bland-ai
Warn
Audited by Socket on Apr 29, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the skill is partly coherent as a Membrane connector guide, and its install path is official npm-based, but it inaccurately claims Bland AI lacks public docs and routes authentication/data through Membrane rather than directly to Bland. That intermediary architecture may be expected for Membrane skills, yet it materially expands trust and enables real-world telephony actions, so overall risk is medium rather than benign.
Confidence: 86%Severity: 68%
Audit Metadata