The Agent Skills Directory

[SAFE]: The skill operates within a controlled orchestration framework, delegating tasks to specialized agents through structured handoffs. This prevents direct execution of arbitrary user input.
[SAFE]: A mandatory 'Verify' stage is implemented, which includes security scanning by the 'Sentinel' agent and code review by the 'Judge' agent, effectively mitigating risks associated with generated content.
[SAFE]: Ingestion of untrusted user data (the project brief) is handled through structured YAML bundles, and the skill includes explicit sanitization steps via verification agents before any launch activity occurs.

bazaar