Follow the SDK's ILLink version into the three lock files that pin it

CI's restore stopped on a commit that changed no dependency and touched none of
these files:

    error NU1004: The package reference Microsoft.NET.ILLink.Tasks version has
    changed from [10.0.10, ) to [10.0.11, ).

◆ NOTHING HERE MOVED. THE RUNNER'S SDK DID. Microsoft.NET.ILLink.Tasks is
referenced implicitly by the SDK — no .csproj asks for it — and its version
tracks the runtime patch band. global.json pins 10.0.100 with
rollForward: latestMinor, so setup-dotnet installs whatever the newest 10.x SDK
is on the day. When that band moved to 10.0.11, every lock file carrying the
entry went stale at once, and RestoreLockedMode is what turns that into a stop
rather than a silent upgrade. Three files carry it: the Android head, Contracts
and Crypto.

Contracts and Crypto are regenerated by --force-evaluate under SDK 10.0.303,
which bundles the same 10.0.11 the runner has; the contentHash is NuGet's, from
that restore. dotnet restore DodoSSH.slnx --locked-mode is clean under it.

◆ THE ANDROID LOCK FILE IS HAND-EDITED AND WAS NOT VERIFIED BY A RESTORE. Same
three lines, same version, same hash the real restore produced — but generated
by an editor, not by NuGet. Installing 10.0.303 rewrote the machine-wide
workload manifests for the 10.0.300 band, after which that project fails
NETSDK1147 asking for wasm-tools (which nothing here targets) under 10.0.302 as
well as 10.0.303, so there was no working local restore left to produce it
honestly. The android workload on that machine is Visual Studio-managed and
repairing it is not a thing to do in passing. This is the hazard the docs
already record — the Android head is outside DodoSSH.slnx, so the locked-mode
gate that keeps the other fourteen lock files honest has never seen it — landing
again, from the other direction. If the android job still fails on NU1004, this
line is the first thing to doubt.

platform-flags.md gains the failure beside the existing lock-file hazards,
including the trap that cost the most time here: --force-evaluate from an SDK
older than the runner's rewrites the lock at the OLD version, changes nothing,
and looks like it worked. Check dotnet --version against the version in the
error before believing a regeneration.

This recurs on every SDK patch that moves the band. That is the accepted cost of
letting the SDK float; pinning an exact SDK trades a recurring lock-file bump
for a recurring toolchain bump and one mandated SDK per contributor.

docs/design-import-gaps.md

@@ -300,7 +300,7 @@ the chrome, hosts and terminals, file transfer, the vault, teams, and preference
 > | The status bar's negotiated cipher, host-key algorithm and key/credential name | ◆ **Shipped, on both surfaces, with three honest deviations.** `ISshConnection` and `ISftpSession` now both carry `Cipher` — the server-to-client algorithm off SSH.NET's own `ConnectionInfo.CurrentServerEncryption`, captured once at construction because a rekey is not an event SSH.NET raises — and `TerminalWorkspace.GetSessionFacts` hands the cipher and the host key's algorithm back to the shell the moment a session opens; `VaultViewModel.TryBuildAuthentication` now threads the authenticating key's or credential's own `Label` into `HostAuthentication.IdentityLabel`, all the way to `MainWindowViewModel`'s surface-aware `SessionCipher`, `SessionHostKeyAlgorithm` and `SessionIdentityLabel`, composed into one `SessionIdentityText` run for the status bar. Three deviations from the mock, not omissions: the algorithm prints exactly as negotiated (`ssh-ed25519`), not the design's shortened `ed25519`, because trimming it would be an edit to a string this client did not choose; the run is plain text rather than the design's clickable element, because there is no pin-details modal for a session that is already open, and drawing a click target for a screen that does not exist would itself be a fabrication; and a typed-password session — nothing filed in the keychain to name — shows the host-key algorithm alone, with no `·` after it, because there is no item behind the dot. |
 > | S3 dimmed in the design's own switcher | **Enabled.** The mock leaves S3 as future work; this application already has bucket browsing, so SSH, SFTP and S3 are a true three-way segment, wired to `IsSshShowing`, `IsTransfersShowing` and `IsBucketsShowing` exactly alike. |
 > | The S3/Buckets screen | **Did not get the session shell in v5b.** `TransfersScreen` serves both SFTP and S3 today and only the SFTP usage in `MainWindow.axaml` sat inside the new tab row/header/status bar/sidebar; the S3 usage was unchanged at the time. **v5c gives it the shell's own look without the machinery** — a 26-pixel padded, bordered, radius-12 container and nothing past that, since a bucket has no tab to close, no host to head a card with and no pin for a sidebar to show; see the v5c section, below. |
-> | No pins destination in the design at all | **Kept anyway.** The rail still carries Pins — `KnownHostsScreen` — because the mock has no screen for approved host keys and this application's has to stay reachable. |
+> | No pins destination in the design at all | **The rail agrees with the design now.** `KnownHostsScreen` is still built and still reachable — from **Host keys** on the Keys screen's own header, which was always the second way in — but the rail's Pins row is gone. It was kept through v5b on the grounds that the mock has no screen for approved host keys, which is a reason for the screen to exist and was never a reason for a rail entry once the keychain had a door to the same place. Two rail rows landing on one screen is a rail that has to be read twice. |
 > | The popover's Settings and Preferences rows, and the design's own Settings-* family of screens | **Landed in v5c.** What was two doors to one room in v5b — Settings and Preferences both opening the same bare `Preferences` screen — is now two of three doors onto their own settings pages: Settings opens General, Preferences opens Preferences, and a third row, Vaults, opens Vaults. All three are real, distinct pages inside one settings mode; see the v5c section, below. |
 > | `· Org` after the user chip's name, and a `Primary` tag on a vault row in the popover | Neither. There is no organisation concept behind a vault — only the vault itself — and no vault is distinguished as primary; the popover's vault rows are the existing shown-vaults toggles, restyled. |
 > | The design's titlebar, which has nowhere for a sync indicator | `SYNCED` stays, on the titlebar's right side, ahead of the window's own minimise/maximise/close buttons — the one thing this titlebar keeps that the design's own does not draw at all. |

docs/manual-checks.md

@@ -28,8 +28,9 @@ a phase had nothing left for a person to do, which is the good outcome rather th
 
 ### 1.1 No screen is sliced at the WebView's left edge · **the important one**
 
-Open two terminals, then visit every nav rail entry in turn — Hosts, Keys, Pins, Snips, Logs — and both of
-the switcher's other two segments, SFTP and S3, at the rail's own head.
+Open two terminals, then visit every nav rail entry in turn — Hosts, Keys, Snips, Logs — and both of
+the switcher's other two segments, SFTP and S3, at the rail's own head. The pins screen is no longer a rail
+entry; reach it from **Host keys** on the Keys screen's header and check it the same way.
 
