The Agent Skills Directory

[COMMAND_EXECUTION]: The skill contains multiple examples of shell commands and scripts designed to inject failures. This includes using the AWS CLI to terminate EC2 instances or reboot RDS databases, and using kubectl to delete pods or drain nodes in Kubernetes clusters. These are the core functional requirements for chaos engineering experiments.
[COMMAND_EXECUTION]: In references/infrastructure-chaos.md, the skill provides examples of using sudo to perform system-level tasks, such as installing the stress-ng utility and modifying the /etc/hosts file to simulate DNS resolution failures. These administrative actions are explicitly tied to the skill's purpose of infrastructure resilience testing.
[EXTERNAL_DOWNLOADS]: The GitHub Actions workflow example in references/chaos-tools.md includes a command to download and apply a Litmus Chaos operator manifest directly from a remote GitHub Pages URL (litmuschaos.github.io). This is a standard installation method for the Litmus framework.
[SAFE]: The skill incorporates extensive safety guidelines, including a mandatory safety checklist, steady-state verification, blast radius caps, and requirements for automated rollback procedures. No malicious obfuscation, hidden exfiltration, or unauthorized persistence mechanisms were detected.

chaos-engineer