The Agent Skills Directory

[COMMAND_EXECUTION]: The skill manages a local Node.js server (server.cjs) and utility shell scripts (start-server.sh, stop-server.sh) to provide a visual interface for brainstorming. These executions are scoped to the skill's specific directories and serve the tool's core documented purpose.
[COMMAND_EXECUTION]: The stop-server.sh script utilizes rm -rf for cleanup of ephemeral session data. It includes a robust safety check to ensure it only deletes directories within the intended 'brainstorm' path segment, preventing accidental file deletion.
[EXTERNAL_DOWNLOADS]: No external code or dependencies are downloaded from the internet. The custom Node.js server implementation relies entirely on built-in modules such as http, fs, path, and crypto.
[DATA_EXFILTRATION]: The server binds to the local loopback interface (127.0.0.1) by default. Communication is strictly limited to the local machine, facilitating interaction between the user's browser and the agent without sending data to external domains.
[PROMPT_INJECTION]: Instructions emphasize obtaining user consent before initiating browser-based sessions. There are no patterns suggesting attempts to bypass agent safety filters or override system behavior.

visual-brainstorming