* feat(shared): steer capability gates and live steered signal schemas - STEERING_SUPPORTED_FLAVORS / isSteeringSupportedForSession gate which agents can deliver queued messages into the active turn (pi, codex, cursor ACP; legacy stream-json cursor excluded) - AgentState.steeringActive, DecryptedMessage.steered and messages-consumed live signal (never persisted by the hub) * feat(cli): queue reservations and steered messages-consumed option - MessageQueue2 gains takeByLocalId/restoreReservation/ beginReservationDispatch/commitReservation so an async steer can reserve a queued row without racing the main loop's turn/start drain - emitMessagesConsumed accepts steered: true to mark mid-turn delivery * feat(codex): mid-turn steer via app-server turn/steer (#888) - CodexAppServerClient.steerTurn + TurnSteerParams/Response types - CodexRemoteLauncher registers the steer-queued-message RPC handler: reserves the queued row, validates it against the active turn (no control commands, matching mode hash), injects via turn/steer with an epoch guard that invalidates in-flight steers on abort/cleanup - steeringActive agent state tracks the active-turn window - hub syncEngine gate opens to codex; messages-consumed relays steered * feat(web): Steered badge and steer gating for codex sessions - HappyUserMessage shows a ↳ Steered badge fed by the live messages-consumed steered signal, preserved across server echoes and refetches (mergeMessages carries the optimistic marker) - SessionChat gates canSteer via isSteeringSupportedForSession instead of the pi-only check - clearStaleQueuedStatus normalizes a queued status on an invoked message - fix(web): drop duplicate showSessionSummaryInChat in markdown test (upstream typecheck breakage) * fix(codex,shared): address bot findings on steer gate and ambiguous turn/steer - STEERING_SUPPORTED_FLAVORS / isSteeringSupportedForSession advertise codex and pi only; cursor joins when its soft-steer handler lands (#1609) - turn/steer now splits dispatch (stdin accepted) from completion (turn finished): the hub RPC acks once dispatch succeeds — never on the concurrent turn's completion, which can exceed the 30s RPC window - queue row commits only after the turn settles; a rejected/aborted steer restores the row so the message still delivers via turn/start, and a dispatched steer is never restored (no duplicate delivery) - steer carries clientUserMessageId (echoed as userMessage.clientId) so ambiguous transport failures can reconcile the thread later - client tests cover dispatch/complete split and stdin-write failure * fix(codex): reconcile dispatched steers before restoring; align error copy - A dispatched turn/steer whose completion fails (disconnect / protocol error) is now reconciled via thread/read by clientUserMessageId before the queued row is restored — the instruction is only re-delivered by turn/start when the thread never received it - Reconcile targets the pinned steer thread, not whichever turn is current when completion fails - syncEngine unsupported-flavor error now matches the capability gate (Pi and Codex only until the cursor handler lands) - launcher tests cover steer success (ack on dispatch), reconcile-accepted and reconcile-rejected outcomes * fix(codex): consume the row at dispatch; drop background reconcile - The hub RPC acks and the queue row is consumed as soon as stdin accepts turn/steer; completion is background-only logging. A dispatched steer is never restored, so the same localId cannot be re-delivered via turn/start after the caller was told the steer succeeded - Dispatch failure (stdin write error) still restores the row and reports failure - steer.completed rejection is always handled (no unhandled rejection on the dispatch-failure path) - tests updated: completion failure after dispatch keeps the row consumed; dispatch failure restores it * fix(codex): distinguish definite rejection from indeterminate completion - Transport-level failures (timeout, abort, disconnect, spawn, protocol) carry an indeterminate marker; explicit JSON-RPC error responses do not - After a dispatched steer, turn completion resolves → commit + consumed; a definite app-server rejection restores the row (instruction was never accepted, so turn/start cannot duplicate it); an indeterminate outcome leaves the row reserved so it can never be delivered twice - Completion handling registers before awaiting dispatch so the dispatch-failure path cannot leak an