stock-data-verifier

Preventing hallucinated stock data requires enforcing a cross-verification protocol where agents must extract original_text quotes, source URLs, and dates from multiple web searches, matching financial metrics within a ±5% tolerance to prevent fabricated figures.

Cross-verify ETF and stock prices by instructing agents to query at least two independent sources plus one primary official source, comparing extracted numeric data within a ±5% tolerance band and returning a standardized FAIL code if variance exceeds the limit.

Yes, you can verify Korean equity data using Naver Finance and KRX as independent sources. The verification process requires extracting original_text quotes and source URLs to cross-check KRW prices, PER, and dividend yields within the ±5% tolerance.

The best way to validate NAV and expense ratios is enforcing a three-source cross-check protocol during data collection, requiring at least two independent sources plus one primary official source, and demanding exact original_text quotes for traceability.

Yes, financial data verification integrates with automated workflows by outputting verified JSON containing matched metrics, original_text, and source URLs, or returning a standardized FAIL code when cross-source variance exceeds the ±5% tolerance band.

Stock data verification returns a FAIL code when the variance between extracted financial metrics from multiple web sources exceeds the strict ±5% tolerance band, indicating incompatible data points that cannot be confidently confirmed as accurate.