The Agent Skills Directory

[PROMPT_INJECTION]: No attempts to override agent behavior or bypass safety filters were detected in the instructions or metadata. The language is technical and focused on shader implementation.
[DATA_EXFILTRATION]: No sensitive file access or unauthorized network operations were found. The skill uses standard Three.js loaders to fetch textures and 3D models from provided URLs, which is standard practice for graphics skills.
[REMOTE_CODE_EXECUTION]: No remote script execution patterns or suspicious package installations were detected. The skill uses legitimate Three.js ecosystem libraries.
[OBFUSCATION]: No hidden or encoded malicious content was found. The binary assets (.glb) are standard 3D files containing geometry data and metadata, and no hidden URLs or acrostic patterns were detected.
[INDIRECT_PROMPT_INJECTION]: The skill has ingestion points for external assets via textureBaseUrl in hybrid-soil-moss-surface.js and GLTFLoader in model-moss-accumulation.js. However, it lacks exploitable capabilities like shell access or file writing, making the attack surface benign.
[DYNAMIC_EXECUTION]: The skill uses Three Shading Language (TSL) and onBeforeCompile to generate shader code at runtime. This is the intended purpose of the skill and does not incorporate untrusted user-supplied strings into the generation logic.
[COMMAND_EXECUTION]: No unauthorized shell commands or subprocess calls were found. The implementation is restricted to 3D rendering and physics simulation logic.

threejs-procedural-materials