Merge dev into main: kap-server steer flag & agent-core-v2 full compaction

AGENTS.md

@@ -24,8 +24,7 @@ This is a TypeScript monorepo built for agent-assisted development. Keep the roo
 - `packages/kaos`: the execution environment and file/process abstractions.
 - `packages/oauth`: Kimi OAuth and managed auth utilities.
 - `packages/telemetry`: shared client-side telemetry infrastructure.
-- `packages/transcript`: the isomorphic transcript rendering data layer — L1 agent-granular store, L2 idempotent operations, L3 `off/turn/block/delta` subscription granularity, L4 framework-free view registry, plus turn-cursor pagination. Pure TypeScript (browser-safe, no engine imports); the sole owner of the transcript contract types (`src/contract/`) and the op-batch sequencing contract.
-- `packages/kap-server`: the Kimi Code server, backed by `@moonshot-ai/agent-core-v2`; exposes sessions over REST + WebSocket (`/api/v1` + `/api/v1/ws`), plus the `/api/v1/debug/*` reflection RPC surface (`--debug-endpoints`, loopback bind + bearer auth).
+- `packages/kap-server`: the Kimi Code server, backed by `@moonshot-ai/agent-core-v2`; exposes sessions over REST (`/api/v1`) and the v3 flat entity message WebSocket protocol (`/api/v3/ws`), plus the `/api/v1/debug/*` reflection RPC surface (`--debug-endpoints`, loopback bind + bearer auth).
 - `packages/remote-control`: the Kimi Remote Control tunnel client — registers this machine with the relay and forwards HTTP/WebSocket traffic to the local server, with a machine-wide single-instance lock; consumed by kap-server (the `/api/v1/remote-control` toggle) and by the CLI (`kimi web --remote-control`).
 - `packages/klient`: the client SDK — a contract-driven facade over agent-core-v2 (`global.*` / `session(id).*` / `agent(id).*`, zod-validated); transport via subpath entry (`@moonshot-ai/klient/ipc|memory`, both return the same `Klient`); also hosts the e2e suites. See `packages/klient/AGENTS.md`.
 - `packages/tree-sitter-bash`: a pure-TypeScript bash parser (no runtime deps, no wasm); `parse(source, { timeoutMs, maxNodes })` runs under a deterministic budget and returns a discriminated `ParseResult` — callers must treat aborted/hasError trees as "cannot analyze" and degrade. Parser only, no safety judgments; see the package README's "Known differences" section.
@@ -48,7 +47,7 @@ This is a TypeScript monorepo built for agent-assisted development. Keep the roo
 
 ## General Coding Rules
 
-- `packages/agent-core-v2`, `packages/kap-server`, and `packages/transcript` are comment-free zones: no comments of any kind — no line/block comments, no JSDoc (not even on exported symbols); the only exception is load-bearing lint-suppression directives (`oxlint-disable` / `eslint-disable`), while other tooling directives (`@ts-expect-error`, …) stay banned. Enforced by `scripts/check-no-comments.mjs` over `.ts`/`.tsx`/`.mts`/`.mjs` under `src/`/`test/`/`scripts/`, which runs as part of `pnpm lint`.
+- `packages/agent-core-v2` and `packages/kap-server` are comment-free zones: no comments of any kind — no line/block comments, no JSDoc (not even on exported symbols); the only exception is load-bearing lint-suppression directives (`oxlint-disable` / `eslint-disable`), while other tooling directives (`@ts-expect-error`, …) stay banned. Enforced by `scripts/check-no-comments.mjs` over `.ts`/`.tsx`/`.mts`/`.mjs` under `src/`/`test/`/`scripts/`, which runs as part of `pnpm lint`.
 - For optional object properties, pass `undefined` directly instead of using conditional spread.
   - YES: `{ user }`
   - NO: `{ ...(user ? { user } : undefined) }`

