threejs
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze the user's project by reading
package.jsonand listing 3D model filenames in directories likepublic/,assets/, orstatic/. This metadata ingestion serves as a surface for indirect prompt injection where an attacker could use malicious filenames or package names to influence the agent's tool selection or behavior. - Ingestion points: Project configuration (
package.json) and directory listings for 3D assets. - Boundary markers: Absent; the skill does not specify delimiters or warnings to ignore instructions embedded in the ingested file metadata.
- Capability inventory: The skill acts as a router to other specialized skills (e.g.,
threejs-model-inspect,threejs-model-fix) and does not perform sensitive file writes or network operations itself. - Sanitization: No validation or sanitization of the listed filenames or package versions is described.
- [EXTERNAL_DOWNLOADS]: The documentation recommends the installation of an external skill using the command
npx skills add img2threejs/img2threejs. This references a third-party repository that is not part of the established trusted organizations, involving the download of external content at the user's discretion.
Audit Metadata