unhandled rejection - client/launcher tests cover explicit rejection (restore), indeterminate outcome (row stays reserved) and dispatch failure * fix(codex): reconcile indeterminate steers instead of a permanent reservation - After an indeterminate completion (disconnect/protocol), reconcile the thread by clientUserMessageId immediately: accepted → commit + consumed, provably rejected → restore, still unreadable → keep the reservation and retry from the main-loop top on later passes (post-reconnect) - A row never sits in dispatching forever: the hub cannot stamp it invoked while the instruction may never have been accepted - tests: indeterminate keeps reserved while thread unreadable; accepted reconciliation consumes; rejected path restores * fix(codex): accept all thread item shapes; retry reconcile; ack through abort - Reconcile matcher accepts userMessage/user_message with clientId/ client_id, matching the shapes the thread parser supports — an accepted steer can no longer be misclassified as rejected - A pending reconciliation schedules a wakeLoop retry, so a temporary app-server outage cannot strand the reservation behind waitForTurnOrRecovery - The success-path ACK no longer checks the steer epoch: the hub already reported steered on dispatch, so commit + messages-consumed must reach it even when an abort resets the queue in between * fix(codex): reinit reconnected app-server; keep reconcile retries alive - thread/read after a disconnect auto-connects a fresh app-server, which must be initialized before any request — reconcile now ensures connect + initialize (isConnected getter added to the client) - every still-unknown loop-top reconciliation schedules the next retry, so recovery without external traffic is eventually observed - launcher mock gains isConnected * fix(codex): timer-driven reconciliation; init tracking; abort-safe ACK - Reconciliation runs on a self-rescheduling 1s timer independent of the main loop (wakes it too), so idle loops and waitForTurnOrRecovery still observe app-server recovery; abort clears nothing implicitly — the ACK path commits and consumes even when the reservation was cancelled - Absence of a durable client id is ambiguous: unmatched reads stay 'unknown' and keep retrying instead of restoring the row - CodexAppServerClient tracks initialized state (reset on disconnect/exit) so ensureAppServerInitialized re-initializes a fresh process before thread/read; initialize failures leave the flag false for the next retry - tests: accepted reconciliation via scheduled timer, indeterminate keeps reserved, explicit rejection restores * fix(codex): bind reconciliation to the launcher lifecycle - runSteerReconciliation clears any armed retry timer on entry and never installs a second one, so loop-top and timer-driven passes cannot multiply - shuttingDown is set when the main loop ends: timers are cleared and the pending map is dropped, so an unresolved steer can never respawn an app-server after cleanup (remote-to-local switch included) * fix(codex): report steered only after app-server acceptance - The handler now awaits steer.completed (the inject-acceptance response): an explicit JSON-RPC rejection surfaces as failed and restores the row for the normal turn/start path instead of a false steered - Transport failure after dispatch reports 'Steer outcome is being reconciled' and keeps the row reserved while the timer-driven thread reconciliation runs - dispatch-failure path also swallows the paired completion rejection * fix(steer): tri-state cancel, clear-safe reservations, bounded acceptance wait - MessageQueue2.cancelByLocalId returns 'in-flight' for a dispatching steer reservation: the hub neither deletes the row nor stamps invoked_at (new CancelMessageResponse 'busy' status; web restores the optimistic row); pushIsolateAndClear and reset/close share cancelReservations so /clear-style commands cannot have a rejected steer resurrect a discarded prompt - turn/steer acceptance wait bounded at 25s (< hub 30s RPC timeout): a lost response is indeterminate and funnels into thread reconciliation instead of stranding the reservation - tests updated for the tri-state cancel contract * fix(codex,web): busy-aware edit flow; bound reconciliation reads - QueuedMessagesBar edit flow treats a 'busy' cancel as unsuccessful: it never prefills the composer when the row is inside an async steer, so a second client cannot send a duplicate - reconcileSteerByClientId