From 573f5d5668ab5d3867294b487dd4c8b80539ca7a Mon Sep 17 00:00:00 2001 From: Jaap-Jan de Wit | DodoTech Date: Thu, 30 Jul 2026 15:17:30 +0200 Subject: [PATCH] Keep the device key in the TPM, behind a consent Windows enforces MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The last of ADR 0007's three pieces, and it does not implement what that ADR originally decided — because writing it exposed a flaw in the decision. The ADR said "a Windows Hello gesture gating a protected blob". That does not deliver what the rest of the document claims for it: a gate inside the process is not a gate. A store that showed a prompt and then read a DPAPI blob would be bypassed by malware that skipped the prompt, read the file and called CryptUnprotectData itself — which is exactly the attacker the whole decision was made against, and exactly the reason DPAPI alone was rejected. The presence requirement has to be a condition of using the key, enforced below the application, or it is decoration. So the device key is encrypted to an RSA key created in the Microsoft Platform Crypto Provider — the TPM — under CngUIProtectionLevels.ProtectKey. Windows requires consent to use that key, so the prompt is not something this code can be talked out of showing. Malware can ask for the key; it cannot answer the dialog. That is strictly stronger than the ADR described, and most of what option D was being saved for: the wrapping key genuinely never leaves hardware. The X25519 device key still lands in memory to open the wrap, because DSH1 fixes that wrap at a curve the TPM cannot do — the remaining gap, and now a smaller step than it was. CngKey is in-box, so this needed no WinRT projection and no Windows target framework. Which is worth stating plainly because the opposite was planned: the piece was scoped as "where the Windows TFM lands", and it turned out a platform guard on one class was enough. Client.App and its two test projects stay on net10.0. Two things were measured on real hardware rather than assumed, and the second changed the shape of the work. The platform provider works here and holds an RSA key — confirmed by creating and deleting one before writing anything that depended on it. And ProtectKey prompts at key *creation*, not only at use. The comment in the first draft of this file said the opposite, with a confident explanation: sealing uses only the public half, so it should be silent. It is not. CngKey.Create blocks on a dialog, because the policy means "protect this key with a PIN" and Windows asks the user to set that up there and then. Found by writing tests around save and forget and watching the suite hang for ten minutes waiting for somebody to type one. That has two consequences worth knowing before touching this file. SaveAsync is user-facing code — it belongs on a UI thread, behind a button somebody pressed, never on a background pass. And almost nothing in the store can be covered automatically: two tests remain, availability and the empty-blob case, both of which provably reach no dialog. Disabling the UI policy to make the rest testable would remove the one property worth having. The interface offers two things and hides both where they cannot work. "Use Windows Hello" appears on the unlock screen only when this machine has a cached wrap and a keystore still willing to release the key; "Use Windows Hello here" appears in the account bar only when the machine can keep a key and has not already registered one, so it is spent once used. Absent rather than disabled, in both cases: a greyed-out button on a machine that never had a TPM reads as something broken, and the passphrase box beside it is not a fallback — it is the ordinary way in. Both unlock paths now share AdoptAsync rather than each opening the known-host store, building the vault and starting auto-sync. The ordering in there is load-bearing and a second copy would be a second chance to get it wrong. The shell's tests drive a fake keystore. Not for speed: the real one prompts on every save and load, so a suite using it would block forever. What the shell has to get right is which buttons appear and what happens when one is pressed, and a fake answers exactly that. It is shared from Client.Session.Tests by source link rather than reimplemented. 