container-escape-techniques

SUSPICIOUS/HIGH-RISK skill. Its stated purpose is explicitly to break out of containers and compromise the host, which is coherent internally but inherently dangerous for an AI agent. The main concern is offensive security enablement and destructive host-level actions, compounded by reliance on unpinned third-party tools and transitive loading of more attack-focused skills. No clear external credential-harvesting endpoint is shown, so this is not confirmed malware, but it is a high-risk exploit skill.

This fragment is an explicit container/Kubernetes escape and host/node compromise playbook that can be used to obtain root-level host execution, steal high-value secrets (shadow passwords, SSH keys, kubelet config, cluster certificates), and establish persistence (SSH authorized_keys and cron backdoors) including reverse-shell style command-and-control. It is not obfuscated and is directly operational for offensive use, making it extremely unsafe if included in any software supply chain artifact.