add-benchmark
Warn
Audited by Socket on Sep 1, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The skill’s purpose is coherent, but it delegates to an unspecified agent and processes arbitrary third-party GitHub repositories while inferring and running repository-derived tests. There is no clear credential theft or malicious exfiltration, yet the combination of untrusted external content plus execution makes this medium risk.
Confidence: 84%Severity: 56%
Audit Metadata