openclaw-awd-arena

SUSPICIOUS. The skill’s stated purpose is coherent for a cyber-competition platform, but its footprint is high risk: it equips AI agents for offensive security actions, executes generated exploit/defense commands, recommends Docker socket access, and relies on unverifiable or unofficial default container images. The documentation-context package installs are low concern, but the image provenance and broad host/container control make the overall skill risky and disproportionate for general agent use.