The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes several local processes including Node.js scripts (runner.mjs, server.mjs), .NET build commands, and the Godot engine. These commands are used to manage the game's lifecycle and playtest automation within the project directory.
[DATA_EXFILTRATION]: All network interactions use curl targeting localhost (port 9901). There are no attempts to access external domains or exfiltrate sensitive data from the machine.
[SAFE]: The instructions are tailored for a specific development environment and use standard local tooling to achieve the stated goal of automated quality assurance, with no signs of malicious patterns or data exfiltration.

autoresearch