track-equity-cumulative-return

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes historical price data from an external source (Yahoo Finance), which represents a potential surface for indirect prompt injection if the external data source were compromised.
  • Ingestion points: scripts/fetch_price_data.py fetches market data using the yfinance library.
  • Boundary markers: No explicit delimiters or instructions are used to distinguish external API data from system instructions within the processing pipeline.
  • Capability inventory: The skill has the ability to write JSON results (scripts/cumulative_return_analyzer.py), generate and save PNG visualization charts (scripts/visualize_cumulative.py), and maintain a local Parquet cache (scripts/fetch_price_data.py).
  • Sanitization: Tickers, years, and indices are strictly validated in scripts/validators.py. The price data itself is checked for basic quality (NaN percentage, non-positive values) in scripts/fetch_price_data.py before calculation.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 10:33 PM