882 tests green, 6 of them new. Zero warnings, dotnet format clean. Not verified, and not verifiable here: the dialogs. Whether the consent prompt appears at the right moments, reads sensibly, and returns to a usable window when declined needs the application run by a person on a machine with a TPM. That is the remaining half of outstanding item #7, and it is now the only thing between this feature and being finished. --- docs/adr/0007-device-key-protection.md | 66 ++++- src/DodoSSH.Client.App/App.axaml.cs | 23 +- .../ViewModels/MainWindowViewModel.cs | 159 ++++++++++-- .../ViewModels/VaultViewModel.cs | 10 + src/DodoSSH.Client.App/Views/MainWindow.axaml | 26 +- src/DodoSSH.Client.Session/ClientPaths.cs | 11 + .../WindowsDeviceKeyStore.cs | 229 ++++++++++++++++++ .../DodoSSH.Client.App.Tests.csproj | 11 + .../ShellFlowTests.cs | 83 +++++++ .../DeviceUnlockTests.cs | 46 ---- .../FakeDeviceKeyStore.cs | 48 ++++ .../WindowsDeviceKeyStoreTests.cs | 75 ++++++ 12 files changed, 699 insertions(+), 88 deletions(-) create mode 100644 src/DodoSSH.Client.Session/WindowsDeviceKeyStore.cs create mode 100644 tests/DodoSSH.Client.Session.Tests/FakeDeviceKeyStore.cs create mode 100644 tests/DodoSSH.Client.Session.Tests/WindowsDeviceKeyStoreTests.cs diff --git a/docs/adr/0007-device-key-protection.md b/docs/adr/0007-device-key-protection.md index 1a89f6e..b8f38f5 100644 --- a/docs/adr/0007-device-key-protection.md +++ b/docs/adr/0007-device-key-protection.md @@ -33,11 +33,47 @@ leaves hardware" would be false under all of them. ## Decision -**A Windows Hello gesture gating a protected blob, with the passphrase kept as a permanent fallback.** +**A TPM-resident key whose use requires the user's consent, with the passphrase kept as a permanent +fallback.** -The gesture is what carries the security value: it requires **user presence** per unlock. Hello cannot -decrypt, so it is used to gate release of the wrapping key, and the passphrase path remains available -unconditionally. +User presence per unlock is what carries the security value. What changed between this decision and its +implementation is *who enforces the presence*, and the change was a correction rather than a refinement. + +> **Amended 2026-07-30.** This section originally read "a Windows Hello gesture gating a protected blob". +> That design does not deliver what the rest of this document claims for it, and the flaw is worth keeping +> on the record: **a gate inside the process is not a gate.** A store that showed a Hello prompt and then +> read a DPAPI blob would be bypassed by malware that skipped the prompt, read the file and called +> `CryptUnprotectData` itself. The presence requirement has to be a condition of *using the key*, enforced +> below the application, or it is decoration. + +So the device key is encrypted to an RSA key created in the **Microsoft Platform Crypto Provider** — the +TPM — under `CngUIProtectionLevels.ProtectKey`. Windows requires consent to use that key, so the prompt is +not something this code can be talked out of showing. Malware can ask for the key; it cannot answer the +dialog, and the attempt is visible. `System.Security.Cryptography.CngKey` is in-box, so this needs no WinRT +projection and **no Windows target framework** — a plain platform guard is enough. + +RSA rather than an agreement algorithm because the payload is 32 bytes and OAEP over 2048 bits carries 190. +That also keeps the DSH1 device wrap unchanged at X25519: the TPM key protects the device key, it does not +replace it. + +Availability is probed by creating a throwaway key and deleting it, not by asking whether the provider is +registered — it is registered on machines with no usable TPM too, and reports itself present right up to +the point where creating a key fails. + +### What was measured, and what it cost + +Two things were verified on real hardware rather than assumed, and one of them changed the design's shape: + +- **The platform provider works** and holds an RSA key: confirmed by creating and deleting one. +- **`ProtectKey` prompts at key *creation*, not only at use.** `CngKey.Create` blocks on a dialog, because + the policy means "protect this key with a PIN" and Windows asks the user to set that up there and then. + +The second has consequences. Registering a device shows a setup dialog and every unlock shows a consent +dialog, which is the right shape for an opt-in feature — but it means **`SaveAsync` is user-facing code** +that belongs on a UI thread behind a button somebody pressed, and it means almost nothing in the store can +be covered by an automated test. That was found by writing those tests and watching a suite hang for ten +minutes waiting for a PIN. Two tests remain: availability, and the empty case that provably reaches no +dialog. ### Why not DPAPI alone @@ -55,9 +91,14 @@ Neither defends the *local malware* case. The gesture does. ### Why not extend the spec (yet) -The only option that delivers what the TPM is usually credited with