What this is
Every card is a real review or request someone actually filed, collected from published sources and rated for how far it sits from anything you would expect a person to complain about. One star, five stars, or no stars at all.
For machines
/api/cards.json: every card, structured/api/cards/<id>.md: one card as a post brief/api/index.md: the whole set in markdown/llms.txt: how to read all of it
Take the .md from any card, hand it to a model, and it has the quote, the rating, the source and the rules for writing the post.