bayesian-bandits

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security concerns identified. The skill implements standard statistical algorithms for multi-armed bandits using trusted scientific libraries.
  • [EXTERNAL_DOWNLOADS]: The skill uses well-known Python packages (marimo, numpy, pandas, matplotlib, scipy) from standard registries. These are documented neutrally as required.
  • [DATA_EXFILTRATION]: No exfiltration patterns detected. The skill mentions data ingestion from local files (e.g., 'data/events.parquet') for its primary purpose of statistical analysis.
  • [PROMPT_INJECTION]: No prompt injection or behavior override patterns found in the instructions.
  • [COMMAND_EXECUTION]: No dangerous command execution or privilege escalation patterns detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 06:43 AM
Security Audit — agent-trust-hub — bayesian-bandits