review-loop
Warn
Audited by Socket on Oct 22, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The core capabilities mostly match a PR-review orchestration skill, and data flows stay largely within GitHub and local repo workflows. The main risks are the mutable raw-GitHub setup script executed via bash in Agent0 mode and the optional autonomous merge behavior; these are proportionate enough to avoid a malicious verdict, but substantial enough to rate the skill medium risk overall.
Confidence: 89%Severity: 62%
Audit Metadata