implementer
Warn
Audited by Socket on Oct 2, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The skill’s core purpose is coherent, but it expands trust to an external CLI running unsandboxed with permission checks disabled and possible autonomous git actions. Because `agy` appears to be an official Google/open-source tool, this is more a high-trust delegation and supply-chain risk than confirmed malware; the main concerns are unrestricted execution, credential/context forwarding, and possible endpoint redirection via environment overrides.
Confidence: 84%Severity: 74%
Audit Metadata