ml-strategy
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements numerical machine learning logic for time-series analysis using established libraries including scikit-learn, pandas, and numpy. All operations are local and algorithmic in nature.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external OHLCV data via the
data_mapparameter inSignalEngine.generate(SKILL.md). While explicit boundary markers for natural language instructions are absent, the skill treats the input strictly as numerical data for statistical computation. The capability inventory across the script is limited to mathematical calculations and machine learning model training; it contains no subprocess execution, network operations, or file system writes. Input is sanitized through thevalidate_datafunction and specific handling ofinfandNaNvalues during feature engineering.
Audit Metadata