Regex bug undercounted resume data 8x until a formatting quirk was caught

A developer maintaining a corpus of 456 role-specific resume examples in TypeScript built a regex script to extract job description fields, but the pattern initially returned only 199 results instead of the expected 1,723. The discrepancy was traced to Prettier's auto-formatting, which sometimes placed description values on a new line, causing the regex — which expected a space after the colon — to silently skip those records. Adding \s* to the pattern to allow optional whitespace fixed the match count. The incident highlights a broader risk: parsers that silently skip malformed input produce wrong answers rather than errors, making bugs harder to detect. The author recommends always having an expected order-of-magnitude figure before reviewing output, and cross-checking multiple extracted counts from the same data source to catch such silent failures early.