cobrapy
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data via metabolic model files in SBML, JSON, and YAML formats.
- Ingestion points: Methods such as
read_sbml_model,load_json_model, andload_yaml_modelare documented inSKILL.mdfor loading models from local file paths. - Boundary markers: Ingestion is restricted to structured scientific data formats which provide inherent data boundaries.
- Capability inventory: Capabilities include flux balance analysis, gene deletion studies, and production strain design.
- Sanitization: The COBRApy library includes internal validation logic to ensure metabolic models are mathematically and structurally consistent.
- [SAFE]: No malicious patterns such as prompt injection, obfuscated commands, or unauthorized network operations were detected. The workflows utilize standard Python scientific libraries for data analysis and visualization.
Audit Metadata