Unified Search
Warn
Audited by Socket on Apr 8, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: the stated purpose is coherent for a web-search skill, but the trust model is not. The skill asks Claude Desktop to launch an insufficiently verified `ubersearch` CLI via `bunx` and forwards provider API keys into that process; under the required scoring rules, an unverifiable external CLI receiving credentials is high risk even without proof of malicious behavior.
Confidence: 87%Severity: 88%
Audit Metadata