UTF-8 CSV validation and normalized rewrite

Validate one non-sensitive UTF-8 or UTF-8-BOM CSV and produce a normalized UTF-8 rewrite when safe.

Verified boundaries

One file, maximum 10 MB and 25,000 rows. No sensitive data, fuzzy identity matching, ambiguous merging, or third-party acceptance guarantee. Source evidence is preserved and exceptions require review.

What to check

Strict validation accepts supported UTF-8 input. Unsupported encodings fail closed instead of being guessed or silently repaired.

Safe next step

Keep the original file and verify row and column counts before using the normalized output. Unsupported encoding requires separate review.

Capability lineage: CAP-CSV-NORMALIZATION-v1

See the exact deliverable · Request private review

Privacy · Terms