synthetic-data-generation
Pass
Audited by Gen Agent Trust Hub on Jul 3, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill specifies several Python dependencies required for data processing, including
lasio,welly,dlisio,segyio,numpy,pandas, andscipy. These are established, domain-specific libraries for handling geoscientific data formats and performing numerical analysis. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface as it accepts user-provided parameters (such as
well_name,formation, andfield_name) which are then used to populate metadata in generated synthetic files (LAS, SEG-Y, and OSDU manifests). - Ingestion points: User inputs for data generation parameters defined in
SKILL.md. - Boundary markers: None explicitly defined for user-supplied strings.
- Capability inventory: The skill performs file writing and metadata generation through domain-specific classes (
LASGenerator,SEGYGenerator,LASMapper). - Sanitization: There is no mention of explicit sanitization for user-provided strings before they are written to file metadata.
- [SAFE]: No malicious patterns, obfuscated code, unauthorized network operations, or persistence mechanisms were identified. The instructions for handling expensive AI image generation (core photos) prioritize user consent and cost awareness.
Audit Metadata