Helps you deliberately break things in production-like environments to find weaknesses before they find you. You'd reach for this when you want to test how your system handles network latency, service failures, resource exhaustion, or other real-world chaos. The approach is straightforward: inject failures, observe what breaks, fix it, repeat. Honestly, it's one of those things that feels reckless until you've had a 3am outage that chaos testing would have caught. Works best when you have monitoring in place to actually see what happens when you start yanking cables.
npx skills add https://github.com/alirezarezvani/claude-skills --skill chaos-engineering