The Agent Skills Directory

[SAFE]: The skill provides legitimate NLE automation workflows and includes proactive instructions to avoid license bypasses and ensure system stability. All external links point to official MAGIX documentation.
[INDIRECT_PROMPT_INJECTION]: The skill defines workflows that ingest untrusted data from project metadata and UI screenshots, which are standard for NLE automation.
Ingestion points: The BatchRenderRegions.cs script in SKILL.md reads region.Label from the timeline; gemini.md directs the agent to process UI screenshots and error dialogs.
Boundary markers: Not explicitly defined for the project data ingestion.
Capability inventory: The skill generates scripts capable of file system directory creation and file writing through render commands.
Sanitization: The provided C# code implements sanitization by iterating through Path.GetInvalidFileNameChars() and replacing them with underscores, effectively mitigating path traversal and illegal filename risks.

vegas-pro