bounds thread/read with a 5s timeout so a connected-but-silent app-server cannot hold the reservation in-flight indefinitely * fix(steer): inFlight-dominated cancel acks; bounded reconciliation - hub cancel-queued-message acks check inFlight before removed: a stale duplicate socket reporting removed can no longer delete the durable row while another socket is dispatching the steer - reconciliation entries expire after 60s and mark delivered: after the rejection window, a dispatched steer that the app-server never proved (client ids dropped on restart) is committed instead of polling thread/read forever - pre-dispatch failures (abort before write included) never enter reconciliation — they restore the row and report failure * fix(steer): persist indeterminate outcomes without replay * fix(steer): make ambiguous delivery restart-safe * fix(steer): recover crash-held rows and preserve retry dedup * fix(steer): ack retries and bound stdin dispatch * fix(steer): reconcile indeterminate dispatches and serialize retries * fix(codex): classify stdin callback failures as indeterminate * fix(steer): recheck indeterminate cancels after ACK * fix(steer): close retry and abort races * fix(steer): serialize live retries and abort admission * fix(steer): distinguish live dispatching from unknown * fix(steer): keep ACK failures held and reconcile busy cancel * fix(steer): distinguish held cancel from removal * fix(store): combine schema v24 migrations * fix(store): reserve schema v25 for steer delivery state * fix(steer): keep held cancel state and notify requeue * fix(steer): release explicitly cancelled unknown reservations * fix(codex): reject cancelled reservations before native steer * fix(codex): make reservation restore atomic with state * fix(codex): terminate abandoned transport writes * fix(steer): own abandoned app-server lifecycle and consume races * fix(codex): confirm dispatch and recover abandoned turns * test(codex): mock abandoned transport callback * fix(codex): clear visible turn state on transport loss * fix(steer): claim retries and cover native delivery state * fix(native): preserve indeterminate state on Android hydration * fix(steer): make retry claims single-winner * fix(steer): serialize concurrent retry claims * fix(socket): tolerate missing steer-state ACK callbacks * fix(native): serialize retry operations * docs(web): document unknown steer delivery and retry controls * fix(steer): handle retry failures and abort-before-connect * fix(steer): reinitialize after transport loss and finish iOS retry errors * fix(steer): preserve indeterminate rows across reconnect gaps * test(web): mock indeterminate queued recovery state * fix(steer): recover consumed ACK tombstones * fix(steer): expose consumed cancel tombstones
HAPI Android Companion
Native Android client (Kotlin + Jetpack Compose) for the HAPI hub. Fully
independent from the web app; shares only the protocol contract
(docs/api/) and the golden fixtures (shared/fixtures/).
- applicationId:
run.hapi.companion· minSdk 26 · target/compileSdk 36 - Toolchain: Gradle 8.14.2 (wrapper) · AGP 8.11.1 · Kotlin 2.1.21 · Compose BOM 2025.05.00 · JDK 17+ (CI uses 21)
Modules
| Module | Type | Responsibility |
|---|---|---|
:core:protocol |
pure Kotlin/JVM (no Android) | Hub wire types (kotlinx.serialization), chat pipeline port (normalize → reduce → tool groups), message-window/pagination logic, versioned patch application, modes catalog, git output parsers, BindLink pairing-link parsing. M1a landed: wire/ (HapiJson, Session/SessionPatch/SessionSummary, DecryptedMessage, AgentState, Machine, 13-type SyncEvent union via SyncEvents.parse, MessagesResponse), catalog/ (flavors + permission/collaboration modes), patch/SessionPatching.kt (exact port of web/src/lib/sessionPatch.ts), all fixture-verified. |
:core:data |
Android library | Transport + persistence. M1b landed — auth/ (JwtPeek, CredentialStore interface + EncryptedPrefsCredentialStore/in-memory, HubUrls origin normalization, HubRegistry roster behind a storage seam, AuthInterceptor + single-flight TokenAuthenticator with ensureFreshToken() and terminal AuthEvents), api/ (HapiApi — plain OkHttp + kotlinx.serialization, one suspend fun per v1 endpoint incl. generated-image bytes via a 256 MB OkHttp cache and the multipart transcription helper; ApiError with (status, code)), HubSession per-hub factory; MockWebServer-tested. M1c landed — sse/: SseEngine (per-key global/session:<id> loops, connection-changed handshake gate with ok/gap resume verdict, per-key Last-Event-ID cursors advanced only after downstream hand-off (at-least-once), 10 s connect deadline, 90 s watchdog, 1 s→30 s→300 s backoff + jitter, background retry deferral + 45 s foreground stale check, one silent 401 re-auth per cycle), OkHttpSseTransport (dedicated client, readTimeout=0, incremental gzip decoding pinned by test, acceptEncodingIdentity fallback), SyncEventRouter → SyncTargets seam; virtual-time tested. Still to come: StateFlow stores + AtomicFile JSON snapshots (M2), FCM registration + WorkManager workers (M4). |
