Agentic Browser 0.1.2

• Analyzed: May 12, 2026, 3:41 AM
• Findings: 1
• Version: 1.0.0
• Hash: 26653e3b4880b6392bc03c657d518b9e36013d845c70210033c00be45dc5eb44
• Source: palm
• External report: View on VirusTotal

Findings (1)

The skill is classified as suspicious due to several high-risk capabilities inherent to its web automation purpose, which could be leveraged for malicious activities if misused. Specifically, the `execute` function allows arbitrary JavaScript execution within the browser context, enabling sensitive data extraction (e.g., `document.cookie` as demonstrated in `references/authentication.md`). The `interact` function supports file uploads, and the `open` function allows proxy configuration, both of which are powerful features. While the `SKILL.md` grants broad `Bash(infsh *)` permissions, this is scoped to the platform's CLI. The initial `curl | sh` command for CLI installation is also a generally risky practice. Although the documentation includes security best practices, the inherent power of these features without strict sandboxing or explicit user consent for each sensitive action warrants a 'suspicious' classification.