The Agent Skills Directory

[SAFE]: No malicious prompt injection detected. The instructions include a security warning for the agent to treat external web content as untrusted and to ignore any embedded instructions (e.g., 'ignore previous instructions'), which is a defensive best practice documenting potential external risks rather than introducing a threat.\n- [SAFE]: The skill utilizes the legitimate 'playwright' library for its functionality. The instructions provided for installation and error handling are standard for this dependency.\n- [SAFE]: No evidence of unauthorized data exfiltration, hardcoded credentials, persistence mechanisms, or obfuscated content was found.

browser