 **Pass:** each screen draws whole, its buttons all clickable, and the nav rail stays up the left edge for
 every one of them. Since v5b's chrome pass the rail is permanent furniture — it no longer collapses for

docs/platform-flags.md

@@ -712,6 +712,30 @@ The lasting hazard is the first paragraph and not the fix. Any change to a share
 to a lock file this repository cannot verify from a machine without the Android workload, and it will go
 on being noticed later than every other one.
 
+**A lock file can go stale with nothing in this repository changing, because `Microsoft.NET.ILLink.Tasks`
+is versioned by the SDK and `global.json` lets the SDK float.** The reference is implicit — nothing in any
+`.csproj` asks for it — and its version tracks the runtime patch band, while `global.json` pins only
+`10.0.100` with `rollForward: latestMinor`. So `setup-dotnet` installs whatever the newest 10.x SDK is on
+the day, and the moment that SDK's band moves, locked-mode restore stops:
+
+```
+error NU1004: The package reference Microsoft.NET.ILLink.Tasks version has changed
+from [10.0.10, ) to [10.0.11, ).
+```
+
+It named `DodoSSH.Client.Android`, `DodoSSH.Contracts` and `DodoSSH.Crypto` — the three lock files that
+carry the entry — on a commit that touched none of them and no dependency at all.
+
+The fix is `--force-evaluate` on those three, **from a machine whose SDK is at least as new as the
+runner's**, which is the part that is easy to get wrong: a `--force-evaluate` from an older SDK rewrites
+the lock at the older version, changes nothing, and looks like it worked. Check `dotnet --version` against
+the version in the error before believing a regeneration.
+
+This will recur on every SDK patch that moves the band. It is the accepted cost of letting the SDK float:
+the alternative is pinning an exact SDK in `global.json`, which trades a recurring lock-file bump for a
+recurring toolchain bump and makes every contributor install one specific SDK. Neither is free, and this
+repository has chosen the floating side deliberately.
+
 **.NET for Android cannot be built on a musl host, and this project's runner is Alpine. Every message the
 toolchain produces on the way to saying so names a missing file that is present.** Three CI rounds went
 into this and the first two fixed symptoms, so the messages are worth reading in the order they arrive.

src/DodoSSH.Client.Android/packages.lock.json

@@ -74,9 +74,9 @@
       },
       "Microsoft.NET.ILLink.Tasks": {
         "type": "Direct",
-        "requested": "[10.0.10, )",
-        "resolved": "10.0.10",
-        "contentHash": "f5VCIE7AJpd5YvzNTeMGVzQIgyE9tX+AreTYwQF+REbu+DZo/2Ae+jNSwhPEYrVz6RRkd7y8ubXjk6Nn6Ka+Cg=="
+        "requested": "[10.0.11, )",
+        "resolved": "10.0.11",
+        "contentHash": "IBf7lbovvjGWVWXZX5cJ/cO0WXbId0Zq4BuSeT94mGZuOAP66oMeH9PTBZ9Jpp3Jb6jtK0qm/NyUbPRo1gC/wQ=="
       },
       "MinVer": {
         "type": "Direct",

src/DodoSSH.Client.App/App.axaml

@@ -913,18 +913,30 @@
       into a grid: equal columns, and a card that grew a third line of tags is taller than its neighbours
       rather than narrower.
 
-      ◆ 224 IS DERIVED, and the arithmetic is written out because getting it wrong is invisible. The grid's
-      column at the window's minimum is 1016 less the rail's 190 and the drawer's 320, which is 506. The
-      scrolling stack inside it takes 16 of margin on each side, and the vertical scrollbar takes its own —
-      call the usable width 474. A WrapPanel fits floor(474 / (Width + 10)) per row, so two columns needs
-      Width no more than 227.
+      ◆ 214 IS DERIVED, and the arithmetic is written out because getting it wrong is invisible. The grid's
+      column at the window's minimum is 1081 less the rail's 255 and the drawer's 320, which is 506. The
+      scrolling stack inside it takes 26 of margin on each side, and the vertical scrollbar takes its own —
+      call the usable width 454. A WrapPanel fits floor(454 / (Width + 10)) per row, so two columns needs
+      Width no more than 217, and 214 is that with the same few pixels of slack the previous number kept.
 
-      The first number here was 248, from the same reasoning with the two margins left out. It laid out
-      cleanly and the layout harness passed it, because the harness asks whether a control is inside the
-      window and not how many of them fit on a line — so the grid quietly became one column wide at exactly
-      the size this application guarantees, which is the shape the cards exist to avoid. The second was 232,
-      derived the same way against the drawer's own 304; v5 widened the drawer to 320 for the ADDRESS field's
-      breathing room, which narrowed the budget this number is drawn from and had to move it down in step.
+      Every number here has moved at least once, and always because something beside the cards did:
+
+      · 248, from this reasoning with the two margins left out. It laid out cleanly and the layout harness
+        passed it, because the harness asks whether a control is inside the window and not how many of them
+        fit on a line — so the grid quietly became one column wide at exactly the size this application
+        guarantees, which is the shape the cards exist to avoid.
+      · 232, derived against the drawer's own 304, which v5 widened to 320 for the ADDRESS field's breathing
+        room — narrowing the budget this number is drawn from and moving it down in step.
+      · 224, which is what that gave. The stated arithmetic still said 1016 and 190 by then: v5b's rail took
+        190 to 255 and the window's minimum 1016 to 1081 in the same pass, so the two changes cancelled and
+        the answer stayed right while the working went stale.
+      · 214, now that HostsScreen's board is inset 26 a side rather than 16 — see that file's own remark on
+        why every screen frames its content the same way. Twenty pixels of board is twenty pixels the cards
+        no longer have, and this is where they come from.
+
+      ◆ THE TEST THAT CATCHES THIS IS NOT THE HARNESS. See
+      ScreenLayoutTests.TheHostsGridKeepsTwoColumnsAtTheMinimumWithTheDrawerOpen, which counts columns
+      because that is the thing this number exists to buy and the thing no fit assertion can see.
     -->
     <Style Selector="Border.tile">
       <Setter Property="Background" Value="{StaticResource Raised}" />
@@ -932,7 +944,7 @@
       <Setter Property="BorderThickness" Value="1" />
       <Setter Property="CornerRadius" Value="12" />
       <Setter Property="Padding" Value="12,10" />
-      <Setter Property="Width" Value="224" />
+      <Setter Property="Width" Value="214" />
       <Setter Property="Margin" Value="0,0,10,10" />
     </Style>
     <Style Selector="ListBoxItem:pointerover Border.tile">

src/DodoSSH.Client.App/Views/HostsScreen.axaml

@@ -98,6 +98,18 @@
 
   <Grid ColumnDefinitions="*,Auto">
 
+    <!--
+      ── 26 DOWN EACH SIDE, the same inset Keychain, Snips, Logs and Pins all take. ──────────────────────
+      Those four say it once, as Margin="26" on their own root; this screen repeats it on each of the four
+      rows below, and it has to. The board's ScrollViewer is the last row and is deliberately full-bleed, so
+      that its scrollbar rides the pane's own edge rather than floating 26 pixels inside it — a root margin
+      would inset the bar with everything else. It would also inset the drawer in the second column, which
+      draws its own edge and wants none.
+
+      It was 16 and 20 until this pass, which put the Hosts header a visible step left of and above every
+      other screen's. Four numbers rather than one is the cost of the two exceptions above; changing one of
+      them means changing all four.
+    -->
     <Grid Grid.Column="0" RowDefinitions="Auto,Auto,Auto,*">
 
