mascot-logo-reconstruction
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill orchestrates a complex modeling pipeline that relies on external data sources such as manifest files, registration reports, and texture packs. This ingestion of untrusted external content represents a surface for indirect prompt injection if the source metadata contains embedded malicious instructions.\n
- Ingestion points: Reads and processes configuration data from
reference_manifest.json,registration_report.json, andatlas_regions.json.\n - Boundary markers: The instructions lack explicit delimiters or specific directives for the agent to ignore instructions that might be embedded within the processed data files.\n
- Capability inventory: The skill is authorized to use
Bashfor shell operations andmcp__blender__execute_blender_codefor dynamic Python execution within the Blender environment.\n - Sanitization: The skill focuses on geometric and structural validation (part counts, silhouette overlays) but does not prescribe sanitization or validation of the text content within the input manifests before processing.
Audit Metadata