jev-answer-review
Warn
Audited by Socket on Oct 2, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: the review workflow and consent language are largely coherent with the stated purpose, and the scanner’s prompt-injection/autonomy flags look like false positives. The main issue is install/execution trust: the skill requires an ambiguously sourced `jev-mcp` server that would receive both hosted credentials and potentially private assessment data, which is disproportionate unless a verified first-party implementation is identified.
Confidence: 84%Severity: 82%
Audit Metadata