       <!--
@@ -107,7 +119,7 @@
         buttons over a board of forty is a pair whose subject the user has to work out. The group's own
         Edit/Move/Delete sit on its own heading's menu for the same reason.
       -->
-      <Grid Grid.Row="0" Margin="16,20,16,16" ColumnDefinitions="Auto,Auto,*,Auto,Auto,Auto">
+      <Grid Grid.Row="0" Margin="26,26,26,16" ColumnDefinitions="Auto,Auto,*,Auto,Auto,Auto">
 
         <TextBlock Grid.Column="0" Text="Hosts" FontSize="33" FontWeight="Bold" LetterSpacing="-0.5"
                    Foreground="{StaticResource Text}" VerticalAlignment="Center" />
@@ -219,7 +231,7 @@
         Ctrl+K is named on it because the palette is the other way to reach a host by typing, and somebody
         who has found this box should know about the one that also connects on Enter.
       -->
-      <Border Grid.Row="1" Margin="16,0,16,16">
+      <Border Grid.Row="1" Margin="26,0,26,16">
         <TextBox x:Name="HostFilter" Text="{Binding HostFilter}" Height="40" CornerRadius="10"
                  FontFamily="{StaticResource MonoFont}"
                  PlaceholderText="Find a host by name, address or note…   ·   Ctrl+K searches and connects" />
@@ -232,7 +244,7 @@
         one of them sits here, above the board, rather than laid over it: a card over the cards would hide
         the very ticks or the very group it is asking about.
       -->
-      <StackPanel Grid.Row="2" Margin="16,0,16,12" Spacing="10">
+      <StackPanel Grid.Row="2" Margin="26,0,26,12" Spacing="10">
 
         <!--
           The conflict log. The merge is only allowed to pick a winner because the value it overrode is kept
@@ -434,7 +446,7 @@
         HostsScreen.axaml.cs.
       -->
       <ScrollViewer Grid.Row="3" x:Name="Scroll" HorizontalScrollBarVisibility="Disabled">
-        <StackPanel Margin="16,0,16,20" Spacing="16">
+        <StackPanel Margin="26,0,26,26" Spacing="16">
 
           <!--
             Named because it is where keyboard focus lands when the terminal gives it back, and because

src/DodoSSH.Client.App/Views/MainWindow.axaml.cs

@@ -54,6 +54,21 @@ internal sealed partial class MainWindow : Window
         };
     }
 
+    /// <summary>
+    /// Colours the system-drawn frame the moment there is a handle to colour it on.
+    /// </summary>
+    /// <remarks>
+    /// <c>OnOpened</c> and not the constructor: the window has no platform handle until it is shown, and
+    /// <see cref="NativeWindowFrame"/> does nothing without one. See that class for what the frame is and
+    /// why <c>BorderOnly</c> still has one.
+    /// </remarks>
+    protected override void OnOpened(EventArgs e)
+    {
+        base.OnOpened(e);
+
+        NativeWindowFrame.MatchTo(this);
+    }
+
     /// <summary>
     /// Asks the Linux backend for the one mode it can actually draw inside this window.
     /// </summary>

