The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references and encourages integration with several third-party Model Context Protocol (MCP) servers hosted on GitHub repositories that are not associated with the established trusted vendor list.
Evidence: Documentation within the "Connected MCP Servers" section provides links to external repositories including github.com/ahujasid/blender-mcp, github.com/AIDC-AI/Pixelle-MCP, github.com/joenorton/comfyui-mcp-server, and github.com/raveenb/fal-mcp-server.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface area for indirect prompt injection as it is designed to ingest and review external data (assets and visual questions) while maintaining file-modification capabilities.
Ingestion points: Data enters the agent context through user-provided arguments such as visual-question and asset-review, as well as file content retrieved via the Read, Glob, and Grep tools.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard potential instructions embedded within the assets or metadata it is tasked to analyze.
Capability inventory: The skill is authorized to use Write and Edit tools across the filesystem (context: fork).
Sanitization: The instructions lack explicit sanitization or validation logic for the content of the external files it processes.

game-art-director