:app |
Android application | Compose UI, navigation, deep links (hapicompanion://bind), FCM service (M4), hand-rolled DI (AppGraph, no Hilt). M1d landed — di/ (AppGraph process singletons: Preferences DataStore-backed HubRegistryStorage, EncryptedPrefsCredentialStore, HubRegistry, auth-terminal fan-out; HubGraph per active hub: HubSession + SseEngine wired to ensureFreshToken, recreated on hub switch; LocalAppGraph CompositionLocal + viewModelFactory helper), feature/pairing/ (landing / zxing ScanContract QR scan / manual entry sharing one PairingViewModel: health + protocol check → POST /api/auth → persist + activate), feature/home/ placeholder (hub switcher + sign-out), Navigation.kt (pairing ⇄ home, auth-terminal → pairing with banner), bind deep-link handling in MainActivity. |
Dependency direction: :app → :core:data → :core:protocol.
Protocol conformance fixtures
:core:protocol is the porting target for web/src/chat/ and is verified
against golden fixtures generated from the web implementation (track K).
The test task already passes the fixtures location as a system property:
// core/protocol/build.gradle.kts
tasks.test {
systemProperty("hapi.fixtures.dir", rootDir.parentFile.resolve("shared/fixtures").absolutePath)
}
Fixture-driven tests (M2) read System.getProperty("hapi.fixtures.dir") —
no further build changes are needed when shared/fixtures/** lands. CI
re-runs this suite whenever android/** or shared/fixtures/** change.
Building
Requires an Android SDK for :app/:core:data (set ANDROID_HOME or
android/local.properties with sdk.dir=...). :core:protocol alone needs
only a JDK.
cd android
./gradlew :core:protocol:test # pure JVM protocol tests (fast)
./gradlew :app:assembleDebug # debug APK
./gradlew :app:installDebug # install on a connected device
Without an Android SDK you can still run the protocol suite by configuring only the needed projects:
./gradlew --no-configuration-cache --configure-on-demand :core:protocol:test
CI (.github/workflows/android.yml) runs the protocol tests and
:app:assembleDebug on every PR touching android/** or shared/fixtures/**.
Pairing
HAPI is self-hosted: the app talks to a hub you run. Pairing = giving the
app a hub URL plus that hub's access token; the app verifies the hub
(GET /health, protocol version), exchanges the token for a JWT
(POST /api/auth), stores the credentials in EncryptedSharedPreferences
(keyed per hub — multiple hubs can be paired, one active at a time), and
lands on the session UI. Three entry points:
- QR scan — the hub prints two QR codes when started with
--relay(also under web Settings → Companion pairing). The in-app scanner accepts both: the companion deeplink (hapicompanion://bind?hub=…&code=…) and the web direct-access URL (…?hub=…&token=…). - Deep link — scanning the companion QR with the system camera opens the
app directly with a confirm screen (
hapicompanion://bindintent filter). - Manual entry — hub URL + access token, for hubs started without
--relay.
Pairing against a local dev hub
# repo root: start the hub (prints the access token + QR codes)
bun run dev
# emulator: the host machine is 10.0.2.2
# Hub URL: http://10.0.2.2:3006
# Access token: from the hub terminal / hub settings.json (CLI_API_TOKEN)
# physical device: use the machine's LAN IP, e.g. http://192.168.1.10:3006
adb shell am start -a android.intent.action.VIEW \
-d "hapicompanion://bind?hub=http%3A%2F%2F10.0.2.2%3A3006&code=<accessToken>" # optional: exercises the deep link
Plain-http LAN/emulator hubs work in all build types: the manifest opts in
to cleartext traffic (android:usesCleartextTraffic="true"), because
self-hosted LAN hubs are the primary pairing target and Android cannot scope
the exemption to local addresses only. Sign-out (home → Sign out) deletes the
stored credentials for that hub and drops it from the roster.
