ml4t-exposure-analysis

Pass

Audited by Gen Agent Trust Hub on Sep 13, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references and imports the ml4t-diagnostic library. Per the author context, this is a vendor-owned resource associated with 'ml4t' and is utilized for professional portfolio evaluation tasks like factor analysis and rolling attribution.
  • [COMMAND_EXECUTION]: The skill configuration includes a path glob **/*kill*.py. While 'kill' is a sensitive term in system administration, in the context of financial portfolio management (alongside patterns like portfolio, risk, and stress), it refers to financial 'kill switches' or emergency risk liquidation scripts and does not indicate malicious intent.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing financial data (returns, factor data, and weights). However, the capabilities exposed are restricted to mathematical calculations using NumPy and Scikit-Learn. There are no functions that execute shell commands or perform network operations based on the input data, mitigating injection risks.
  • [DATA_EXFILTRATION]: No network operations or sensitive file access patterns were detected. The code snippets perform local statistical transformations on provided dataframes and arrays.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 13, 2026, 12:25 PM