The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided mechanic descriptions and reads internal project documentation, creating an ingestion surface for potentially untrusted data.
Ingestion points: Data enters the agent context through user-supplied arguments (e.g., "[mechanic or system to design]") and by reading project-level files in the @docs/ and @templates/ directories.
Boundary markers: The instructions provide detailed templates and persona-based constraints to guide the output, but do not include explicit delimiters or warnings to ignore instructions embedded within the ingested data.
Capability inventory: The skill is authorized to use file system tools (Read, Glob, Grep, Write, Edit) to manage design documentation, but lacks network or arbitrary command execution capabilities.
Sanitization: There are no specific instructions for sanitizing, escaping, or validating the user-provided content before processing.
[SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were identified. The skill's design and tool usage are consistent with its intended role as a documentation and systems design assistant.

game-designer