threejs-loaders

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The SKILL.md shows loaders fetching arbitrary external assets (e.g., loader.load("model.glb"), texture/load("texture.jpg"), fetch("model.glb"), plus CDN decoder paths like https://www.gstatic.com/... and manager.setURLModifier) and explicitly reads GLTF contents (gltf.userData, cameras, animations) and uses them to drive scene setup and actions, so untrusted third-party files could materially influence runtime behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill sets decoder/transcoder paths that fetch and execute remote decoder code at runtime (required for decoding compressed assets), e.g. https://www.gstatic.com/draco/versioned/decoders/1.5.6/ and https://cdn.jsdelivr.net/npm/three@0.183.0/examples/jsm/libs/basis/, so these external URLs are runtime-executed dependencies.