InputGuardrail¶

This docs was updated at: 2026-02-23

com.paragon.agents.InputGuardrail · Interface

Validates user input before agent processing.

Input guardrails are executed at the start of each agent interaction, before the LLM is called. They can be used to:

Filter out sensitive information (passwords, PII)
Validate input format or length
Enforce topic restrictions
Rate limit based on context state

Usage Example¶

InputGuardrail noPasswords = (input, ctx) -> {
    if (input.toLowerCase().contains("password")) {
        return GuardrailResult.failed("Cannot discuss passwords");
    }
    return GuardrailResult.passed();
};
Agent agent = Agent.builder()
    .addInputGuardrail(noPasswords)
    .build();

See Also

OutputGuardrail
GuardrailResult

Since: 1.0

Methods¶

`validate`¶

GuardrailResult validate(@NonNull String input, @NonNull AgenticContext context)

Validates the user input.

Parameters

Name	Description
`input`	the user's input string
`context`	the current agent context (for state-based validation)

Returns

the validation result

`named`¶

static @NonNull InputGuardrail named(@NonNull String id, @NonNull InputGuardrail impl)

Wraps a guardrail implementation with a named ID for blueprint serialization.

Use this when defining guardrails as lambdas that need to be serializable. The guardrail is registered in the GuardrailRegistry and can be reconstructed during deserialization.

InputGuardrail guard = InputGuardrail.named("no_passwords", (input, ctx) -> {
    if (input.contains("password")) return GuardrailResult.failed("No passwords!");
    return GuardrailResult.passed();
});