Milestones (track B of the native-clients plan)
- M0 — this scaffold: modules, version catalog, CI, placeholder screen.
- M1 — foundations: wire types + modes catalog; auth +
HapiApi(MockWebServer-tested);SseEnginereconnect state machine + versioned patches (gzip streaming verified); pairing UI +hapicompanion://binddeep link. - M2 — read-only chat: chat pipeline port gated on fixtures all-green; session list;
MessageWindowStoreport; Markdown renderer; read-only chat screen (LazyColumn(reverseLayout = true)). - M3 — interaction: composer (optimistic send/queue/steer/drafts), permission approvals UX, session controls (mode/model/abort/resume/rename/archive), new session, dictation.
- B-M3ce landed — voice dictation: mic button in the composer (
RECORD_AUDIOrequested at first use),MediaRecorder→ m4a/AAC, provider discovery viaGET /api/voice/transcription/providers(firststandard-capable provider; a hub without one shows a notice), upload through the multipartPOST /api/voice/transcription, transcript appended at the composer text with a space separator;DictationControlleris a plain seam over recorder + API, JVM-tested with fakes. Slash commands: typing a lone/tokenopens a dropdown merging the session'smetadata.slashCommandsnames with theGET /slash-commandsRPC list (RPC entries win dedupe; exact → prefix → contains filtering), tap inserts/name(the skills$trigger is deferred). Session ops: list long-press sheet and chat top-bar overflow gain Rename (PATCH /sessions/:id, optimistic name with roll-forward on failure), Delete (confirm; 409-while-active surfaced), and Reopen for inactive sessions (POST /reopen; a superseding id reuses the supersede path — window seed + draft move + navigate-replace; 422 missing-metadata formatted); chat shows an inactive-session bar ("send to resume, or Reopen").
- B-M3ce landed — voice dictation: mic button in the composer (
- M4 — FCM push (register → notification actions via expedited WorkManager) + files/git viewer, Scratchlist, usage/storage stats.
- B-M4a landed — FCM push + notification actions.
:core:datapush/:PushPayload(data-only contract v1 decoding: type/severity/notifySummaryparsing, channel routing,type-<sessionId>coalescing tags, unknown type/contractVersion degrade to plain title/body),DeviceRegistrar(registers the FCM token with every paired hub on start/pairing/onNewToken, DataStore-persisteddeviceIdUUID, WorkManager retry seam, best-effort unregister on sign-out before credentials are wiped),PushHubAccess+PushActionRunner(workers build aHubSessionon demand from stored credentials — noHubGraphneeded in background — and resolve the owning hub: active hub first, other paired hubs on 404 session-miss).:app:push/PushBinding(Firebase availability gate — nogoogle-services.json→ all push paths no-op),fcm/(HapiFirebaseMessagingService,NotificationChannels—permission_requestsHIGH /ready/task_notifications,PushNotificationsbuilder with severity accents + suppress-when-open rule,NotificationActionReceiver→ expeditedPermissionActionWorker(Allow/Deny → approve/deny{}) andSendMessageWorker(RemoteInput reply →{text, localId}) with pending → done/"Already handled"/failed notification states), WorkManager on-demand init +HapiWorkerFactory, notification tap → internalMainActivityintent route → chat.
- B-M4a landed — FCM push + notification actions.
- M5 — polish: zh-CN i18n, OLED/Material You theming, predictive back, LeakCanary pass, Play listing + self-build docs.
- B-M5a landed — zh-CN localization + in-app language switching (see "Internationalization" below).
Internationalization (B-M5a)
The app ships English (default) and Simplified Chinese
(app/src/main/res/values-zh-rCN/strings.xml). Every user-visible string
lives in resources; both files carry the same key set (lint
MissingTranslation is the gate).
