occupancy-and-detection
Pass
Audited by Gen Agent Trust Hub on Jun 26, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill lists standard ecological and data science dependencies such as the R package "unmarked" and the Python library "pandas". These are widely used, reputable tools for scientific computing.
- [DATA_EXPOSURE]: The skill operates on user-provided data files (CSV format). It does not attempt to access sensitive system directories, credentials, or environment variables. All outputs are written to local directories specified by the user.
- [COMMAND_EXECUTION]: The R and Python scripts perform mathematical computations and data transformations. They use standard library functions for file I/O and do not execute arbitrary shell commands or external binaries.
- [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for external data.
- Ingestion points: "detection_history.csv" and "site_covariates.csv" processed by "scripts/occupancy_analysis.R" and "scripts/occupancy_analysis.py".
- Boundary markers: Not applicable, as the data is used for numerical modeling rather than prompt construction.
- Capability inventory: Local file read/write access. No network access or privilege escalation capabilities.
- Sanitization: Both scripts implement input validation (e.g., "validate_detection_history" in Python) to ensure the data format matches the requirements of occupancy models.
Audit Metadata