skyvern
Warn
Audited by Socket on Jun 13, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The core purpose and capabilities are broadly aligned: official PyPI install, same-org API endpoint, and browser automation features match the skill description. Risk is elevated because the skill enables autonomous actions on arbitrary third-party sites, may handle credentials and sensitive page data through a hosted service, and official docs mention an optional third-party `mcp-remote` bridge that expands credential trust. Not malware, but high-impact and medium-risk for agent use.
Confidence: 100%Severity: 60%
Audit Metadata