Core Concepts¶

Task Types¶

AI Task¶

• Handles intelligent processing using AI capabilities.

  graph LR
      Input --> AI[AI Processing] --> Output

Human Task¶

• Manages human interventions and approvals.

  graph LR
      Request --> Human[Human Decision] --> Response

App Task¶

• Integrates with external tools and applications.

  graph LR
      Data --> App[Tool Integration] --> Result

Coder Task¶

• Executes custom code artifacts.

  graph LR
      Input --> Code[Code Execution] --> Output

Parameter Types¶

Input Parameters¶

Data that flows into tasks:

• Basic types (string, number, boolean)
• Complex types (array, object)
• Special types (artifacts)

Output Parameters¶

Results produced by tasks:

• Basic data types
• File outputs
• Media outputs
• Structured data

Assignment Rules¶

Rule Types¶

1. Exception

   • Error handling
   • Special cases

2. Human Feedback

   • Approvals
   • Reviews

Assignment Types¶

1. User Assignment

   • Direct user allocation
   • Individual responsibility

2. Team Assignment

   • Group allocation
   • Round-robin distribution

Workflow Concepts¶

Dependencies¶

graph TD
    A[Task A] --> B[Task B]
    B --> C[Task C]

• Sequential execution
• Parallel processing
• Conditional flows

Error Policies¶

1. RAISE

   • Stop execution
   • Report error

2. IGNORE

   • Continue execution
   • Log warning

3. RETRY

   • Attempt recovery
   • Configurable attempts

Best Practices Overview¶

1. Task Design

   • Clear naming
   • Proper description
   • Error handling

2. Parameter Configuration

   • Type selection
   • Validation rules
   • Default values

3. Assignment Setup

   • Clear responsibilities
   • Backup assignees
   • Team distribution