search' for concept-heavy queries or 'keyword_search' for highly specific technical terms. Note that large 'top_k' values may increase token consumption, so tune this value based on your specific context window needs. Always ensure your Dify API key has read access to the specific datasets you intend to query, and maintain consistent naming conventions for your datasets to simplify identification when using auto-discovery.