src/DodoSSH.Client.App/Views/NativeWindowFrame.cs

@@ -0,0 +1,133 @@
+using System.Runtime.InteropServices;
+using Avalonia.Controls;
+using Avalonia.Media;
+
+namespace DodoSSH.Client.App.Views;
+
+/// <summary>
+/// Paints the frame Windows still draws around a <c>BorderOnly</c> window in the application's own
+/// colour, so the top edge stops reading as a leftover system titlebar.
+/// </summary>
+/// <remarks>
+/// <para>
+/// <b>The symptom this exists for:</b> a pale strip across the very top of the window, a few pixels
+/// tall and plainly not part of the application — most obvious on a machine with "show accent colour
+/// on title bars and window borders" turned on, where it comes out blue against a near-black shell.
+/// </para>
+/// <para>
+/// It is not <c>TitleBar.axaml</c> leaking and it is not a margin. It is DWM, and the reason it is
+/// there is visible in Avalonia's own Win32 backend: <c>WindowImpl.UpdateWindowProperties</c> gives a
+/// <see cref="WindowDecorations.BorderOnly"/> window <c>WS_BORDER | WS_THICKFRAME</c> and then calls
+/// <c>DwmExtendFrameIntoClientArea</c> with one-pixel margins on all four sides. So the compositor
+/// owns a hairline of every edge of this window, and it fills that hairline with the system's caption
+/// and border colours — which are chosen by the user's personalisation settings and have no reason to
+/// resemble <c>CanvasColor</c>. The window is the wrong place to look for the pixels; they were never
+/// painted by anything in this tree.
+/// </para>
+/// <para>
+/// The fix is to tell DWM what colour to use rather than to try to cover it. <c>DWMWA_BORDER_COLOR</c>
+/// and <c>DWMWA_CAPTION_COLOR</c> arrived in Windows 11 21H2 and are exactly that; both are set to the
+/// window's own background, so the hairline still exists — the resize grip is on it, and the drop
+/// shadow hangs off it — and simply cannot be seen. Deliberately <em>not</em> <c>DWMWA_COLOR_NONE</c>,
+/// which removes the border outright: on a dark desktop that leaves a near-black window with no edge
+/// at all, which trades one visual defect for another.
+/// </para>
+/// <para>
+/// Windows 10 gets the dark-mode attribute and nothing else, and that is the whole of what is
+/// available there: the two colour attributes are unsupported, <c>DwmSetWindowAttribute</c> answers
+/// <c>E_INVALIDARG</c>, and the calls do nothing. Hence the ignored return values — every attribute
+/// here is an improvement where it lands and a no-op where it does not, so there is nothing for a
+/// caller to handle and nothing worth logging on a path that runs once at startup.
+/// </para>
+/// </remarks>
+internal static class NativeWindowFrame
+{
+    /// <summary>Windows 11 21H2 and later: the colour of the frame border.</summary>
+    private const int BorderColorAttribute = 34;
+
+    /// <summary>Windows 11 21H2 and later: the colour of the caption, including the extended frame.</summary>
+    private const int CaptionColorAttribute = 35;
+
+    /// <summary>
+    /// Windows 10 1903 and later: draw the frame in the dark palette.
+    /// </summary>
+    /// <remarks>
+    /// Redundant on Windows 11, where the two colour attributes above name the colours outright, and it
+    /// is set anyway because it is the only one of the three that Windows 10 honours. The build before
+    /// 1903 used attribute 19 for this; that is not chased here, because a border on an OS release that
+    /// left support in 2020 is not worth a second interop call.
+    /// </remarks>
+    private const int DarkModeAttribute = 20;
+
+    /// <summary>
+    /// Matches <paramref name="window"/>'s system-drawn frame to the colour it paints itself.
+    /// </summary>
+    /// <remarks>
+    /// Call once the window has a handle — <c>OnOpened</c> is the first such moment. Calling earlier
+    /// finds no platform handle and silently does nothing, which is the defect this replaced: the strip
+    /// is only visible once the window is on screen, so a call that ran too early looks like a fix that
+    /// does not work rather than a fix that never ran.
+    /// </remarks>
+    internal static void MatchTo(Window window)
+    {
+        // Every attribute below is a DWM one, and DWM is Windows. Elsewhere the frame is drawn by the
+        // platform's own compositor and there is nothing here to say to it.
+        if (!OperatingSystem.IsWindows())
+        {
+            return;
+        }
+
+        if (window.TryGetPlatformHandle()?.Handle is not { } handle || handle == IntPtr.Zero)
+        {
+            return;
+        }
+
+        Set(handle, DarkModeAttribute, 1);
+
+        // The window's own Background rather than a named resource, so the frame cannot drift from the
+        // canvas when the palette moves. A brush that is not solid — a gradient, or nothing set at all —
+        // has no single colour to match, and leaving the system's own is better than inventing one.
+        if (window.Background is not ISolidColorBrush { Color: var canvas })
+        {
+            return;
+        }
+
+        var reference = ColorRef(canvas);
+        Set(handle, BorderColorAttribute, reference);
+        Set(handle, CaptionColorAttribute, reference);
+    }
+
+    /// <summary>
+    /// Sets one integer-valued DWM attribute, and discards the answer.
+    /// </summary>
+    /// <remarks>
+    /// The discard is the point of this method existing rather than being three call sites. Every
+    /// attribute here is unsupported on some Windows this application runs on, and unsupported means
+    /// <c>E_INVALIDARG</c> and no change — which is the intended outcome on that OS, not a failure, so
+    /// there is nothing for the caller to do with the <c>HRESULT</c> and nothing worth logging once at
+    /// startup. Written once, with the reasoning, rather than left implicit at each call.
+    /// </remarks>
+    private static void Set(IntPtr window, int attribute, int value) =>
+        _ = DwmSetWindowAttribute(window, attribute, ref value, sizeof(int));
+
+    /// <summary>
+    /// Packs <paramref name="color"/> into a Win32 <c>COLORREF</c>.
+    /// </summary>
+    /// <remarks>
+    /// <c>0x00BBGGRR</c> — blue in the high byte, not red, and the alpha byte must be zero. Getting the
+    /// order wrong produces a plausible-looking wrong colour rather than an error, which is the kind of
+    /// bug that survives a glance at the window.
+    /// </remarks>
+    private static int ColorRef(Color color) => color.R | (color.G << 8) | (color.B << 16);
+
+    /// <remarks>
+    /// <c>DllImport</c> rather than <c>LibraryImport</c>, for the reason
+    /// <see cref="NativeKeyboardFocus"/> gives at its own P/Invoke: the generated form needs
+    /// <c>AllowUnsafeBlocks</c> across a project that handles key material, and this signature is
+    /// blittable, so there is no marshalling for it to improve.
+    /// </remarks>
+#pragma warning disable SYSLIB1054
+    [DllImport("dwmapi.dll")]
+    private static extern int DwmSetWindowAttribute(IntPtr window, int attribute, ref int value, int size);
+#pragma warning restore SYSLIB1054
+}

src/DodoSSH.Client.App/Views/NavRail.axaml

@@ -40,8 +40,9 @@
     Both are still one click away; see the popover below the user chip. The chip itself carries the signed-
     in identity this application actually has — a display name and, where the server sent one, an email —
     which is also new: the titlebar drew an account name and a vault chip before this pass and does not any
-    more. See TitleBar.axaml and design-notes/v5b-fidelity-notes.md for the deviations this rail keeps on
-    purpose: Pins, which the mock has no screen for at all, and the S3 segment above.
+    more. See TitleBar.axaml and design-notes/v5b-fidelity-notes.md for the one deviation this rail still
+    keeps on purpose: the S3 segment above. Pins was the other, and it is gone — see the remark where that
+    row used to sit, between Keys and Snips.
 
     Buttons rather than a TabStrip or a ListBox, still, for the reason the v3 remark gave: all three hold
     the selection themselves, so a click would move the highlight before the shell decided anything, and a
@@ -128,25 +129,16 @@
           </Button>
 
           <!--
-            KEPT — the mock has no screen for approved host keys at all; see the file-level remark. push_pin
-            is the same codepoint HostsScreen.axaml already draws for a host's own pin badge, reused rather
-            than picked afresh so the one concept reads as one glyph everywhere it appears.
+            ◆ NO Pins ROW. The pins screen is still here and still reached in one click — from "Host keys"
+            on the Keys screen's own header, which is where a list of approved host keys belongs: they are
+            keychain material, and that button was already the second way to reach them. Two rail rows away
+            from each other, both landing on the same screen, is a rail that has to be read twice.
 
-            ◆ U+F10D, not U+E946, which both sites drew until this pass and which no glyph in the embedded
-            face answers to: the cmap of Assets/Fonts/MaterialIcons (Material Icons 1.017, 2019) skips E944
-            and E946, so this row and the hosts screen's own pin badge were both drawing a tofu box. F10D is
-            where push_pin lives in that vintage, verified against the file rather than against a codepoints
-            table for a later release of the font.
+            It is also the last of the rail's own deviations from the mock to go. The row was kept in v5b on
+            the grounds that the design has no screen for approved host keys at all — see the file-level
+            remark — which is true of the design and was never a reason for a rail entry once the keychain
+            had a door to the same place.
           -->
-          <Button Classes="flat nav" Classes.active="{Binding IsKnownHostsShowing}"
-                  Command="{Binding ShowScreenCommand}"
-                  CommandParameter="{x:Static vm:ShellScreen.KnownHosts}"
-                  ToolTip.Tip="Host keys you have approved, and how to withdraw one">
-            <StackPanel Orientation="Horizontal" Spacing="10">
-              <TextBlock Classes="navicon" Text="&#xF10D;" />
-              <TextBlock Classes="navlabel" Text="Pins" />
-            </StackPanel>
-          </Button>
 
           <Button Classes="flat nav" Classes.active="{Binding IsSnippetsShowing}"
                   Command="{Binding ShowScreenCommand}"

src/DodoSSH.Client.Shell/ViewModels/VaultViewModel.cs

@@ -11247,10 +11247,7 @@ internal sealed partial class VaultViewModel(
         }
         catch (TimeoutException)
         {
-            Abandon(
-                attempt,
-                "The terminal did not start, so nothing was connected. The Microsoft Edge WebView2 "
-                + "runtime is probably missing or blocked; install it and try again.");
+            Abandon(attempt, RendererNeverStarted);
         }
         catch (SshHostKeyUnknownException exception)
         {
@@ -11275,6 +11272,29 @@ internal sealed partial class VaultViewModel(
         }
     }
 
