Skip to content
TakoAPI
Back to skills

dify-kb-search

by xiaowenzhou

Search & Research
873Downloads reported by clawskills.sh
0Views
0 likesSign in to like this skill
—Not yet rated
Your rating

Brief

Search Dify Knowledge Base (Dataset) to get accurate context for RAG-enhanced answers.

Description

What This Skill Does

Connects AI agents to Dify knowledge bases via the Dataset API for RAG context retrieval. Supports listing available datasets and searching them with hybrid, semantic, or keyword strategies. Returns scored content chunks for use in downstream AI responses.

Dify's Dataset API is not natively accessible from agents, so this skill removes the need to write and maintain custom retrieval integration.

When to Use It

  • Searching internal product docs before generating a support reply
  • Pulling API reference context for an agent answering developer questions
  • Building a Q&A bot over company knowledge without reindexing data elsewhere
  • Auto-discovering available datasets when a dataset ID is unknown
  • Switching between semantic and keyword search on the same query to compare results

Example Workflow

INPUT

User asks: What are the system requirements?

AGENT
  1. 1

    Call dify_list to enumerate available datasets

  2. 2

    Select the most relevant dataset by name

  3. 3

    Call dify_search with the user's query, top_k=5, and hybrid_search method

  4. 4

    Extract content chunks from the results array

  5. 5

    Pass the ranked chunks as context to the response generation step

OUTPUT

Scored content chunks from the knowledge base, ready to inject into a prompt as RAG context

Requirements

  • DIFY_API_KEY environment variable (from Dify Settings → API)
  • DIFY_BASE_URL environment variable (URL of your Dify instance)
  • Dify account or self-hosted Dify instance with at least one dataset configured

Installation

clawhub install xiaowenzhou/dify-kb-search

Run the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.

Links