Interface IError

Detailed information about a specific validation error

Each error provides granular, actionable information about validation failures, designed to be immediately useful for both human developers and AI systems. The error structure follows a consistent format that enables precise identification and correction of type mismatches.

This error format is particularly valuable for AI function calling scenarios, where LLMs need to understand exactly what went wrong to generate correct parameters. The combination of path, expected type name, actual value, and optional human-readable description provides the AI with comprehensive context to make accurate corrections, which is why ILlmFunction.validate() can achieve such high success rates in validation feedback loops.

The value field can contain any type of data, including undefined when dealing with missing required properties or null/undefined validation scenarios. This allows for precise error reporting in cases where the AI agent omits required fields or provides null/undefined values inappropriately.

Real-world examples from AI function calling:

{
  path: "$input.member.age",
  expected: "number",
  value: "25"  // AI provided string instead of number
}

{
  path: "$input.count",
  expected: "number & Type<'uint32'>",
  value: 20.75  // AI provided float instead of uint32
}

{
  path: "$input.categories",
  expected: "Array<string>",
  value: "technology"  // AI provided string instead of array
}

{
  path: "$input.id",
  expected: "string & Format<'uuid'>",
  value: "invalid-uuid-format"  // AI provided malformed UUID
}

{
  path: "$input.user.name",
  expected: "string",
  value: undefined  // AI omitted required property
}