Say when a vault has moved, so nobody waits out the minute

The delta pull was cheap enough to run on a timer and the client did, once a
minute. That is fine for a machine and wrong for two people: an edit a colleague
makes is up to a minute stale, which is long enough for both of them to make it
and produce a conflict neither needed to have. Shortening the interval is the
obvious answer and the wrong one — it costs a request per client per interval
whether or not anything happened, and it converges on a busier server that is
still late.

So the server now says so. A client holds a WebSocket open at GET /api/v1/events,
subprotocol dodossh.events.v1, and gets a line down it when something it can read
has changed. ADR 0012 has the reasoning; three parts of it are worth repeating
here, because they are what everything else rests on.

**What crosses the socket is a notice, never data.** A frame names a vault and
how far its change log has got. No item, no ciphertext, not even which item it
was. The client's answer is the delta pull it would have run anyway, so there is
still exactly one code path that applies a change to a keychain, and it is not
this one. Pushing the items themselves would save a round trip and fork that path
in two, with the cursor, the merge and the tombstone rules duplicated across both
— ADR 0003 put every mutation through one write path for that reason, and this
keeps every read on one for the same one. It also makes a dropped notice
harmless, which is what lets the fan-out below be as simple as it is.

**Polling stays, and is what guarantees a pass.** The minute timer is unchanged.
A network that eats WebSockets, a server with Events:Enabled off, an older
server, a proxy that will not upgrade, a notice dropped under backpressure —
every one of those leaves a client behaving exactly as it did before this commit.
Nothing is reachable only over the socket and nothing is meant to become so;
VaultViewModel's AutoSyncInterval remark now says that where somebody changing it
will read it.

**The bearer token authorises the upgrade, unlike the relay's ticket.** Not an
inconsistency with ADR 0004: the relay's socket is a byte pipe whose whole
authorization decision — which host, which IPs, which port — is made before it
opens and never revisited, and it is the extraction seam for a process that must
hold no ACL code. This one is a view of the caller's own vault list and has to
keep answering "what may this account read" for as long as it is held. A ticket
would carry that answer in a token and be wrong the moment the account's access
changed. The two bounds that arrangement needs are met rather than waved at: the
socket is closed at the token's exp with close code 4401 and the client comes
straight back with a fresh one, and the vault set is re-resolved every few
minutes as well as on the changes known to affect it. Both bound *metadata*,
because a notice contains nothing else and reading a vault still needs a key this
server has never held.

**The fan-out.** VaultEventHub is a singleton holding the sockets this node
accepted; publishing walks them and asks each whether it cares, rather than
keeping a vault-to-subscriber index that every re-subscription would have to move
entries between under a lock publishing also takes. At a few hundred sockets per
node and an event rate bounded by how often people edit keychains, the walk is
not measurable and its races are obvious. Per-connection queues are bounded and
drop the *oldest*: a notice means "pull vault X, which is at least at sequence
N", so the newest subsumes what it displaces and the client's answer is identical
either way — which is what lets the publish path be void, never block, and never
fail.

Announced from the endpoint rather than from SyncService, and that placement is
the point: by then the push has committed and released the per-vault advisory
lock. From inside it would name a sequence no reader can see yet and would hold
the lock that serialises writers across a socket write. Only the highest
*applied* sequence, so a batch of pure conflicts announces nothing, and a
duplicate — already announced when it first landed — announces nothing either.

Grants and membership publish too, and those take the *recipient* rather than the
actor. This is what AdmitNewVaultsAsync has been apologising for since sharing
shipped — "the recipient is handed nothing, there is no push channel" — and the
README with it. A vault shared with somebody now turns up as it is shared. The
comment and the README paragraph both say what is true now, and both keep saying
that the pass is what *discovers* the vault, because a client with no socket has
to arrive at the same place.

**On the client**, VaultEventStream is really a reconnection policy wrapped round
a ClientWebSocket: a dropped socket is the ordinary case here — laptops sleep,
proxies time out, tokens expire, servers are redeployed — so nothing in it treats
a failure as exceptional, and every path ends in "wait, then dial again". A
connection that lived long enough to say hello resets the backoff, so a laptop
that woke, worked, and lost its network an hour later does not inherit a
minute-long wait it has already proved it need not take. A 4401 close skips the
backoff entirely and asks the token provider again, which is the whole reason
that close code is distinct. A server that does not advertise the events feature
gets IdleVaultEventStream, which never delivers — so IVaultServer.Events is never
null and every caller stays on one shape, because the correct behaviour without a
socket is the behaviour with a silent one.

