The Agent Skills Directory

[SAFE]: The skill performs its stated purpose of software architecture design without any hidden malicious behaviors. All outputs are documented and written to a dedicated local directory.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from project briefs (ingestion points: ideator-output/project-brief.md or user input) to generate architectural designs. While boundary markers and sanitization mechanisms are absent, the risk is minimal because the agent's capabilities (inventory: write, exec) are restricted to creating a static output directory and generating documentation.
[COMMAND_EXECUTION]: The skill executes a simple mkdir command to initialize its output folder. This command is statically defined and does not incorporate user-supplied variables, eliminating the risk of command injection.

constructor