The Agent Skills Directory

[SAFE]: The skill uses the mcp__ue58-mcp__execute_python_code tool to automate Unreal Engine 5.8 tasks. The Python logic provided for creating assets (ImgMediaSource, MediaPlayer, Material) and spawning actors (CineCameraActor) follows standard Unreal Engine developer patterns.
[SAFE]: No external network requests, hardcoded credentials, or data exfiltration attempts were found. The skill operates locally within the Unreal Engine environment and the user's specified media directories (e.g., D:/Plates).
[SAFE]: No obfuscation techniques, hidden commands, or attempts to bypass safety filters were detected in the instructions or reference documentation.
[INDIRECT_PROMPT_INJECTION]: The skill includes workflows for processing external data, such as CSV shot lists and image sequence directories, which represents a standard ingestion surface.
Ingestion points: reference/multi_shot_production.md (CSV file ingestion) and reference/foreground_plate_workflow.md (filesystem directory scanning for image sequences).
Boundary markers: No explicit LLM boundary markers are used in the data ingestion scripts, though the skill relies on standard Unreal Engine asset validation.
Capability inventory: The skill possesses file system access through standard Python libraries (os, csv) and full Unreal Editor control via the unreal module.
Sanitization: Basic path validation and regex-based wildcard generation are utilized to ensure sequence paths are formatted correctly before processing.

unreal-vfx-automation