classification-helper

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXPOSURE] (SAFE): The script reads local CSV data and writes model/plot files to user-specified paths. No network operations or hardcoded credentials were found.
  • [REMOTE_CODE_EXECUTION] (SAFE): No use of eval(), exec(), or subprocesses to run external code. Dependencies are standard, well-known data science libraries.
  • [DYNAMIC_EXECUTION] (SAFE): The script uses pickle.dump to serialize trained models. While pickle.load can be dangerous with untrusted data, saving models with pickle.dump is standard practice and poses no risk in this context.
  • [PROMPT_INJECTION] (SAFE): No natural language prompts or LLM instructions are present in the skill files.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:02 PM