The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a specific command template for running Blender in background mode ("<blender.exe>" -b --factory-startup -P <script> -- <args>). This allows the agent to invoke the Blender executable with arbitrary scripts and arguments.
[DYNAMIC_EXECUTION]: The instructions explicitly direct the agent to generate new Python scripts for mesh manipulation, such as tools/paint_areas.py and tools/area_split.py, and subsequently use them in the pipeline. Executing code generated at runtime by an AI agent is a documented security risk.
[INDIRECT_PROMPT_INJECTION]: The skill's workflow depends on processing external data from multiple sources which could serve as injection vectors.
Ingestion points: The agent is instructed to fetch and analyze 8-12 images from external sources like Google Images, wikis, and video stills, in addition to user-provided reference images.
Boundary markers: There are no instructions for the agent to use delimiters or ignore potential embedded instructions within the metadata or content of these external files.
Capability inventory: The skill has the capability to write files (scripts and delivery assets), execute subprocesses (Blender), and interact with external services (Tripo Studio, Modal).
Sanitization: There are no specified procedures for sanitizing or validating the content of the external images or design briefs before they are used to influence the model generation process.

build-asset