dreamina-3d-from-blender
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes externally provided scene data and artifact receipts, which introduces a potential surface for indirect prompt injection.\n
- Ingestion points: User-provided scene files are processed via
inspect_sceneinSKILL.md, and metadata receipts are ingested by thevalidate_artifactfunction inscripts/handoff_validator.py.\n - Boundary markers: The workflow specifies clear state transitions and includes a mandatory validation step to ensure artifacts match expected hashes before proceeding.\n
- Capability inventory: The skill possesses capabilities to execute scene inspection tools and perform network-based video submissions through the
McpDesignClient.\n - Sanitization: Structural validation and cryptographic integrity checks (SHA256) are performed by
scripts/handoff_validator.pyto verify that the artifact has not been tampered with.
Audit Metadata