+    /// <summary>What a renderer that never attached is reported as.</summary>
+    /// <remarks>
+    /// <para>
+    /// The wait is translated rather than reported for the reason <see cref="OpenSessionAsync"/> gives —
+    /// <see cref="TimeoutException"/> says only "The operation has timed out" — and the whole value of the
+    /// translation is naming where to look. Which is why it cannot be one sentence: the desktop's answer is
+    /// a runtime this application does not install, and the phone has no such runtime and no such answer.
+    /// Telling somebody on a handset to install Microsoft Edge WebView2 is worse than saying nothing, at the
+    /// one moment they are trying to work out what went wrong.
+    /// </para>
+    /// <para>
+    /// A runtime check rather than a constructor parameter, for the reason
+    /// <c>MainWindowViewModel.GestureWait</c> records at length: which renderer is behind the terminal is a
+    /// fact about the platform this assembly is running on, not about one installation of it.
+    /// </para>
+    /// </remarks>
+    private static string RendererNeverStarted =>
+        OperatingSystem.IsAndroid()
+            ? "The terminal did not start, so nothing was connected. Android's WebView is probably "
+                + "disabled or updating; check it in Settings and try again."
+            : "The terminal did not start, so nothing was connected. The Microsoft Edge WebView2 "
+                + "runtime is probably missing or blocked; install it and try again.";
+
     /// <summary>Says, in one place, that an attempt ended without a session and why.</summary>
     /// <remarks>
     /// The reason goes to two places on purpose. The status line is where somebody watching this screen is

src/DodoSSH.Client.Shell/WebAssets/terminal.js

@@ -255,8 +255,31 @@ function createSession(sessionId) {
   // WebGL where it is available. Falling back rather than failing matters because a software
   // renderer is slow but usable, whereas a blank pane is not — and remote desktops and VMs
   // routinely have no usable GPU context.
+  //
+  // ◆ THE CONTEXT-LOSS HANDLER IS THE HALF THAT WAS MISSING, AND ON A PHONE IT IS THE WHOLE THING.
+  //
+  // The addon does not recover from a lost GPU context by itself, and it does not fail loudly either:
+  // it stays loaded over a dead context and draws nothing at all. What that looks like from outside is
+  // a terminal that is connected, still accepting keystrokes, still acknowledging output — and blank.
+  // xterm's own guidance is to dispose the addon and let the DOM renderer take over, which is what this
+  // does; the addon is not reloaded afterwards, because a pane that lost the context once is on a
+  // surface that will do it again and thrashing between renderers is worse than being slow.
+  //
+  // Losing it is ordinary on Android and nearly unheard of on Windows, which is why this went unnoticed
+  // for so long. Collapsing the renderer sets the native view to GONE — see
+  // AndroidNativeControlHostImpl.HideWithSize — and a WebView with no surface has no GL context. The
+  // shell collapses it every time a tab starts connecting, every time the connect sheet opens and every
+  // time the app is backgrounded, so on a phone the first loss arrives within seconds of the first
+  // session. WebView2 hides a child HWND instead and keeps rendering throughout; see
+  // docs/platform-flags.md.
   try {
-    term.loadAddon(new WebglAddon.WebglAddon());
+    const webgl = new WebglAddon.WebglAddon();
+
+    // Subscribed before loadAddon, because loadAddon is what activates the addon and a context that is
+    // already gone can be reported from inside that call.
+    webgl.onContextLoss(() => webgl.dispose());
+
+    term.loadAddon(webgl);
   } catch (error) {
     console.warn('WebGL renderer unavailable; falling back to canvas.', error);
   }
@@ -296,10 +319,17 @@ function activate(sessionId) {
 // caller, because more than one path reaches here: a minimised window, and a splitter dragged to the edge
 // once splits land.
 //
-// It is *not* what protects the vault's lock screen, which an earlier version of this comment claimed.
-// Collapsing the host's WebView hides a native child window without resizing it, so this page's viewport
-// does not change, no observer fires and this function is never called — measured with a live shell, and
-// confirmed by removing the guard and finding the lock cycle equally clean. See docs/platform-flags.md.
+// It is *not* what protects the vault's lock screen on the desktop, which an earlier version of this
+// comment claimed. Collapsing WebView2 hides a native child window without resizing it, so this page's
+// viewport does not change, no observer fires and this function is never called — measured with a live
+// shell, and confirmed by removing the guard and finding the lock cycle equally clean. See
+// docs/platform-flags.md.
+//
+// On the phone it *is* load-bearing, and that is the one place the two heads differ here. Android hides a
+// native child by setting it GONE, and a GONE view is skipped by its parent's layout — so collapsing the
+// renderer really does take this page's viewport to nothing, the observer really does fire, and without
+// the guard every lock, every connect sheet and every trip to the background would reflow the remote pty
+// to 2x1 and mangle the scrollback it wrapped.
 const MINIMUM_FITTABLE_PIXELS = 40;
 
 function resize(session, sessionId) {

src/DodoSSH.Contracts/packages.lock.json

@@ -22,9 +22,9 @@
       },
       "Microsoft.NET.ILLink.Tasks": {
         "type": "Direct",
-        "requested": "[10.0.10, )",
-        "resolved": "10.0.10",
-        "contentHash": "f5VCIE7AJpd5YvzNTeMGVzQIgyE9tX+AreTYwQF+REbu+DZo/2Ae+jNSwhPEYrVz6RRkd7y8ubXjk6Nn6Ka+Cg=="
+        "requested": "[10.0.11, )",
+        "resolved": "10.0.11",
+        "contentHash": "IBf7lbovvjGWVWXZX5cJ/cO0WXbId0Zq4BuSeT94mGZuOAP66oMeH9PTBZ9Jpp3Jb6jtK0qm/NyUbPRo1gC/wQ=="
       },
       "MinVer": {
         "type": "Direct",

src/DodoSSH.Crypto/packages.lock.json

@@ -16,9 +16,9 @@
       },
       "Microsoft.NET.ILLink.Tasks": {
         "type": "Direct",
-        "requested": "[10.0.10, )",
-        "resolved": "10.0.10",
-        "contentHash": "f5VCIE7AJpd5YvzNTeMGVzQIgyE9tX+AreTYwQF+REbu+DZo/2Ae+jNSwhPEYrVz6RRkd7y8ubXjk6Nn6Ka+Cg=="
+        "requested": "[10.0.11, )",
+        "resolved": "10.0.11",
+        "contentHash": "IBf7lbovvjGWVWXZX5cJ/cO0WXbId0Zq4BuSeT94mGZuOAP66oMeH9PTBZ9Jpp3Jb6jtK0qm/NyUbPRo1gC/wQ=="
       },
       "MinVer": {
         "type": "Direct",