is to add a `SealTo` algorithm over -a curve the TPM can do — `alg_id = 4` over P-256 — so the device private key never exists in process -memory at all. That is **the recorded target**, not this decision. +The device *wrapping* key now genuinely never leaves the TPM, which is most of what option D promised. What +remains is that the X25519 device key itself is reassembled in process memory to open the wrap, because DSH1 +fixes that wrap at a curve the TPM cannot do. + +Closing that last gap means adding a `SealTo` algorithm over a curve the TPM can do — `alg_id = 4` over +P-256 — so the device key never exists outside hardware at all. That is **the recorded target**, not this +decision, and it is now a smaller step than it was: the keystore plumbing, the endpoint and the unlock path +would all be unchanged. It is cheaper than "change a frozen spec" sounds, because a device wrap row is read only by the device that created it: not by another client, and not by the server. The envelope already carries `alg_id` @@ -97,11 +138,12 @@ would have become false under DPAPI alone. A gesture is still something the atta ### Operational -- **Hello is not always available.** No biometric hardware falls back to a Hello PIN, which is - TPM-bound and rate-limited and still satisfies the presence requirement. Some machines have no Hello - at all. The passphrase path is therefore required, not a nicety. -- **Hello keys are invalidated when the PIN is reset**, so the blob must be treated as losable at any - time; losing it degrades to a passphrase prompt and never to a locked-out vault. +- **A TPM is not always there.** A machine without one gets a store that reports itself unavailable, so + unlock keeps asking for the passphrase and neither affordance appears in the interface. The passphrase path + is therefore required, not a nicety. +- **The stored key must be treated as losable at any time** — a reset PIN, a cleared TPM, a replaced key. + Every loss degrades to a passphrase prompt and never to a locked-out vault, which is why every failure in + the store returns null rather than throwing and why the three unlock statuses all end in the same advice. - **Registering a device is a separate act from enrolling one.** `EnrollmentService.AddDevice` runs only during enrollment, so every already-enrolled account — which is all of them — needs an endpoint to add a device wrap while unlocked. Producing the wrap requires the bundle, so the client proves possession diff --git a/src/DodoSSH.Client.App/App.axaml.cs b/src/DodoSSH.Client.App/App.axaml.cs index 425a2ac..1b8496a 100644 --- a/src/DodoSSH.Client.App/App.axaml.cs +++ b/src/DodoSSH.Client.App/App.axaml.cs @@ -67,11 +67,17 @@ internal sealed partial class DodoSshApp : Application var browser = new SystemBrowserLauncher(); + // Chosen once, here, because it is a property of the machine and not of any session. A computer with + // a usable TPM gets the store that keeps a device key behind a Windows consent prompt; anything else + // gets one that reports itself unavailable, so unlock keeps asking for the passphrase. See ADR 0007. + var deviceKeys = DeviceKeyStores.ForThisMachine(paths); + var viewModel = new MainWindowViewModel( paths, caches, workspace, knownHosts, + deviceKeys, async (url, cancellationToken) => await ServerConnection .SignInAsync(url, browser, TimeProvider.System, cancellationToken) .ConfigureAwait(false), @@ -84,11 +90,22 @@ internal sealed partial class DodoSshApp : Application // discarding the task here is safe rather than merely convenient. _ = viewModel.StartAsync(CancellationToken.None); + WireShutdown(desktop, viewModel, workspace, caches); + } + + /// + /// Shutdown is deferred rather than blocked on. Sessions hold SSH connections and a listening socket, and + /// blocking the UI thread on their disposal is how an application comes to take several seconds to close — + /// or deadlocks, if any of that disposal needs the UI thread. + /// + private static void WireShutdown( + IClassicDesktopStyleApplicationLifetime desktop, + MainWindowViewModel viewModel, + TerminalWorkspace workspace, + ClientCacheFactory caches) + { var shuttingDown = false; - // Shutdown is deferred rather than blocked on. Sessions hold SSH connections and a listening - // socket, and blocking the UI thread on their disposal is how an application comes to take several - // seconds to close — or deadlocks, if any of that disposal needs the UI thread. desktop.ShutdownRequested += async (_, e) => { if (shuttingDown) diff --git a/src/DodoSSH.Client.App/ViewModels/MainWindowViewModel.cs b/src/DodoSSH.Client.App/ViewModels/MainWindowViewModel.cs