playwright-human-in-the-loop
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust security model that distinguishes between low-risk and high-risk browser operations, ensuring that actions involving resource deletion, permission changes, or financial costs require explicit human approval via the AskUserQuestion tool.
- [SAFE]: Specific safety guidelines are embedded to prevent the agent from automatically entering secrets or clicking destructive buttons, which effectively mitigates common automation risks.
- [SAFE]: The instructions utilize verification steps, such as browser snapshots, to ensure the agent remains aware of the page state before and after actions, reducing the likelihood of errors in complex management interfaces.
- [SAFE]: No malicious patterns, obfuscation, or unauthorized external communications were detected; the skill functions entirely through defined MCP tool interfaces with localized human-in-the-loop controls.
Audit Metadata