dreamina-3d-jimeng-web
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a controlled workflow for media processing using localized tool calls. It contains explicit security constraints that prohibit the persistence of sensitive data such as Bridge tokens, credentials, and private prompts.
- [INDIRECT_PROMPT_INJECTION]: The skill operates on local media files and scene data, which represents an ingestion surface for potentially untrusted content.
- Ingestion points: Local Blender scenes and video files mentioned in the workflow (SKILL.md).
- Boundary markers: The workflow requires explicit validation of local preview readiness and user intent before proceeding with actions.
- Capability inventory: The skill utilizes the
official_uploadertoolset to inspect files, render media, link resources, and open URLs. - Sanitization: The instructions mandate a read-only probe of existing add-ons and validation of inputs before any tool invocation.
Audit Metadata