# Tripwire

> A guardrail signal that halts or blocks an agent run when input, output, or a tool action fails a policy check.

- Category: Evaluation & Safety
- Stage: Rising
- Also called: Guardrail tripwire, Block signal
- Canonical: https://www.joelcsaji.com/ai-dictionary/tripwire

## Plain English

It is the red wire that stops the workflow before the risky action continues.

## Example

A guardrail detects a request for disallowed financial advice and triggers a tripwire before the final answer is sent.

## Why it matters

Tripwires make safety logic explicit, observable, and testable in agent systems.

## Related

- [Guardrails](https://www.joelcsaji.com/ai-dictionary/guardrails)
- [Action-Time Authorization](https://www.joelcsaji.com/ai-dictionary/action-time-authorization)
- [Human-in-the-Loop](https://www.joelcsaji.com/ai-dictionary/human-in-the-loop)

## Sources

- [OpenAI Agents SDK: Guardrails](https://openai.github.io/openai-agents-python/guardrails/)