token-audit
Warn
Audited by Socket on Sep 24, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The skill’s purpose and permissions are mostly coherent for a token-audit workflow, and there is no clear credential theft or exfiltration behavior. However, it executes external CLIs via unpinned `npx`, and the Terrazzo command does not match the official documented invocation, creating a meaningful supply-chain/provenance inconsistency for a skill that could otherwise have stayed purely local and read-only.
Confidence: 88%Severity: 55%
Audit Metadata