container-security

Warn

Audited by Socket on Sep 15, 2026

3 alerts found:

Securityx3
SecurityMEDIUM
SKILL.md

SUSPICIOUS. The skill is internally coherent for container/Kubernetes security assessment, but its footprint is high-risk because it equips the agent with offensive security capabilities against live clusters and hosts, including escape testing and privilege escalation. Install sources are mostly official, lowering malware concerns, yet the unpinned raw Kubescape installer and active exploit workflows keep overall security risk high.

Confidence: 92%Severity: 81%
SecurityMEDIUM
payloads/container_escape_poc.md

This is security-testing documentation containing several high-impact container escape and privilege-escalation procedures. It does not itself establish persistence, exfiltrate data, or behave as package malware, but the included commands are operationally dangerous and can compromise a host or Kubernetes cluster when prerequisites are met. It should be restricted to authorized isolated environments, with the exploit snippets replaced by read-only detectors where possible.

Confidence: 99%Severity: 92%
SecurityMEDIUM
references/container_escape.md

This is an offensive container-escape reference containing highly dangerous, directly executable commands for host compromise, persistence, reverse-shell access, Docker socket abuse, cgroup escape, namespace entry, and Kubernetes privilege escalation. It is not itself package malware or obfuscated code because the fragment contains no automatic execution mechanism or concealed payload, but using these commands outside an authorized lab or engagement could cause severe compromise.

Confidence: 99%Severity: 94%
Audit Metadata
Analyzed At
Sep 15, 2026, 04:19 PM
Package URL
pkg:socket/skills-sh/hardw00t%2Fai-security-arsenal%2Fcontainer-security%2F@cdb7e011ebad32386f890ae4a3dcefa0f42429ac100d7a5407b0c49637a8c424