tests/DodoSSH.Client.App.Layout.Tests/ScreenLayoutTests.cs

@@ -1596,7 +1596,7 @@ public sealed class ScreenLayoutTests : IAsyncLifetime
     /// <remarks>
     /// <para>
     /// v5b's redraw changes what this test has to hold. Three button shapes live in the rail now rather
-    /// than one: the switcher's three segments, each a third of the rail's own content width; the six item
+    /// than one: the switcher's three segments, each a third of the rail's own content width; the five item
     /// rows below it and the user chip at the foot, both the rail's full content width. A single
     /// across-the-board width assertion the way the v3 version of this test made one would either be wrong
     /// for the segments or have to loosen until it caught nothing, so each shape gets its own count and its
@@ -1604,13 +1604,18 @@ public sealed class ScreenLayoutTests : IAsyncLifetime
     /// </para>
     /// <para>
     /// The rail runs vertically, so what runs out at the window's minimum is still height — a switcher plus
-    /// six rows plus a user chip have to leave room for each other in the same space the v3 rail's seven
+    /// five rows plus a user chip have to leave room for each other in the same space the v3 rail's seven
     /// plain rows did. Both counts are asserted in both directions for the reason the old test's was: an
     /// entry silently dropping off the bottom would still pass every other assertion here.
     /// </para>
+    /// <para>
+    /// Five and not six since Pins left the rail: the pins screen is reached from "Host keys" on the Keys
+    /// screen, which was always the other way in. Exact rather than a bound, so putting a row back is a
+    /// decision somebody makes here rather than something that slips in.
+    /// </para>
     /// </remarks>
     [Fact]
-    public async Task TheNavRailHoldsItsSwitcherSixDestinationsAndTheUserChipAtTheWindowsMinimum()
+    public async Task TheNavRailHoldsItsSwitcherFiveDestinationsAndTheUserChipAtTheWindowsMinimum()
     {
         await LayoutHarness.OnTheUiThreadAsync(
             () =>
@@ -1630,7 +1635,7 @@ public sealed class ScreenLayoutTests : IAsyncLifetime
 
                     segments.Count.ShouldBe(3, "SSH, SFTP and S3");
                     rows.Count.ShouldBe(
-                        6, "the mode-dependent first row, then Hosts, Keys, Pins, Snips and Logs");
+                        5, "the mode-dependent first row, then Hosts, Keys, Snips and Logs");
 
                     foreach (var segment in segments)
                     {

tests/DodoSSH.Client.Ssh.Tests/SshServerFixture.cs

@@ -1,4 +1,6 @@
+using System.Net.Sockets;
 using System.Security.Cryptography;
+using System.Text;
 using DotNet.Testcontainers.Builders;
 using DotNet.Testcontainers.Containers;
 using Xunit;
@@ -40,6 +42,21 @@ public sealed class SshServerFixture : IAsyncLifetime
 
     private const int SshPort = 2222;
 
+    /// <summary>
+    /// How many connections in a row the server has to answer before this fixture calls it ready.
+    /// </summary>
+    /// <remarks>
+    /// Twenty-five, and the number is measured rather than picked. Probing a fresh container 200 times with
+    /// penalties left at the image's default, the first <c>Not allowed at this time</c> came back at probe
+    /// 18 and 183 of the 200 were refused; with <c>PerSourcePenalties no</c> applied, none of 200 were. Ten
+    /// was tried first and is useless — it sits below the threshold, so the guard passed happily against a
+    /// server that was still penalising. See <see cref="WaitUntilServingAsync"/>.
+    /// </remarks>
+    private const int RequiredStreak = 25;
+
+    /// <summary>How long to keep trying before giving up on the server entirely.</summary>
+    private static readonly TimeSpan ReadyTimeout = TimeSpan.FromSeconds(60);
+
     private readonly SemaphoreSlim sftpGate = new(1, 1);
 
     private IContainer? container;
@@ -80,101 +97,94 @@ public sealed class SshServerFixture : IAsyncLifetime
             .Build();
 
         await container.StartAsync();
-        await AllowTcpForwardingAsync();
+        await ReconfigureAsync();
+        await WaitUntilServingAsync();
     }
 
