skills/simota/agent-skills/clay/Gen Agent Trust Hub

clay

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses subprocess.run to execute various local CLI tools for 3D asset processing, including Blender, instant-meshes, toktx, cwebp, and npx. These are utilized in the references/game-pipeline.md and references/texture-baking.md files to automate retopology, texture compression, and mesh optimization.- [EXTERNAL_DOWNLOADS]: The skill fetches 3D models and assets from numerous external repositories and marketplaces, such as Sketchfab, Poly Pizza, Kenney, and the Smithsonian Institution (as detailed in references/api-integration.md). It also calls several external 3D generation APIs, including Meshy, Tripo, and Rodin.- [REMOTE_CODE_EXECUTION]: The skill invokes npx to run the @gltf-transform/cli utility, which involves downloading and executing code from the npm registry at runtime.- [DATA_EXFILTRATION]: The skill is designed to read API keys from environment variables and transmit them to external 3D service providers via HTTPS requests using the httpx library. This is a legitimate requirement for the service integrations but represents a data transfer capability.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it is designed to ingest and process untrusted external data, such as 3D model files (FBX, OBJ, glTF) and images, which are then used as inputs for processing scripts and API calls.
  • Ingestion points: Processes externally sourced 3D model files and user-provided images (references/api-integration.md, references/quality-validation.md).
  • Boundary markers: No specific delimiters or "ignore instructions" warnings are applied to the data processing logic.
  • Capability inventory: The skill possesses the capability to execute shell commands via subprocess.run and perform network requests via httpx (references/game-pipeline.md).
  • Sanitization: No explicit sanitization or validation of the internal content of processed 3D files is demonstrated in the provided code snippets.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 01:01 AM
Security Audit — agent-trust-hub — clay