computer-use-agents
Warn
Audited by Socket on Sep 10, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The skill is largely coherent with its stated purpose and strongly emphasizes sandboxing and confirmation, but it still teaches high-impact autonomous computer control, forwards screenshots/page data to model APIs, and includes a weakly-sanitized shell execution path plus file-editing capabilities. This is not confirmed malware, but it is a high-risk skill whose execution footprint can affect real systems if used outside the recommended sandbox.
Confidence: 91%Severity: 71%
Audit Metadata