The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines a sequence of commands for the agent-browser tool to interact with web pages, such as open, wait, and snapshot. These are standard operations for web automation and debugging.
[REMOTE_CODE_EXECUTION]: The skill utilizes the eval command to execute JavaScript snippets within the browser context (e.g., checking HTML5 form validity). This is a standard debugging technique restricted to the target web page's environment and intended for diagnostic inspection.
[DATA_EXFILTRATION]: While the protocol includes steps to check cookies and localStorage for authentication tokens, security is addressed through explicit "Judgment Gates" that instruct the agent to pause and ask the user for permission before logging sensitive information found in network payloads. No unauthorized exfiltration patterns were detected.

agent-browser-debug