The Agent Skills Directory

[SAFE]: The skill enforces a rigorous "Question -> Options -> Decision -> Draft -> Approval -> Write" cycle for every document section. By using the AskUserQuestion tool at every decision point, it prevents the agent from taking autonomous actions without human oversight.
[SAFE]: File system interactions are limited to local project paths (e.g., design/, docs/, production/). The skill also includes a robust validation phase that checks for data consistency against a local registry (entities.yaml) and surfaces any conflicts to the user rather than overwriting data silently.
[SAFE]: Multi-agent orchestration is used safely to provide specialized domain expertise. All outputs from delegated agents (like creative-director or systems-designer) are presented to the user for evaluation before being incorporated into the final document.
[SAFE]: No remote code execution, external network requests, obfuscation, or sensitive data harvesting patterns were detected. The skill is entirely focused on the stated purpose of assisting with game design documentation within a local environment.

design-system