zhihu
Warn
Audited by Socket on Aug 24, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: The skill's search capability is well aligned and low-risk, but the read/write side expands to full-session cookie access and autonomous public posting. There is no clear malware behavior or remote installer chain, yet the combination of raw account cookies, third-party connection handling, and public account actions makes the overall footprint higher-risk than a typical API-only skill.
Confidence: 84%Severity: 62%
Audit Metadata