> ## Documentation Index
> Fetch the complete documentation index at: https://docs.skrut.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Tests

> The scenarios Skrut runs against your agent.

A **test** is a multi-turn conversation scenario that Skrut executes against your agent to evaluate a specific behaviour.

## How scenarios are generated

Skrut generates test scenarios from two sources:

1. **Your agent's knowledge base** — grounded scenarios that probe for hallucinations and broken refusals.
2. **Red-team templates** — adversarial patterns (prompt injection, jailbreaks, PII leaks) applied to your agent's context.

## Grading

Every turn in a conversation is graded on a 0–100 scale across:

* **Accuracy** — did the agent answer correctly?
* **Refusal** — did it correctly refuse out-of-scope or harmful requests?
* **Consistency** — does it contradict itself across turns?

The per-turn grades roll up into a **Readiness Score** for the full run.