     /// <summary>
-    /// Lets this server open the direct-tcpip channels a forward is made of.
+    /// Turns off the hardening this suite trips over, and makes the running server re-read its config.
     /// </summary>
     /// <remarks>
     /// <para>
-    /// ◆ <b>The image ships <c>AllowTcpForwarding no</c>, and nothing says so at the point it bites.</b> A
-    /// dynamic forward starts perfectly happily — it is a local listener, and opening it asks the server
-    /// nothing — and then every connection through it is refused when the channel is opened. SSH.NET
-    /// reports that as <c>SOCKS5: General failure</c> from the proxy, which names neither the server nor
-    /// the setting, and is what the first run of <c>LoopbackProxyTests</c> collected.
+    /// ◆ <b><c>PerSourcePenalties no</c> is the fix for the flake this suite had for months, and the other
+    /// two settings here are not.</b> OpenSSH 9.8 added per-source penalties and 10.x has them on by
+    /// default; this image runs 10.3. A source address that keeps disconnecting without authenticating is
+    /// penalised, and while the penalty holds every connection from it is answered with the clear-text line
+    /// <c>Not allowed at this time</c> and then closed.
     /// </para>
     /// <para>
-    /// Patched after start rather than baked in, because the image's entrypoint writes its configuration
-    /// itself on every boot — a mounted file would be overwritten before sshd read it. sshd re-reads on
-    /// <c>SIGHUP</c> and applies the result to connections made after that, and the readiness wait has
-    /// already run, so nothing here races the boot.
+    /// <b>This suite generates exactly that traffic, by design.</b> This client's first contact with an
+    /// unknown host is a connection deliberately refused at the host key — which is a disconnect with no
+    /// authentication attempt — and several tests do nothing else:
+    /// <c>RefusingTheHostKey_AbortsTheConnection</c>, <c>AnUntrustedHost_IsRefusedExactlyAsAShellWouldBe</c>,
+    /// and every helper that learns a host key by being turned away first. Enough of them close together and
+    /// sshd stops talking to the test host altogether, for a while, and then starts again.
+    /// </para>
+    /// <para>
+    /// From the client that is <c>SshConnectionException: The connection was closed by the remote host</c>
+    /// within milliseconds — no banner, nothing to say which of the many reasons it was. It hits whichever
+    /// class is running when the penalty lands and spares the rest, which is why it read as random and why
+    /// the class it hit lost <em>every</em> connection it made rather than a random few. The one test in that
+    /// class that expects a refusal passed throughout, for the wrong reason.
+    /// </para>
+    /// <para>
+    /// ◆ <b>Two earlier diagnoses were wrong, and are recorded here so they are not tried again.</b>
+    /// <c>MaxStartups</c> was blamed on the reasoning that xUnit runs test classes in parallel, so ten
+    /// unauthenticated connections would be in flight at once — but every class that touches this server
+    /// shares <see cref="SshCollection"/>, and xUnit's unit of parallelism is the collection, so they run one
+    /// after another and never have more than a connection or two open. The reload window was blamed next,
+    /// and a wait for the banner to answer was written and removed as unproven; it was unproven because the
+    /// banner does answer, right up until the penalty lands.
+    /// </para>
+    /// <para>
+    /// The line is appended rather than replaced in place, unlike the two below it, because the image's
+    /// config does not mention the keyword at all — there is no line to replace, and sshd takes the first
+    /// value it finds for a keyword that appears more than once.
+    /// </para>
+    /// <para>
+    /// ◆ <b><c>AllowTcpForwarding</c> is what a dynamic forward needs</b>, and the image ships it off as
+    /// hardening. Without it a forward opens perfectly happily — a local listener asks the server nothing —
+    /// and then every connection through it is refused when the channel is opened. SSH.NET reports that as
+    /// <c>SOCKS5: General failure</c>, which names neither the server nor the setting, and is what the first
+    /// run of <c>LoopbackProxyTests</c> collected. That suite is also the alarm if this method ever silently
+    /// stops working.
+    /// </para>
+    /// <para>
+    /// <c>MaxStartups</c> is raised for the reason it should have been in the first place rather than as a
+    /// fix for anything: the compiled-in default refuses connections at random past ten unauthenticated ones
+    /// in flight, and a throttle is hardening a test server has no business reproducing. It is kept, not
+    /// because it was ever shown to matter here, but because removing it would be a second change riding
+    /// along with this one.
+    /// </para>
+    /// <para>
+    /// Both are replaced in place rather than appended, because sshd_config takes the <em>first</em> value
+    /// it finds for a keyword: an appended line would be dead the day the image ships an uncommented one of
+    /// its own.
     /// </para>
     /// <para>
     /// ◆ <b><c>/config/sshd/sshd_config</c>, and there are two.</b> The image also carries
-    /// <c>/etc/ssh/sshd_config</c>, which looks like the file to patch, reads identically, and is not the
-    /// one the running server was started with — patching it changes the text and nothing else, which is a
-    /// fix that appears to work and leaves the failure exactly where it was. Measured with <c>find</c>
-    /// rather than assumed, after the first version of this method did precisely that.
+    /// <c>/etc/ssh/sshd_config</c>, which looks like the file to patch, reads identically, and is not the one
+    /// the running server was started with — patching it changes the text and nothing else, which is a fix
+    /// that appears to work and leaves the failure exactly where it was.
     /// </para>
     /// <para>
-    /// It is on for the whole assembly rather than for the one test that needs it. Forwarding is off in
-    /// this image as hardening, not as a behaviour worth reproducing: nothing else here opens a channel of
-    /// any kind, so allowing it changes what exactly one suite can do and what none of the others see.
-    /// </para>
-    /// <para>
-    /// ◆ <b><c>MaxStartups</c> is raised here too, against a flake this suite has and that this change is
-    /// a mitigation for rather than a proven cure.</b> The distinction is stated because the evidence
-    /// stops short of the claim, and a later reader deserves to know which.
-    /// </para>
-    /// <para>
-    /// What is established: sshd's compiled-in default is <c>10:30:100</c> — past ten
-    /// <em>unauthenticated</em> connections in flight it refuses new ones at random, thirty percent of the
-    /// time, rising to always at a hundred — and the image ships the line commented out, so that default
-    /// was what ran. xUnit runs test classes in parallel and most classes here open a connection, so ten
-    /// in flight is reachable in the opening seconds. A refused connection presents to the client as
-    /// <c>SshConnectionException: The connection was closed by the remote host</c> within tens of
-    /// milliseconds, on whichever test connects at the wrong moment — which is exactly the observed
-    /// failure, seen in CI and reproduced locally.
-    /// </para>
-    /// <para>
-    /// What is <em>not</em> established is that this limit is the only cause, because the flake rate could
-    /// not be measured reliably. On the development machine the identical unmodified suite ran 85/85 clean
-    /// and, an hour later, failed 13 runs out of 15 — Docker throughput on that host swings far enough to
-    /// swamp the effect being measured. Any before/after comparison taken there is noise, and two were,
-    /// before that was noticed.
-    /// </para>
-    /// <para>
-    /// It is committed anyway, on the narrower argument that it is right regardless: a connection throttle
-    /// is hardening this suite has no interest in reproducing. It exists to test an SSH client, not to
-    /// survive a rate limit, and a test server that drops connections at random is a bad test server
-    /// whether or not it is the cause of this particular flake.
-    /// </para>
-    /// <para>
-    /// <b>Not fixed by serialising the suite</b>, which would have hidden it and cost the parallelism, and
-    /// not by retrying the connect, which would have made the client's own reconnect behaviour untestable
-    /// by burying it in the fixture. The limit is a property of a hardened server that this suite has no
-    /// interest in reproducing — it exists to test an SSH client, not to survive a throttle.
-    /// </para>
-    /// <para>
-    /// Replaced in place rather than appended, because sshd_config takes the <em>first</em> value it finds
-    /// for a keyword: an appended line would be dead the day the image ships an uncommented one of its own.
-    /// </para>
-    /// <para>
-    /// ◆ <b>The reload window is the other candidate, and it is deliberately not guarded against.</b>
-    /// <c>SIGHUP</c> makes sshd close its listeners and re-execute itself, and <c>pkill</c> returns when
-    /// the signal is delivered rather than when that has finished — so in principle a connection made
-    /// immediately afterwards is refused, producing this same exception. A wait that opened connections
-    /// until the server answered with its banner three times running was written, and then removed: it
-    /// could not be shown to change anything either, and a fixture carrying two unproven fixes for one
-    /// symptom is worse than one, because the next person has to disprove both.
-    /// </para>
-    /// <para>
-    /// If this flake returns, that is the next thing to try. Two things to know before trying it: the two
-    /// causes are indistinguishable from the client, so a fix can only be judged by a repeat run and never
-    /// by whether the next run passes — and the repeat run has to happen somewhere with stable Docker
-    /// throughput, which the development machine is not. Better still, make sshd say why: raise its
-    /// <c>LogLevel</c> here, disable Ryuk so the container outlives the run, and read
-    /// <c>docker logs</c>. A <c>MaxStartups</c> refusal names itself there; a reload does not.
+    /// ◆ <b>Patched after boot and reloaded, rather than injected before it — which was tried and does not
+    /// work.</b> This image family runs <c>/custom-cont-init.d</c> scripts, which look like the right hook
+    /// and are not: the container's own log puts <c>sshd is listening on port 2222</c> <em>before</em>
+    /// <c>[custom-init] Files found, executing</c>, so a script there edits a file the running server has
+    /// already read. It leaves a config that greps correctly and a server behaving as though it had never
+    /// been touched — the same trap as the wrong file, one layer up. Measured from the log, after a version
+    /// of this fixture did exactly that and failed twenty-eight tests.
     /// </para>
     /// </remarks>
-    private async Task AllowTcpForwardingAsync()
+    private async Task ReconfigureAsync()
     {
         var result = await container!.ExecAsync([
             "sh",
             "-c",
             "sed -i 's/^AllowTcpForwarding no/AllowTcpForwarding yes/' /config/sshd/sshd_config"
                 + " && sed -i 's/^#*MaxStartups .*/MaxStartups 200/' /config/sshd/sshd_config"
+                + " && printf '\\nPerSourcePenalties no\\n' >> /config/sshd/sshd_config"
                 + " && pkill -HUP sshd",
         ]);
 