apps/kimi-inspect/AGENTS.md

@@ -10,7 +10,7 @@ A left icon rail (`src/components/NavRail.tsx`) switches top-level views:
 - **Global message search** (`src/components/SearchView.tsx`) — cross-session full-text search over `POST /api/v1/search`, cursor-paged via a manual Load more; an exact-match checkbox maps to the API's `mode: 'literal'` substring search, which ignores sort and orders newest-first; a `live`/`index` badge on the results shows which server route served them (in-memory session transcript vs the persisted index).
 - **Model Catalog** (`src/components/ModelCatalogView.tsx`) — every Provider with its Models and the default marker, via `IModelCatalog` / `IModelService` channel proxies, with per-model ping and session creation actions.
 - **App Services** (`src/components/AppServicesView.tsx`) — the app-scope Service reflection, full width, joined by the **Workspace Services** view (`src/components/WorkspaceServicesView.tsx`) — the workspace-scope counterpart with a left sidebar directory browser (`src/components/WorkspaceDirBrowser.tsx` — server-side fs browsing over the App-scope `IHostFolderBrowser`, marking entries that are registered workspaces with their `IWorkspaceTrust` trust state, and registering a picked folder on demand via `IWorkspaceService.createOrTouch`), its proxies riding the `/workspace/:id` route, which materializes the handler on demand via `IWorkspaceLifecycleService.handlerFor`.
-- **DI view** (`src/components/DiInspectionView.tsx`) — the engine's Service × Effect × DI debug surface over the App-scope `IDebugLedgerService` / `IDebugGraphService` / `IDebugEventsService` / `IDebugCascadeService`: the unit tree = ledger tree with unprovide / update / dispose triggers, the dependency DAG as Miller columns (`di/DiGraphPanel.tsx`), the event-subscription ledger (unit-book `on:<name>` entries + per-bus listener counts, `di/DiEventsPanel.tsx`), the cascade history, and the waiting area; the five panels poll on a short interval and refresh eagerly off the global `event.di.unit_changed` WS frame via `src/activity/di.ts`, which invalidates the `['di']` react-query prefix.
+- **DI view** (`src/components/DiInspectionView.tsx`) — the engine's Service × Effect × DI debug surface over the App-scope `IDebugLedgerService` / `IDebugGraphService` / `IDebugEventsService` / `IDebugCascadeService`: the unit tree = ledger tree with unprovide / update / dispose triggers, the dependency DAG as Miller columns (`di/DiGraphPanel.tsx`), the event-subscription ledger (unit-book `on:<name>` entries + per-bus listener counts, `di/DiEventsPanel.tsx`), the cascade history, and the waiting area; the five panels poll on a short interval over the `/api/v1/debug` RPC (no push channel — a settled trigger invalidates the `['di']` react-query prefix so every panel converges).
 
 The **Agent scope** stays in the Chat view's right dock (`src/components/RightPanel.tsx` — Audit / Agent / State / Session tabs) across two of them:
 
@@ -25,7 +25,7 @@ Built on its own old-klient-style channel layer (`src/channel/`: the VS Code `Pr
 
 ## Session activity
 
-Session-level coarse status is the one exception to no-push: `src/activity/` holds a second `/api/v1/ws` client (`GlobalEventsWs`) that subscribes to nothing and consumes the server-pushed global facts — `event.session.work_changed` updates a per-session activity map (`SessionActivityHub` + subscribe/version store, seeded on connect/reconnect from `GET /api/v1/sessions`), while `event.session.created` / `event.session.archived` / `session.meta.updated` / `event.workspace.*` invalidate the `['sessions']` / `['v2-sessions']` / `['workspaces']` queries (an archive also drops the session's live activity entry, since no further `work_changed` frames will correct a stale badge); the session tree rows render `running` / `approval` / `question` / `failed` badges from it via `useSessionActivities` (live facts override the REST `activity.status`).
+Session-level coarse status is the one exception to no-push: `src/activity/` holds a second `/api/v3/ws` client (`GlobalEventsWs`) that subscribes to nothing and consumes the server-pushed global messages — the hub fans them out to every established connection with no subscribe frame — where every `session` message (created / updated / archived / deleted) embeds the full SessionInfo whose `busy` / `main_turn_active` / `pending_interaction` / `last_turn_reason` update a per-session activity map (`SessionActivityHub` + subscribe/version store, seeded on connect/reconnect from `GET /api/v1/sessions`), while the message subtypes plus `workspace` created / updated / deleted invalidate the `['sessions']` / `['v2-sessions']` / `['workspaces']` queries (an archive or delete also drops the session's live activity entry, since no further messages will correct a stale badge); the session tree rows render `running` / `approval` / `question` / `failed` badges from it via `useSessionActivities` (live facts override the REST `activity.status`). Session/agent-grained traffic (`session.state`, entities, deltas) stays subscribe-gated server-side and never arrives on this socket; the DI view has no push feed at all — its panels poll the `/api/v1/debug` RPC.
 
