dify-workflow

Overview

Dify is an open-source platform for building LLM applications with a visual workflow builder. It supports chatbots, agents, text generators, and complex workflows with knowledge bases, tools, and conditional logic.

Capabilities

• Build LLM apps visually (chatbot, agent, text generator, workflow)
• Create and manage knowledge bases with document upload
• Design workflows with nodes: LLM, knowledge retrieval, code, HTTP, conditional
• Deploy apps via API, embed, or shareable link
• Manage prompts, variables, and conversation memory
• Integrate with OpenAI, Azure, local models via Ollama

When to Use

Trigger phrases:

• "dify workflow"

• "Dify AI workflow platform — LLM apps, knowledge bases, agents, workflow orchestr"

• Building LLM applications without coding

• Needing visual workflow design for AI pipelines

• Wanting knowledge base RAG without infrastructure setup

• Prototyping AI features quickly

• Deploying AI apps as APIs

When NOT to Use

• Task is outside your authorization scope
• You need to implement controls (use implementing-* skills)
• Task is about analysis, not action (use analyzing-* skills)
• You don't have access to target systems
• Task requires compliance expertise (consult professionals)
• Task is about defense, not offense (use defensive skills)

Pseudo Code

# Example workflow for this skill
def execute(input_data):
    # Step 1: Validate input
    if not input_data:
        raise ValueError("Input data is required")

    # Step 2: Process core logic
    result = process(input_data)

    # Step 3: Validate output
    validate_output(result)

    return result

App Types

Type	Use Case
Chatbot	Multi-turn conversation with memory
Agent	Tool-using autonomous assistant
Text Generator	Single input → output transformation
Workflow	Complex multi-step pipeline

Workflow Node Design

Start → Knowledge Retrieval → LLM → Conditional → HTTP Request → End
                                    ↓
                              Code Processing → End