deep-interview
Warn
Audited by Socket on Aug 20, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the core interview and artifact-writing behavior is mostly aligned with the stated purpose, but the skill also acts as an orchestration hub that hands work to other skills, creating transitive trust and moderate prompt-injection exposure from repository/user content. No direct malware, credential harvesting, remote download-execute path, or clear exfiltration endpoint is present.
Confidence: 82%Severity: 52%
Audit Metadata