Skip to main content

Overview

This guide covers all validation rules that must be satisfied when creating or updating an agent. These rules ensure the integrity and proper functioning of your workflow.

General Workflow Validation

Assignment Rules

  1. Minimum Requirement
    • At least one assignment rule must exist

Task Dependencies

  1. Cyclic Dependencies
    • No circular dependencies allowed between tasks
  2. Valid Connections
    • Tasks must be properly connected
    • Output parameters must match input requirements

Task-Specific Validation

1. AI Task Validation

2. Human Task Validation

3. App Task Validation

4. Coder Task Validation

Parameter Validation

Input Parameters

1. Basic Types

2. Array Type

3. Object Type

Output Parameters

1. Basic Types

2. File Types

Data Source Validation

1. Memory Data Source

  • Default data source
  • No additional validation required

2. Data Lake Source

Type-Specific Validation

1. String Values

  • Must be valid string
  • Length constraints if specified

2. Numeric Values

3. Boolean Values

  • Must be true/false
  • String “true”/“false” accepted

4. Array Values

  • Must be valid JSON array
  • Items must match specified type
  • Nested validation for complex types

5. Object Values

  • Must be valid JSON object
  • Required properties must exist
  • Property types must match specifications

Common Validation Patterns

1. Required Fields

2. Conditional Requirements

3. Type Compatibility

Error Messages

Common validation errors and their meanings:

Best Practices

  1. Pre-validation
    • Validate input data before submission
    • Check type compatibility
    • Verify required fields
  2. Error Handling
    • Handle validation errors gracefully
    • Provide clear error messages
    • Log validation failures
  3. Testing
    • Test edge cases
    • Verify all required fields
    • Check type conversions