Interface GuardrailInputOutcome

Result of running input guardrails.

Contains the potentially modified input and all evaluation results. Check evaluation.action to determine if processing should continue.

interface GuardrailInputOutcome {
    sanitizedInput: AgentOSInput;
    evaluation?: null | GuardrailEvaluationResult;
    evaluations?: GuardrailEvaluationResult[];
}

Index

Properties

sanitizedInput evaluation? evaluations?

Properties

sanitizedInput

sanitizedInput: AgentOSInput

Input after all sanitization (may be modified from original)

`Optional` evaluation

evaluation?: null | GuardrailEvaluationResult

The last evaluation result (for backwards compatibility)

`Optional` evaluations

evaluations?: GuardrailEvaluationResult[]

All evaluation results from all guardrails