The Agent Skills Directory

[SAFE]: The skill contains legitimate instructions and code snippets for using the Three.js library to handle animations, skeletal rigs, and morph targets. No malicious patterns, such as prompt injection, persistence mechanisms, or unauthorized privilege escalation, were detected in the instructions or scripts.
[INDIRECT_PROMPT_INJECTION]: The skill includes examples for loading external 3D model data, which represents a potential injection surface for untrusted content.
Ingestion points: The GLTFLoader.load method in SKILL.md ingests external .glb files.
Boundary markers: The instructions do not define specific boundaries or validation logic for the content of the loaded models.
Capability inventory: The code snippets demonstrate capabilities limited to the Three.js rendering context, such as updating object properties and animation playback. There are no capabilities for file system modification, shell execution, or network data exfiltration.
Sanitization: External data (such as animation clip names or morph target names) is used directly within the rendering logic without explicit sanitization.

threejs-animation