index ae549c8..a448c77 100644 --- a/src/DodoSSH.Client.App/ViewModels/MainWindowViewModel.cs +++ b/src/DodoSSH.Client.App/ViewModels/MainWindowViewModel.cs @@ -68,6 +68,13 @@ internal sealed partial class MainWindowViewModel : ObservableObject, IAsyncDisp /// private readonly VaultKnownHostStore knownHosts; + /// + /// Whatever this machine can keep a device key in, chosen once at composition. An interface because the + /// answer is a platform decision — see ADR 0007 — and because a machine with no TPM gets a store that + /// reports itself unavailable rather than a null this state machine would have to check for. + /// + private readonly IDeviceKeyStore deviceKeys; + private readonly SignInHandler signIn; private readonly TimeProvider clock; private readonly Argon2Profile? passphraseProfile; @@ -91,6 +98,7 @@ internal sealed partial class MainWindowViewModel : ObservableObject, IAsyncDisp ClientCacheFactory caches, TerminalWorkspace workspace, VaultKnownHostStore knownHosts, + IDeviceKeyStore deviceKeys, SignInHandler signIn, TimeProvider clock, Argon2Profile? passphraseProfile = null) @@ -99,6 +107,7 @@ internal sealed partial class MainWindowViewModel : ObservableObject, IAsyncDisp this.caches = caches; this.workspace = workspace; this.knownHosts = knownHosts; + this.deviceKeys = deviceKeys; this.signIn = signIn; this.clock = clock; this.passphraseProfile = passphraseProfile; @@ -113,6 +122,14 @@ internal sealed partial class MainWindowViewModel : ObservableObject, IAsyncDisp [ObservableProperty] private bool isBusy; + /// Whether the unlock screen should offer a gesture instead of the passphrase. + [ObservableProperty] + private bool canUnlockWithDevice; + + /// Whether an unlocked vault should offer to register this machine. + [ObservableProperty] + private bool canRegisterDevice; + /// /// The address dotnet run --project src/DodoSSH.Api actually serves, so the first launch after /// a clone works without the user having to know a port. This was https://localhost:7217, which @@ -219,6 +236,13 @@ internal sealed partial class MainWindowViewModel : ObservableObject, IAsyncDisp ServerUrl = profile.ServerUrl; State = ShellState.Locked; StatusMessage = $"Enrolled against {profile.ServerUrl}."; + + // Both halves have to hold: a wrap in the cache, and a machine still willing to hand the key + // back. Offering the button without the second would prompt for a key that is not there; without + // the first it would prompt for a wrap that is not there. Neither failure is one a user could + // make sense of, so the button simply does not appear. + CanUnlockWithDevice = profile.DeviceWrappedPrivateKey is not null + && await deviceKeys.IsAvailableAsync(cancellationToken).ConfigureAwait(true); } catch (Exception exception) when (exception is not OperationCanceledException) { @@ -362,34 +386,119 @@ internal sealed partial class MainWindowViewModel : ObservableObject, IAsyncDisp Passphrase = string.Empty; - // Before the vault view model, so the first connection after an unlock already knows which - // host keys this user has approved. Reading them is one listing; doing it here rather than - // lazily is what keeps it off the SSH handshake thread. - try - { - await knownHosts.OpenAsync(outcome.Session!, cancellationToken).ConfigureAwait(true); - } - catch - { - // Nothing owns the session yet, so nothing else would ever dispose it — and an - // undisposed session is vault keys left in memory for the life of the process, which is - // precisely what unlocking must be able to undo. - await outcome.Session!.DisposeAsync().ConfigureAwait(true); - throw; - } - - Vault = new VaultViewModel(outcome.Session!, workspace, knownHosts, () => connection); - State = ShellState.Unlocked; - - await Vault.LoadAsync(cancellationToken).ConfigureAwait(true); - - // After the first load, so the list is on screen before anything talks to a server. The - // loop is started from the UI thread deliberately: every pass resumes here, which is what - // keeps the observable collections single-threaded. - Vault.StartAutoSync(); + await AdoptAsync(outcome.Session!, cancellationToken).ConfigureAwait(true); }).ConfigureAwait(true); } + /// Opens the vault with this machine's device key instead of the passphrase. + /// + /// No Task.Run, unlike the passphrase path: there is no Argon2 to pay for here, and the work that + /// does block is a Windows consent dialog which belongs on the UI thread anyway. + /// + [RelayCommand] + private async Task UnlockWithDeviceAsync(CancellationToken cancellationToken) + { + await RunAsync( + "Waiting for Windows…", + async () => + { + var outcome = await