The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local CLI utilities, specifically codex and pstack scripts, to spawn and manage worker sub-agents. These commands are integral to the workflow's orchestration logic for parallel design exploration.
[PROMPT_INJECTION]: No malicious injection patterns or safety bypass attempts were found. The skill defines specific personas and constraints to guide the agent's architectural reasoning.
[SAFE]: The skill does not perform external network requests, access sensitive user credentials, or employ obfuscation techniques. All referenced tools and scripts are treated as local environment components.

architect