Skip to main content

Circuit Breaker Labs GitHub Actions

Integrate AI safety evaluations directly into your CI/CD pipeline with GitHub Actions that connect to the Circuit Breaker Labs API.

What are Circuit Breaker Labs Actions?

Circuit Breaker Labs provides GitHub Actions for automated AI model safety testing. Each action maps to a specific evaluation endpoint in the Circuit Breaker Labs API, allowing you to test system prompts and fine-tuned models for safety vulnerabilities as part of your development workflow.

Single-Turn Evaluations

Test system prompts and fine-tuned models with single-turn conversations

Multi-Turn Evaluations

Evaluate models across multi-turn conversation flows

System Prompts

Validate system prompt safety with OpenRouter models

Fine-Tuned Models

Test OpenAI fine-tuned models for safety compliance

Key Features

Configurable Thresholds

Set failure thresholds for actions and individual test cases

Test Case Groups

Run specific test case groups or all available tests

Detailed Reports

Get comprehensive failure reports with safety scores

CI/CD Integration

Seamlessly integrate with your GitHub workflows

Available Actions

All four actions provide comprehensive AI safety testing for different use cases:

Get Started

Quickstart Guide

Set up your first AI safety evaluation in under 5 minutes

Build docs developers (and LLMs) love