The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project-local files such as game concepts and existing design docs, which represents a surface for indirect prompt injection. This is the intended behavior for analysis but lacks explicit input sanitization.
Ingestion points: design/gdd/game-concept.md and design/gdd/*.md.
Boundary markers: None defined to isolate external data.
Capability inventory: Write, Edit, and Task tools.
Sanitization: None mentioned.
[SAFE]: The skill functions as a workflow template for game development, coordinating between design documents and sub-agents using local tools and predefined gates. It does not attempt network exfiltration or unauthorized system access.

map-systems