The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell scripts (start-server.sh, stop-server.sh) to manage a local Node.js server (server.cjs). This server is used to host a visual companion that displays mockups and diagrams during the brainstorming process.
[SAFE]: The server implementation follows security best practices for local development tools. It binds to the loopback interface (127.0.0.1) by default, utilizes path.basename to prevent path traversal when serving files from the session directory, and includes an automatic shutdown mechanism that triggers if the parent agent process terminates or after 30 minutes of inactivity.
[PROMPT_INJECTION]: The instructions utilize strict operational constraints (e.g., <HARD-GATE> markers) to enforce a specific workflow. These are legitimate pedagogical instructions designed to ensure the agent completes design work before implementation, rather than attempts to bypass safety filters or ignore prior instructions.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user interaction data (clicks and selections) from a local state file generated by the browser companion. While this represents an ingestion of external data, the risk is minimal as the data is structured JSON generated by the skill's own trusted helper script, and the process occurs within a local execution environment controlled by the user.

brainstorming