threejs-materials-lighting
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No patterns detected that attempt to override agent instructions or bypass safety filters.
- [DATA_EXPOSURE]: No hardcoded credentials or access to sensitive local file paths (e.g., .ssh, .env) were found. Asset loading paths (e.g., 'assets/logo.png') are standard for development.
- [OBFUSCATION]: No obfuscated code, base64-encoded instructions, or hidden characters were detected.
- [REMOTE_CODE_EXECUTION]: No remote script downloads or piped shell executions. Standard ESM imports from 'three' and 'three/addons' are used appropriately for a Three.js environment.
- [DYNAMIC_EXECUTION]: No use of eval(), exec(), or runtime compilation techniques.
- [PRIVILEGE_ESCALATION]: No commands requesting administrative or root access (e.g., sudo) are present.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external asset files (textures, HDR maps) through Three.js loaders. This represents a standard development surface with no exploitable capability chains like file system writes or network exfiltration identified.
- [DYNAMIC_CONTEXT_INJECTION]: No usage of the dynamic shell execution syntax ('!command') was found in the skill metadata.
Audit Metadata