metforge-analysis
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is focused on reproducible atmospheric science calculations using industry-standard libraries such as MetPy, xarray, and SciPy.
- [DATA_EXPOSURE]: The skill contains an explicit security safeguard instructing the agent to 'Never record credentials or secrets' when appending records to logs or provenance files.
- [COMMAND_EXECUTION]: The skill permits the use of Slurm (
sbatch) for distributed computing tasks, which is an expected and standard practice for the scientific workflows described. - [PROMPT_INJECTION]: The instructions are clearly defined for scientific tasks and do not contain any patterns attempting to override agent safety or system instructions.
Audit Metadata