Continuity

What happens if I stop

One person builds and runs this. That is the single biggest weakness in the whole thing, and pretending otherwise would be the wrong way to run a warning channel.

The page tells you itself

After ten days with no completed scan, a banner appears on every page saying plainly that the person running this may no longer be able to — including the possibility that they are dead.

That check runs inside your own browser, against the timestamp of the last scan. It does not ask the server whether the server is working, because a system that has stopped cannot report that it has stopped. It keeps working even when everything behind it is gone.

Ten days is deliberately long. A banner that appears every time a scheduler drifts is noise, and noise on a warning page teaches people to ignore the one banner that matters.

Nothing restarts itself

If I stop, the scans keep running on their schedule until a key expires or a free quota runs out. Then the site goes still — while still looking normal to anyone who does not check the date. That gap between “looks fine” and “is fine” is the thing the banner exists to close.

If you want to continue it

There is a written handover — architecture, what runs where, which credentials exist and what each is for — kept in the public site repository rather than the private one, because a handover note nobody can reach is not a handover note.

Read CONTINUITY.md — written for a stranger with none of the context.

The complete published record — every scan result, the full alert history, the corrections — is in public files with full version history. Anyone can audit or fork the track record without asking permission.

One request: if you rebuild this, do it under your own name. The corrections log and the independence claim on the front page are promises made by a specific person, and a warning channel that changes hands invisibly is the exact failure this project argues against.

What is still not solved

Stated rather than hidden — a continuity plan that oversells itself is worse than none:

• No second person holds the credentials. This is the real single point of failure, and none of the above fixes it. It only makes the aftermath survivable.
• No organisation owns anything. The domain and the accounts sit with one individual, which makes an orderly handover harder than it should be.
• The keyword lists have no succession path. They are private, they are the hard-won part, and nobody else has a copy.

These are listed here so that nobody has to discover them at the worst possible moment.

Questions, or an offer to help with any of the above: [email protected].