ui-verification
Warn
Audited by Socket on Oct 3, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The skill’s capabilities are mostly aligned with its stated UI-testing purpose, and the autonomy-abuse flag looks like a false positive. However, it routes browser control and credentials through an external MCP wrapper, creating moderate supply-chain and credential-forwarding risk even though the component appears Amazon-aligned and officially documented.
Confidence: 83%Severity: 53%
Audit Metadata