​
Validation Rules

​
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
    {
    "rule_type": "Human Feedback|Exception",
    "assignment_type": "User|Team",
    "assignee_id": "valid_id"
}

​
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

{
    "name": "required|string",
    "instructions": "required|string",
    "input_parameters": "required|array",
    "expected_output": "required|array"
}

​
2. Human Task Validation

{
    "name": "required|string",
    "instructions": "required|string",
    "error_policy": "optional|RAISE|IGNORE|RETRY"
}

​
3. App Task Validation

{
    "name": "required|string",
    "instructions": "required|string",
    "provider": "required|string",
    "tool_name": "required|string",
    "tool_id": "required|integer",
    "input_parameters": "required|array",
    "expected_output": "required|array"
}

​
4. Coder Task Validation

{
    "name": "required|string",
    "instructions": "required|string",
    "code_artifact_id": "required|integer",
    "input_parameters": "required|array",
    "expected_output": "required|array"
}

​
Parameter Validation

​
Input Parameters

​
1. Basic Types

{
    "name": "required|string",
    "type": "required|STRING|INTEGER|FLOAT|BOOLEAN",
    "description": "required|string",
    "required": "boolean",
    "source": "required|task_config|system_generated|human_input",
    "value": "conditional|based_on_type"
}

​
2. Array Type

{
    "type": "ARRAY",
    "items": {
        "type": "required|STRING|INTEGER|FLOAT|BOOLEAN|ARRAY|OBJECT",
        "items": "required_for_nested_array",
        "properties": "required_for_object"
    }
}

​
3. Object Type

{
    "type": "OBJECT",
    "properties": [
        {
            "name": "required|string",
            "type": "required|STRING|INTEGER|FLOAT|BOOLEAN",
            "description": "required|string",
            "required": "boolean"
        }
    ]
}

​
Output Parameters

​
1. Basic Types

{
    "name": "required|string",
    "type": "required|BOOLEAN|INTEGER|FLOAT|STRING|DATETIME",
    "description": "required|string"
}

​
2. File Types

{
    "type": "required|TEXT|MARKDOWN|HTML|CODE|PDF|CSV|XML|FILE|PPT",
    "description": "required|string"
}

​
Data Source Validation

​
1. Memory Data Source

  • Default data source
  • No additional validation required

​
2. Data Lake Source

{
    "data_source": "DATA_LAKE",
    "data_lake_id": "required|string",
    "type": "must_be_ARRAY"
}

​
Type-Specific Validation

​
1. String Values

  • Must be valid string
  • Length constraints if specified

​
2. Numeric Values

# Integer
if type == "INTEGER":
    must be valid integer or convertible string

# Float
if type == "FLOAT":
    must be valid float or convertible string/integer

​
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

{
    "required_field": "value is mandatory",
    "optional_field": "can be null"
}

​
2. Conditional Requirements

{
    "type": "ARRAY",
    "items": "required when type is ARRAY",
    "properties": "not allowed when type is ARRAY"
}

​
3. Type Compatibility

{
    "output": "type STRING",
    "connected_input": "must accept STRING"
}

​
Error Messages

Common validation errors and their meanings:
ErrorMeaningSolution
Missing Required FieldMandatory field not providedAdd required field
Invalid TypeValue type mismatchCorrect value type
Cyclic DependencyCircular task referenceRestructure workflow
Invalid AssignmentMissing or invalid assignment ruleAdd valid assignment

​
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