alterlab-timesfm

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads model weights (~800 MB) from Hugging Face's official repository (google/timesfm-*) upon first use. This is a standard and expected behavior for using large foundation models and involves a well-known, trusted service.
  • [COMMAND_EXECUTION]: The scripts/check_system.py script executes benign system commands such as sysctl and vm_stat specifically to determine available RAM and hardware resources on macOS. This is used for safety to prevent system crashes during model loading.
  • [DYNAMIC_EXECUTION]: The skill uses __import__() within its diagnostic script to programmatically verify whether the required timesfm and torch libraries are installed. The module names are hardcoded, preventing arbitrary code loading.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied CSV files for forecasting. While this represents a data ingestion surface, the processing is limited to numerical analysis and forecasting without evidence of untrusted data influencing sensitive system operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 02:36 PM