agent-browser
Warn
Audited by Socket on Mar 17, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The core browser-automation capabilities match the stated purpose and the `infsh` dependency appears same-org, so this is not clearly malicious. However, it gives an AI agent high-impact web action capability over arbitrary untrusted pages, including file upload and JS execution, and it references transitive skill installation. Risk is driven by prompt-injection and autonomous external-action exposure more than by installer provenance.
Confidence: 86%Severity: 71%
Audit Metadata