bland-ai
Warn
Audited by Socket on Apr 26, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the core capability fits a Bland AI integration, and the CLI comes from an official registry with a consistent publisher relationship. However, the skill is internally inconsistent about Bland's documentation, routes all access through Membrane as an intermediary, and enables impactful telephony actions through a broad proxy-capable third-party CLI. This is better classified as a medium-risk wrapped integration than a clean direct Bland skill.
Confidence: 88%Severity: 58%
Audit Metadata