The shell's background loop now selects between the timer and a notice, and both
waits are held across iterations. That is load-bearing rather than tidy:
PeriodicTimer permits one outstanding WaitForNextTickAsync and throws on a
second, and an abandoned channel read stays registered and consumes the next
notice written. Either defect leaves the first notice working and every one after
it silently lost, which is why NoticesKeepWakingTheLoop_NotJustTheFirst pushes
three and not one. Notices are coalesced over a quarter of a second, so one
person's save — a host and its log entry are two items — and a colleague clearing
a folder each cost one pass rather than a dozen.

**The kind is a string, not an enum**, and that is a compatibility decision.
UseStringEnumConverter throws on a value it does not know, so a newer server
sending a kind an older client had never heard of would not add an unreadable
frame — it would break that client's socket outright. A string is ignored
instead. ProblemCodes is the same shape for the same reason.

**Tested on both sides, through the real pipeline.** The endpoint suite opens a
genuine socket against TestServer and proves a push produces a notice, that
another account's push does not reach it, that a ping is answered, and that a
frame this server cannot parse does not end the connection. Two of those assert
on *ordering* rather than on absence within a timeout — the stranger's write goes
first, so a socket that leaked would have announced it before the one the test
waits for — because "nothing arrived in two seconds" is a test that passes on a
slow machine for the wrong reason. And ANoticeCarriesNoCiphertext asserts on the
bytes that crossed the wire rather than on the record's fields, since the latter
would only prove that this type has no payload member, which is a tautology; the
former is what catches a field added later without anybody thinking about
disclosure.

The client suite drives VaultEventStream through an injected connector, because
the one thing a test cannot do to a real network is make it fail on cue — and
failure is the entire subject. The shell suite proves a notice produces a pull
inside ten seconds against a sixty-second timer, so the timer cannot be what
caused it.

**Two limits, stated rather than left to be discovered.** Fan-out is in-process,
so a deployment running more than one API replica only pushes for writes its own
replica handled and the rest arrive on the timer. IVaultEventPublisher is the
seam a PostgreSQL LISTEN/NOTIFY backplane implements and it is deliberately not
implemented: an untested backplane is worse than a documented gap, and multiple
replicas degrade to the behaviour before this commit rather than breaking. And a
client is notified of its own writes; it pushed, so it already pulled, and the
extra pass finds nothing. Suppressing that echo correctly needs a per-device
identity on the socket, and the same user's other machines must still be told.

Manual checks phase 15 covers what no test here can reach, which is the network
in between: a proxy that will not upgrade, one that drops an idle socket without
telling either end, a laptop lid, a token expiring. Every one of those is
invisible inside a test host, and every check there passes only if the change
arrives quickly *and* still arrives with the socket taken away.

ADR 0012 also fixes one thing about the shared terminal session this is the
transport for, so it need not be renegotiated later: session data will be binary
frames on this same socket, because base64 in a JSON envelope is the wrong shape
for the one payload here that is continuous rather than occasional. Two questions
it explicitly does not answer by implication — whether those bytes go through the
API at all, and what end-to-end encryption means when the second party watches a
stream rather than holding a key — are ADR 0001 questions and get their own
decision.

1512 tests pass. DodoSSH.SystemTests was not run — it needs the whole compose
stack — so the end-to-end path is unverified for this change beyond what the
manual checks describe.

README.md

@@ -350,6 +350,42 @@ Both default to your personal vault and neither moves on its own, because an ite
 visible to everybody holding that vault's key. Choosing a vault in the host editor also decides which groups
 it can be filed under: a group is an item like any other and lives in exactly one vault.
 
