narration
Warn
Audited by Socket on Nov 5, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
The narrative-writing purpose and cue rules are coherent, and no credential collection or suspicious data routing is described. The main risk is unpinned, global installation of multiple companion skill sets from separate repositories, which expands the agent's instruction trust surface; this warrants caution but does not establish malicious intent.
Confidence: 90%Severity: 62%
Audit Metadata