unfreeze

Warn

Audited by Socket on Mar 28, 2026

1 alert found:

Security
SecurityMEDIUM
SKILL.md

SUSPICIOUS: the skill is internally coherent but its stated purpose is to disable an existing safety mechanism by removing path-boundary restrictions. There is no exfiltration or supply-chain behavior, yet the local state change weakens protections and materially increases downstream agent access.

Confidence: 94%Severity: 88%
Audit Metadata
Analyzed At
Mar 28, 2026, 12:12 PM
Package URL
pkg:socket/skills-sh/smashandclash%2Fagent-studio%2Funfreeze%2F@be20d1deda9d9ba58c7d28fc2e775c8fde299922
Security Audit — socket — unfreeze