CSV Risk Scan (Local)
Scan a CSV for patterns that Excel often rewrites on open.
Safety note: This file contains CSV/formula-injection examples (cells that start with
= / @). Treat it as untrusted.
Run a local risk scan
Select a CSV file. The analysis runs in your browser — nothing is uploaded.
Sample kit (download)
What it demonstrates
- Leading zeros removed: IDs like
00123become123. - Precision loss: long IDs (16+ digits) are converted to scientific notation and altered.
- Locale/date guessing:
03/04/2026changes meaning across machines. - Formula injection: values starting with
=/@are interpreted as formulas.
How to run the test
- Download the CSV.
- Open it directly in Excel (double-click).
- Observe the damage.
- Now compare: run it through SterileCSV and open the generated
.xlsx.
SterileCSV output (generated)
We ran this exact CSV through SterileCSV with default safe settings (formula-like values stored as text; ambiguous dates kept as text).
Download output XLSX Download summary JSON
- Leading zeros preserved: IDs stay
00123(not123). - Long IDs preserved: long identifiers remain exact (not scientific notation).
- Formula injection neutralized: cells that begin with
=/@are stored as text (not executed).