insight-engine
Warn
Audited by Socket on Aug 29, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The core research-synthesis behavior is coherent and proportionate, with no credential collection or exfiltration. The main risks are transitive skill installation instructions and unpinned GitHub-based skill loading through the `skills` CLI, plus exposure to indirect prompt injection from untrusted research text while retaining file-write ability.
Confidence: 88%Severity: 53%
Audit Metadata