Messy CSV and Excel exports create manual cleanup work, inconsistent records, and uncertainty about what changed. I turn one structured file into a clean, reviewable output using explicit rules agreed before processing. You receive a cleaned CSV or Excel file in the agreed format, a separate exception file with a reason for each failed rule, a change log, and a validation/reconciliation report with input, accepted, and exception counts. Base scope: one CSV/XLSX table up to 5,000 rows and 20 columns, with up to five explicit cleanup or validation rules. Turnaround is two business days after the complete file, rules, and funded project are confirmed. One correction is included if delivery does not follow the agreed rules; changed requirements are new scope. Process: you provide the file, output format, column definitions, and specific rules with examples; we confirm acceptance checks; I apply the rules and isolate exceptions instead of guessing; then I validate row counts and outputs. Excludes semantic research, enrichment, fuzzy matching, OCR or PDF extraction, scraping, restricted-source extraction, database migration, API integration, large manual fact checking, and accounting, medical, or legal judgment. Disclose protected or high-risk data before sharing any file. The supporting CSV sample is an Independent Demonstration Project using synthetic data, not client work.