The Agent Skills Directory

[SAFE]: The skill incorporates extensive human-in-the-loop checkpoints using the AskUserQuestion tool at every phase, ensuring the user remains in control of the creative direction and data written to the file system.
[SAFE]: Multi-agent coordination via the Task tool (spawning technical, art, and creative directors) is used for validation and feedback, following a robust architectural pattern for complex agentic workflows.
[INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for potential indirect prompt injection as it reads existing design documents (design/gdd/game-concept.md, design/gdd/game-pillars.md) to resume brainstorming sessions.
Ingestion points: design/gdd/game-concept.md, design/gdd/game-pillars.md, and production/review-mode.txt.
Boundary markers: No explicit delimiters are specified for content read from these files.
Capability inventory: The skill possesses Write, WebSearch, and Task capabilities.
Sanitization: There is no evidence of sanitization for the content read from the local files, although the collaborative nature of the skill requires the user to interactively validate the synthesized outputs.

brainstorm