Opener() + .UnlockWithDeviceAsync(deviceKeys, cancellationToken) + .ConfigureAwait(true); + + StatusMessage = outcome.Message; + + if (!outcome.IsUnlocked) + { + // A declined gesture leaves the passphrase box exactly where it was, which is the whole + // fallback: the user types instead. Nothing about the screen changes but the message. + return; + } + + await AdoptAsync(outcome.Session!, cancellationToken).ConfigureAwait(true); + }).ConfigureAwait(true); + } + + /// + /// Registers this machine so a later launch can unlock with a gesture. + /// + /// + /// Needs a network, because the wrap has to reach the server — a wrap that exists only here would be + /// lost with the cache file and could never be revoked. Needs an unlocked vault too, because only an + /// open session can seal the bundle. + /// + [RelayCommand] + private async Task RegisterDeviceAsync(CancellationToken cancellationToken) + { + if (Vault is not { } vault || connection is null) + { + StatusMessage = "Sign in first: registering this machine has to reach the server."; + return; + } + + await RunAsync( + "Waiting for Windows…", + async () => + { + var name = Environment.MachineName; + + var registered = await vault.Session + .RegisterDeviceAsync(connection.Account, deviceKeys, name, cancellationToken) + .ConfigureAwait(true); + + if (!registered) + { + StatusMessage = "This machine has nowhere to keep a device key."; + return; + } + + CanRegisterDevice = false; + StatusMessage = $"'{name}' can now unlock without your passphrase."; + }).ConfigureAwait(true); + } + + /// + /// Takes ownership of a freshly opened session, whichever door opened it. + /// + /// + /// Shared by both unlock paths rather than duplicated, because the ordering in here is load-bearing and + /// a second copy would be a second chance to get it wrong. + /// + private async Task AdoptAsync(VaultSession session, CancellationToken cancellationToken) + { + // Before the vault view model, so the first connection after an unlock already knows which host keys + // this user has approved. Reading them is one listing; doing it here rather than lazily is what + // keeps it off the SSH handshake thread. + try + { + await knownHosts.OpenAsync(session, cancellationToken).ConfigureAwait(true); + } + catch + { + // Nothing owns the session yet, so nothing else would ever dispose it — and an undisposed + // session is vault keys left in memory for the life of the process, which is precisely what + // unlocking must be able to undo. + await session.DisposeAsync().ConfigureAwait(true); + throw; + } + + Vault = new VaultViewModel(session, workspace, knownHosts, () => connection); + State = ShellState.Unlocked; + + // Offered only where it can actually be honoured: a machine that can keep a key, and a profile that + // has not already registered one. Asked once here rather than recomputed, because the answer + // involves a TPM probe. + CanRegisterDevice = session.Profile.DeviceWrappedPrivateKey is null + && await deviceKeys.IsAvailableAsync(cancellationToken).ConfigureAwait(true); + + await Vault.LoadAsync(cancellationToken).ConfigureAwait(true); + + // After the first load, so the list is on screen before anything talks to a server. The loop is + // started from the UI thread deliberately: every pass resumes here, which is what keeps the + // observable collections single-threaded. + Vault.StartAutoSync(); + } + /// /// Closes the vault and forgets every key it held. Open shells keep running. /// diff --git a/src/DodoSSH.Client.App/ViewModels/VaultViewModel.cs b/src/DodoSSH.Client.App/ViewModels/VaultViewModel.cs index 2c96901..4e24d03 100644 --- a/src/DodoSSH.Client.App/ViewModels/VaultViewModel.cs +++ b/src/DodoSSH.Client.App/ViewModels/VaultViewModel.cs @@ -208,6 +208,16 @@ internal sealed partial class VaultViewModel( private Task? autoSyncLoop; private bool disposed; + /// + /// The open vault this view model is showing. + /// + /// + /// Exposed for the operations only an open session can perform — sealing the bundle to a new device key, + /// chiefly — which the shell drives rather than this view model. Ownership does not move: this type + /// disposes it, and a caller must not. + /// + internal VaultSession Session => session; + /// The hosts to show, unpushed local state included. internal ObservableCollection Hosts { get; } = []; diff --git a/src/DodoSSH.Client.App/Views/MainWindow.axaml b/src/DodoSSH.Client.App/Views/MainWindow.axaml index 40a4941..b6dd128 100644 --- a/src/DodoSSH.Client.App/Views/MainWindow.axaml +++ b/src/DodoSSH.Client.App/Views/MainWindow.axaml @@ -50,6 +50,16 @@ IsVisible="{Binding !IsOnline}" />