Adding a string
- Add it to
app/src/main/res/values/strings.xmlwith a feature-prefixed key matching the existing convention (chat_,sessions_,files_,scratchlist_,pairing_,settings_,new_session_,notif_,tool_for tool-card titles). Dynamic values use positional format args (%1$s,%2$d); count-dependent copy uses explicit_one/_manykeys (the deliberate house style — no<plurals>). - Add the zh-CN twin to
values-zh-rCN/strings.xml. Terminology source of truth is the web corpusweb/src/lib/locales/zh-CN.ts— reuse its product terms (会话 session, 机器 machine, 权限模式 permission mode, 工作树 worktree, 草稿夹 scratchlist, 语音输入 dictation, 用量 usage, 智能体/代理 agent). Technical identifiers (model ids, flavor names like Claude/Codex, permission-mode catalog labels, CLI flags) stay untranslated, matching the web's choices. - Reference it: composables via
stringResource(R.string...). ViewModels stay string-free — transient notices are semantic sealed types (ChatNotice,ScratchlistNotice,PairingError,DictationErrorKind) resolved at the UI layer; where a ViewModel genuinely composes display text it takes a small Strings seam (FilesStrings,FileViewerStrings,NewSessionStrings) whose defaults are the English values (JVM tests construct without arguments) and whose production instance is resource-resolved in the Navigation holders. Server-provided error text passes through verbatim.
Language switching
Settings → App language offers Follow system (default) / English /
简体中文. The choice persists in LanguagePrefs (DataStore) and applies
immediately via AppCompatDelegate.setApplicationLocales:
MainActivityextendsAppCompatActivity(theme parentTheme.AppCompat.DayNight.NoActionBar) so per-app locales work back to API 26; on API 33+ the frameworkLocaleManagertakes over (the app also declaresandroid:localeConfigfor the system App-languages screen).- The manifest opts into appcompat's
autoStoreLocales(AppLocalesMetadataHolderServicemeta-data), which re-applies the stored choice synchronously on cold start. - Surfaces that resolve strings from the application context — FCM
notifications, WorkManager result updates, notification-action receivers —
wrap their context with
localizedForAppLanguage(AppGraph.appLanguage)(di/LocaleContexts.kt), since per-app locales only retarget activity contexts below API 33.
Out of scope on purpose: :core:protocol presentation strings
(getEventPresentation, tool-group activity titles) stay English — the web
does not translate them either, and terminology parity with the web wins.
Firebase / push
FCM needs a Firebase project binding, which is deliberately optional:
the com.google.gms.google-services plugin is applied conditionally
(only when app/google-services.json exists — see app/build.gradle.kts),
so the repo always builds green without any Firebase config. Without one,
FirebaseApp never initializes, PushBinding.isAvailable reports false,
and every push code path (registration, FCM service, workers, the
notification-permission prompt) no-ops — the app behaves like pre-M4a.
To enable push:
- Official builds: CI injects the default Firebase project's
google-services.jsonbefore assembling (the file is gitignored;app/google-services.json.exampledocuments the expected shape). - Self-builds: create your own Firebase project, add an Android app
with your
applicationId(defaultrun.hapi.companion), downloadgoogle-services.jsonintoandroid/app/, and rebuild. - Hub side: point the hub at the same Firebase project —
FCM_SERVICE_ACCOUNT_PATH+FCM_PROJECT_ID(docs/api/native-companion-contract.md). The device registers itself with every paired hub (POST /api/devices/register) on pairing, app start, and token rotation, and unregisters on sign-out.
Multi-hub note: the FCM payload does not name the sending hub (contract v1), so notification actions resolve it — the workers try the active hub first, then the other paired hubs when a hub answers 404 for the session. Single-hub setups always hit on the first try. Tapping a notification opens the session against the active hub.
Planned for v1.x: runtime FirebaseOptions handed out by the hub, so
self-builds get push without baking a config into the APK. That lands
entirely behind the existing app/.../push/PushBinding.kt seam.