Reviewed-by for Tau Ceti

1. Find the declaration, open it and press Review this: a GitHub form opens with its name and the commit shown filled in. Submit it to say it is the intended mathematical notion, and, if you like, what you compared it with and which ways a definition goes wrong you checked. Saying why is optional for people; AI agents must.
2. A bot records the review in the evidence store of this repository, answers on the issue and closes it. There is no pull request, and this page updates within a few minutes.
3. A review is keyed by the declaration's meaning hash, which changes when the declaration, or anything it rests on, changes meaning, and not when it is renamed, reformatted or re-proved. When it changes, the mark stays but is greyed, as earlier version if the declaration itself was rewritten, or as something it rests on changed since if it reads the same but a definition it uses moved. A renamed declaration keeps its marks. Comment /withdraw on your review's issue to take it back.
4. Each declaration counts its reviews, people apart from AI agents; Who lists them, with dates, versions and evidence.
5. Named shows only the named results and notable definitions: the ones the roadmaps' status files and Voyager's announcements single out, rather than the API and proof steps around them.

Tests are what a declaration passes, rather than a mark: key results listed as its tests with List a test, and tests anyone proposes with Suggest a test: a property it should have, which stays open until someone proves it in Tau Ceti and comments /met <the declaration that proves it> on its issue (or /failed, if it turns out false: then report the problem). A test passes while it is in Tau Ceti at the pinned commit without sorry. Which shows each with its statement.

If a declaration is wrong, press Report a problem instead and say why, and how to fix it if you know. The report is the issue for fixing it: it stays open, and the page flags the declaration, until its reporter or a maintainer closes it as completed (fixed) or as not planned (not a problem), or comments /fixed <commit>, /intended or /invalid.

Marking many at once: comment lines like Reviewed-by: TauCeti.X.y — what you checked, Test: TauCeti.X.y — TauCeti.X.y_zero — what it checks or Named: TauCeti.X.y — its name — a sentence on issue #1. AI agents use the same routes and name the agent, model and session; their marks are shown apart from people's.

Reviewed-by

it is the intended mathematical notion