@@ -183,7 +193,107 @@ public sealed class SshServerFixture : IAsyncLifetime
             throw new InvalidOperationException(
                 $"Could not reconfigure the test server: {result.Stderr}");
         }
+    }
 
+    /// <summary>
+    /// Blocks until the server answers <see cref="RequiredStreak"/> connections in a row with its banner.
+    /// </summary>
+    /// <remarks>
+    /// <para>
+    /// ◆ <b>This is a guard rather than a wait, and what it guards against is
+    /// <c>PerSourcePenalties</c> coming back.</b> Reconfiguring above turns it off; this proves it is off,
+    /// immediately and by name, instead of letting the suite discover it later as an unrelated-looking
+    /// failure in whichever class happened to be running.
+    /// </para>
+    /// <para>
+    /// <b>Consecutive, and deliberately with no pause between them.</b> Each probe opens a connection, reads
+    /// the identification string and disconnects without authenticating — which is exactly the shape of
+    /// connection <c>PerSourcePenalties</c> punishes, and exactly what this suite does all day: a first
+    /// contact with an unknown host is a connection this client deliberately refuses at the host key.
+    /// <see cref="RequiredStreak"/> back to back is therefore not a soak test, it is the specific
+    /// provocation, sized above the measured threshold on purpose, and it costs well under a second when the
+    /// setting is off.
+    /// </para>
+    /// <para>
+    /// It is also the one check that can tell a listening socket from a running server. The container's own
+    /// readiness — a log line and <c>netstat</c> showing <c>:2222</c> — passes on a container whose sshd has
+    /// gone: the socket is published by a host-side proxy that accepts before it has anything to forward to,
+    /// so a dead server presents as a connection accepted and closed rather than as one refused.
+    /// </para>
+    /// <para>
+    /// Probed from the host rather than with <c>docker exec</c>, deliberately: that is the path the tests
+    /// take, proxy included, and penalties are counted per source address — from inside the container the
+    /// source would be the loopback rather than the address every test connects from.
+    /// </para>
+    /// </remarks>
+    private async Task WaitUntilServingAsync()
+    {
+        // TimeProvider.System rather than DateTimeOffset.UtcNow, which this repository bans so that time can
+        // be faked — and rather than a fake, because what is being waited on is a real container starting.
+        var deadline = TimeProvider.System.GetUtcNow() + ReadyTimeout;
+        var streak = 0;
+        var last = "no probe ran";
+
+        while (streak < RequiredStreak)
+        {
+            if (TimeProvider.System.GetUtcNow() >= deadline)
+            {
+                throw new InvalidOperationException(
+                    $"The test server did not answer {RequiredStreak} connections in a row within "
+                    + $"{ReadyTimeout}. The last probe said: {last}. If it says \"Not allowed at this "
+                    + "time\", sshd is penalising this source address and PerSourcePenalties is no longer "
+                    + "being turned off — see ReconfigureAsync.");
+            }
+
+            var (answered, what) = await ProbeAsync();
+            last = what;
+
+            if (answered)
+            {
+                streak++;
+                continue;
+            }
+
+            // Only pause when it is not working. Back-to-back probes are the point while they succeed;
+            // hammering a server that has not finished starting is just noise.
+            streak = 0;
+            await Task.Delay(TimeSpan.FromMilliseconds(200));
+        }
+    }
+
+    /// <summary>Opens a socket and reads far enough to see OpenSSH's identification string.</summary>
+    /// <remarks>
+    /// The description comes back with the answer because the interesting failures are not exceptions. A
+    /// penalised source is told <c>Not allowed at this time</c> in clear text before the socket closes, and
+    /// a suite that only knew "no banner" would have to go and find that out again — which is what happened
+    /// the first time, at some length.
+    /// </remarks>
+    private async Task<(bool Answered, string What)> ProbeAsync()
+    {
+        try
+        {
+            using var probe = new TcpClient();
+            using var timeout = new CancellationTokenSource(TimeSpan.FromSeconds(5));
+
+            await probe.ConnectAsync(Host, Port, timeout.Token);
+
+            var buffer = new byte[64];
+            var read = await probe.GetStream().ReadAtLeastAsync(
+                buffer, 4, throwOnEndOfStream: false, timeout.Token);
+
+            var answered = read >= 4 && "SSH-"u8.SequenceEqual(buffer.AsSpan(0, 4));
+
+            return (
+                answered,
+                answered
+                    ? "SSH-"
+                    : $"{read} bytes: "
+                        + Encoding.ASCII.GetString(buffer, 0, Math.Max(read, 0)).ReplaceLineEndings(" "));
+        }
+        catch (Exception exception) when (exception is SocketException or OperationCanceledException or IOException)
+        {
+            return (false, $"{exception.GetType().Name}: {exception.Message}");
+        }
     }
 
     /// <summary>
@@ -225,12 +335,17 @@ public sealed class SshServerFixture : IAsyncLifetime
     /// </summary>
     /// <remarks>
     /// <para>
-    /// Shared rather than opened per test, and that is a limit of the server rather than an optimisation.
-    /// sshd's <c>MaxStartups</c> drops connections at random once enough are part-way through a handshake,
-    /// and this client's first contact with an unknown host is a connection deliberately <em>refused</em> at
-    /// the host key — so a suite that opened its own session per test made two handshakes per test and
-    /// pushed the whole assembly over the threshold. What that looks like is unrelated tests failing with
-    /// "the connection was closed by the remote host", a different few each run.
+    /// Shared rather than opened per test. This was once explained as a way of staying under the server's
+    /// <c>MaxStartups</c> throttle, on the belief that the suite ran its classes in parallel and made two
+    /// handshakes per test — this client's first contact with an unknown host is a connection deliberately
+    /// <em>refused</em> at the host key, so every session costs two. The parallelism was not real: every
+    /// class here shares one collection and xUnit runs collections, not classes, in parallel. See
+    /// <see cref="WaitUntilServingAsync"/>, which is where that mistake was found and what the failure it
+    /// was blamed for turned out to be.
+    /// </para>
+    /// <para>
+    /// It stays shared regardless, on the plainer argument: one session is enough, and a handshake per test
+    /// would be seconds of the suite's runtime spent proving nothing this file has not already proved.
     /// </para>
     /// <para>
     /// Safe to share because an SFTP session holds no per-test state: every test here works in a directory