+### Changes that do not wait
+
+A client holds a WebSocket open to the server — `GET /api/v1/events`, subprotocol
+`dodossh.events.v1` — and the server sends a line down it whenever something you can read has moved. The
+client's answer is the same delta pull it would have run on its timer, only now rather than in up to a
+minute. Two things you can see: an edit somebody else makes appears while you are looking at the list, and
+a vault shared with you turns up as soon as they share it.
+
+**What is on that socket is a notice, not your data.** A frame says which vault changed and how far its
+change log has got, and nothing else: no item, no ciphertext, not even which item it was. That is the
+decision the rest of this rests on, and it is deliberate twice over — the server has nothing else it
+*could* send, and keeping it that way means there is still exactly one path that applies a change to your
+keychain, so the socket can be wrong or absent without anything being applied incorrectly.
+
+**Polling is still there and is still what guarantees a pass.** The minute timer is unchanged. A network
+that eats WebSockets, a server with `Events:Enabled` off, an older server, a proxy that will not upgrade, a
+notice dropped because your machine was too slow to read it — every one of those leaves you with exactly
+what this product did before the socket existed. Nothing is reachable only this way, and nothing is
+supposed to become so.
+
+Three limits are worth knowing rather than discovering:
+
+- **One node.** Fan-out is in-process, so a deployment running more than one API replica only pushes for
+  writes that its own replica handled. The rest arrive on the timer. The seam for a PostgreSQL
+  `LISTEN`/`NOTIFY` backplane is in place and is not implemented, because an untested backplane would be
+  worse than a documented gap.
+- **The socket does not outlive your access token.** It is closed at the token's expiry and the client
+  reconnects with a fresh one, which is a gap you will not see. That, plus re-reading your vault list every
+  few minutes, is what bounds how long a withdrawn grant can keep producing notices — and what it bounds is
+  *metadata*, because reading a vault needs a key the server has never held.
+- **You are told about your own writes.** Your client pushed, so it has already pulled; the extra pass finds
+  nothing. Notices are coalesced over a quarter of a second so that a burst is one pass rather than a dozen.
+
+The reasoning, including why this is a WebSocket rather than server-sent events and where a shared terminal
+session will attach to it, is in [ADR 0012](docs/adr/0012-realtime-push.md).
+
 ### The Android head
 
 `src/DodoSSH.Client.Android` is a phone-first head that shares every view model with the desktop one — the
@@ -618,6 +654,20 @@ keychain plus a terminal — and the spike that gates all of it.
   directories, an interrupted **upload** starts again rather than resuming (an object cannot be written from
   the middle), and a rename is a copy and a delete rather than one atomic operation. Downloads do resume — a
   ranged GET is part of the protocol, which is the one place a bucket beats SFTP.
+
+  *Realtime done:* a WebSocket the client holds open, over which the server says which vault has moved so a
+  pull happens now rather than within the minute. What crosses it is a notice and never an item, which is
+  what keeps one code path applying changes and makes a dropped socket cost latency and nothing else — the
+  timer is unchanged and is still the guarantee. Two limits are stated rather than implied: fan-out is
+  in-process, so a multi-replica deployment falls back to the timer for writes another replica handled, and
+  a socket is closed at its access token's expiry rather than outliving the credential that authorised it.
+  See [Changes that do not wait](#changes-that-do-not-wait) and
+  [ADR 0012](docs/adr/0012-realtime-push.md).
+
+  It is also the transport a **shared terminal session** will use — one person's shell, watched or driven by
+  somebody else. Nothing of that exists yet, and ADR 0012 records the one decision made early so it need not
+  be renegotiated: session data will be binary frames on this same socket, because base64 in a JSON envelope
+  is the wrong shape for the one payload here that is continuous rather than occasional.
 - **M3 — shared vaults**, sharing, ACLs. *Done.* Membership with roles, a public-key directory, the
   append-only key log served for clients to verify against, shared vaults, and vault key grants
   wrapped by a client and stored opaquely by the server. `VaultAccessService` now resolves team
@@ -641,10 +691,12 @@ keychain plus a terminal — and the spike that gates all of it.
   the rotation is re-sealed as it is pushed, so nothing reaches the server under a superseded key at all.
   See [ADR 0010](docs/adr/0010-vault-key-rotation.md).
 
-  **A vault shared with you arrives on the next synchronisation pass**, within the minute, with no sign-in
-  and nothing to press. There is no push channel, so each pass asks the server which vaults this account can
-  reach before syncing the ones it already knows — which is also how a vault that has been deleted, or one
-  whose grant was withdrawn, stops being listed.
+  **A vault shared with you arrives at once**, with no sign-in and nothing to press. Each pass asks the
+  server which vaults this account can reach before syncing the ones it already knows — which is also how a
+  vault that has been deleted, or one whose grant was withdrawn, stops being listed — and the server now
+  says so the moment somebody wraps a key to you rather than leaving it for the next pass. Without a
+  reachable socket that becomes "within the minute", which is what it always was; see
+  [Changes that do not wait](#changes-that-do-not-wait).
 
   **Ownership transfer is here, and it is one write rather than two.** The member you name becomes owner
   and you become an admin, in a single transaction — because ownership is sole, so promoting first leaves