Contributing

The dataset is the public artifact. The fastest way to make it better is to open a pull request that adds or corrects a row, and links the primary source.

What to send

1. A single new row in data/regulations.json (CSV mirror regenerates from JSON).
2. The primary source URL in source_url.
3. A short plain-language summary in your own words (no copy-paste from a law-firm alert).
4. A last_verified date — the day you actually opened the primary source.

How to send it

Open a PR against main on GitHub. One row per PR is fine; a small batch of related rows is also fine. The maintainer will run the data-check script and review the primary source before merging.

What I won't accept

• Rows whose source_url is a secondary source (law-firm blog, news article, a general AI tracker like IAPP / multistate.ai).
• Summary text copy-pasted from a source — the summary is the dataset's editorial contribution and must be written from scratch.
• Anything that smells like legal advice. Aggregation, not advice.

See METHODOLOGY for the full data-quality rules and the status taxonomy.