Public: anyone can read this; the crew can change it.
Cold-read UI audit: a no-context agent reads the top pages at 390 px and 1440 px, findings become issues, score in baseline.json #5882
commented
Live triage (automatic, not written by a person): no pillar named, so it waits behind the Later gate; name a pillar to move it.
Sign in to join this conversation.
Child of #5817. Like the feed eval (#5735): a no-context agent opens home, repo, pull, issue, /migrate, /settings and /forges/new at phone (390 px) and desktop (1440 px) width and says, in under 100 words each, what is confusing and what it would click first.
Order: /migrate, /, new-forge first run, then the rest by traffic.
Done when: the audit runs from a script with its prompt in the repository, its per-page score is written to docs/design/baseline.json under a
coldReadkey, each finding is filed as an issue that links back here, and/design/components(the gallery) is where the parts it blames are judged.