bili-ai-feed
Warn
Audited by Socket on Jun 16, 2026
2 alerts found:
AnomalySecurityAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the skill’s general purpose is coherent, but trust and data-flow integrity are weakened by routing through a third-party provider with inconsistent auth documentation and unclear official Bilibili endpoint support. No confirmed malware or download-execute behavior is present, but the credential path and mandatory broad web research make this a medium-risk skill.
Confidence: 100%Severity: 60%
Securityreferences/investigator-config.json
MEDIUMSecurityMEDIUM
references/investigator-config.json
Audit Metadata