ITBench

Run `npx skills add itbench-hub/ITBench --all -g -y` in your terminal to install all skills in this suite globally.

ITBench measures how well automated systems handle real-world IT tasks like fixing Kubernetes outages, checking security compliance, and reducing cloud costs. It provides ready-made incident scenarios and a public leaderboard for comparison.

The fault-scaffolding skill walks you through describing an incident, picking a target service, and writing the Ansible injection code. The scenario-scaffolding skill then applies that fault to a service and generates the ground truth files.

Yes. Scenarios run on Kubernetes, and the repository includes setup playbooks for Minikube, Kind, and Kops clusters. Fully managed hosted environments are also available so you can skip local setup.

Yes. You run your system against the managed scenarios on your own machine and submit the outputs for evaluation. Results are published on the SRE and CISO leaderboards.