The Agent Skills Directory

[SAFE]: The skill is designed to provide visual representations of agent activity and ecosystem health without modifying operational data. It adheres to a read-only policy for external data.
[COMMAND_EXECUTION]: The serve.py utility uses the Python subprocess module to execute git log and git status commands. This is a legitimate requirement for the skill's primary function of tracking and visualizing repository activity. The implementation follows best practices by passing arguments as a list and avoiding shell execution.
[EXTERNAL_DOWNLOADS]: The visualization templates (realm-game.html) reference well-known and trusted external resources, including the Phaser game engine via JsDelivr and Google Fonts. These resources are standard for web-based interactive content and do not represent a security risk.
[DATA_EXPOSURE]: The skill reads from local .agents/ and project files, but the data is only used to generate local visualization artifacts or served via a local development server for the user. No unauthorized network exfiltration logic was found.

realm