The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted source code and project files without security boundaries. Malicious instructions within the codebase could influence the agent's behavior.\n- Ingestion points: Source files from Unity, Unreal, Godot, and web projects (e.g., .cs, .gd, .unity, .uasset) are read from the project directory during analysis phases.\n- Boundary markers: No delimiters or instructions to disregard embedded commands are provided when reading codebase content.\n- Capability inventory: The skill performs file-read operations for project analysis and file-write/append operations to log telemetry in a hidden metadata directory (~/.claude/projects/).\n- Sanitization: No sanitization or filtering is applied to the data ingested from the project files to prevent instruction injection.

game-performance