The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructs the agent to use Python scripts for automating modeling operations, batching construction tasks, and managing render processes.
Evidence: SKILL.md and references/blender-operations.md explicitly define a workflow that utilizes "deterministic Blender Python for repeatable construction or render batches" and "scoped, reproducible Python for repeated construction."
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted user data, including product briefs and reference specifications, which are used to influence the modeling logic and generated scripts.
Ingestion points: templates/product-brief.md and user-provided reference images/descriptions.
Boundary markers: The templates/accepted-constraints.md ledger is used to differentiate between "AUTHORITATIVE USER FACTS" and "PHOTO-INFERRED" data.
Capability inventory: The skill can generate and execute Python code, perform file operations (saving .blend, .glb, and .png files), and configure system-level execution via Blender CLI arguments.
Sanitization: The instructions do not specify automated validation or sanitization of input data before its use in modeling parameters or script generation.
[COMMAND_EXECUTION]: The workflow involves interacting with the host system to locate the Blender executable and configuring command-line arguments for background processing.
Evidence: references/blender-operations.md provides guidance on arranging CLI arguments so intended files load before scripts run and supplying discovered executable paths to the environment.

guitar-gear-modeling