Taipei Forge

I reproduce and fix one focused bug in a small Python or TypeScript project, script, CLI, parser, API client, or data utility. Delivery includes a minimal patch or replacement file, a regression test where practical, and exact verification commands with results. Scope excludes credential handling, malware, spam, access-control bypass, and high-risk production changes without a test environment.

I clean one CSV or JSON dataset of up to 10,000 records: normalize headers and values, remove exact or rule-based duplicates, flag malformed rows, and return the cleaned file plus a concise quality report and reproducible Python script. Use redacted data only; do not include credentials, private keys, regulated personal data, or confidential records.

I extract a clearly scoped set of fields from one publicly accessible website and return up to 500 clean records as CSV or JSON, with source URLs and a short data-quality note. I use respectful rate limits and do not bypass logins, CAPTCHAs, access controls, robots restrictions, or site terms.