AUTHOR’S WORD — 5 rows here show a completed status that nothing derived The pill on these rows was typed, not measured: prod-deploy-rate, preview-provenance, canceled-billing, window-premise, audit-answer. Every other status on this page comes from that row’s own PR, commit or decision links; these name nothing a machine can read, so nothing can contradict them when the work changes. To make the same claim checkable, write one of these on the row and it derives on the next tick: prs (a PR ref — read from GitHub), completion_evidence (owner/repo@<full 40-char sha> — checked against the default branch), shipped_without_pr (why no PR can exist), or decision (the id that settled it). A row that is not owed a build says so with unbuilt; one a decision retired says so with canceled.

Vercel build audit — do we need a build every time?

Can u do an audit, do we need to do a build every time? We only deploy once every 30 mins. On release. Not on push or merge

Status is DERIVED from each row’s own PR, decision and prod-evidence links by bin/refresh-tracker vercel-build-audit — never typed. Last refreshed never.

Overall

—

Shipped

—

Merged

—

In flight

—

Open

—

Waiting on you

Groups — click to filter

Every row opens: when it was added, why it exists, and the links that prove it. Rows waiting on a human carry a Decide → link straight to the answerable question on the decisions page. Statuses are re-derived from live PR, decision and prod-evidence state by bin/refresh-trackers; nothing here is typed by hand, and merged is not shipped until a prod_evidence permalink says the fix was exercised in production.