A red restore drill, or none in eight days, is one line on the board and a push to the owner's phone #1183

Part of #1147 (P0 Trust): a drill nobody hears about is not a drill.

Done looks like.

• [ ] The board shows one line when the last drill is red or older than eight days, naming the step that failed; it clears on the next green drill.
• [ ] The same cause sends one push notification, not one per tick.
• [ ] Tests: red drill → one line and one push; green drill clears it; eight days with none → line.
• [ ] The line is reachable through the MCP (health and asks name it).
• [ ] A look (screenshot) of the page that shows it, taken from the pull's preview, is in the pull.
• [ ] A test that goes red in this pull is fixed in this pull, whoever's it is.

Files. src/snapshot/drill.ts, src/crew/notify.ts, the board's warning lines, a test.

Risks. Duplicate pushes across a roll (two processes): record the notice in the same transaction as the check.

Fits. Mason.

Before pushing. bun x tsc --noEmit and the test files named here, quoted in the pull. One pull, green on its first push.

One pillar: this serves exactly the one it is linked under. Done means reachable on the page, through MCP, in a test and by eye (#334).

Moved off P0's umbrella (#586): this waits on the restore drill #1182, which is now deferred behind the Later gate (#595), so it waits there too.

Correction: this is back under P0 (#586) and now waits on #1568, which replaces #1182.

Parked by the crew: the preview's deploy failed and the decider saw no point in another try (pull #1718 closed). Planner: split this smaller, or rewrite what done looks like; a person can also say what to try.

Parked by the crew: the check failed ("the drill's verdict > a failing drill makes the card naming the step, and the next passing drill clears it") and the decider saw no point in another try (pull #1721 closed).

Parked twice since a person last spoke here, so it is the planner's now. Every reason:

1. the preview's deploy failed and the decider saw no point in another try (pull #1718 closed)
2. the check failed ("the drill's verdict > a failing drill makes the card naming the step, and the next passing drill clears it") and the decider saw no point in another try (pull #1721 closed)

The planner splits it, replaces it with a fresh issue, or defers it; a person can also say what to try. Nothing starts here until then, and the board lists it.

Deferred behind the Later gate (#595). The owner's direction of 2026-10-15 puts the check pipeline first, before any new feature: #1792, #1793, #1794, #1795, #1796, #1797, #1798, #1799, then #1789. One try died in the preview deploy and the other on its own drill test; with checks now the priority, restore-drill alerting waits until the pipeline (#1795–#1797) gives it a stable run.

replaced by #1804. Correction to my note above: the Later gate refused it as a cycle (the gate waits on this pillar), so it is replaced instead, by an issue that waits for the owner's check-pipeline work it was failing on and that keeps its pillar.