 ## Dev server
 

apps/kimi-inspect/README.md

@@ -35,8 +35,8 @@ there is no fallback data source.
   workspace handlers materialize on demand).
 - **DI** — the engine's Service × Effect × DI debug surface, four panels fed
   by the App-scope debug Services (`IDebugLedgerService` / `IDebugGraphService`
-  / `IDebugCascadeService`) and refreshed eagerly off the `event.di.unit_changed`
-  WS frame:
+  / `IDebugCascadeService`) over the `/api/v1/debug` RPC, polling on a short
+  interval:
   - **Unit tree** — scope → unit → ledger entries (label, five-state
     `Pending / Activating / Active / Unloading / Failed`, uid, `pinned` flag,
     unit error object), with **unprovide / update / dispose** triggers.
@@ -53,5 +53,5 @@ there is no fallback data source.
   `GET /api/v1/debug/channels` enumerates every scoped Service — there is no
   whitelist; new Services appear automatically.
 - There is no Service-event push channel besides the global events listed
-  above; panels fetch/refresh on demand (react-query, 15 s poll) plus the
-  `event.di.unit_changed` invalidation for the DI view.
+  above; panels fetch/refresh on demand (react-query, 15 s poll — the DI view
+  polls on its own short interval).

apps/kimi-inspect/src/activity/store.test.ts

@@ -58,6 +58,37 @@ function seedFetch(items: Record<string, unknown>[]): typeof fetch {
   })) as unknown as typeof fetch;
 }
 
+function sessionMessage(
+  subtype: 'created' | 'updated' | 'archived' | 'deleted',
+  session: Record<string, unknown> & { id: string },
+): Record<string, unknown> {
+  return {
+    type: 'session',
+    timestamp: Date.now(),
+    subtype,
+    session: {
+      workspace_id: 'wd_example_0123456789ab',
+      title: 'session',
+      created_at: new Date().toISOString(),
+      updated_at: new Date().toISOString(),
+      busy: false,
+      metadata: { cwd: '/tmp/example' },
+      agent_config: { model: 'test-model' },
+      usage: {
+        input_tokens: 0,
+        output_tokens: 0,
+        cache_read_tokens: 0,
+        cache_creation_tokens: 0,
+        context_tokens: 0,
+      },
+      permission_rules: [],
+      message_count: 0,
+      last_seq: 0,
+      ...session,
+    },
+  };
+}
+
 describe('SessionActivityStore', () => {
   it('applies work facts and notifies with a version bump', () => {
     const store = new SessionActivityStore();
@@ -112,17 +143,13 @@ describe('SessionActivityHub', () => {
 
     expect(hub.store.get('s1')).toEqual(facts({ busy: true, mainTurnActive: true }));
     expect(hub.store.get('s2')?.pendingInteraction).toBe('approval');
-    // The hello goes out with no subscriptions — global facts flow regardless.
-    const hello = JSON.parse(instances[0]!.sent[0]!) as {
-      type: string;
-      payload: { subscriptions: string[] };
-    };
-    expect(hello.type).toBe('client_hello');
-    expect(hello.payload.subscriptions).toEqual([]);
+    // Nothing goes out — v3 global messages flow to every connection with no
+    // subscribe frame.
+    expect(instances[0]!.sent).toEqual([]);
     hub.close();
   });
 
-  it('applies live work_changed frames by session id', () => {
+  it('applies live session messages by session id', () => {
     const { ctor, instances } = makeFakeWsCtor();
     const hub = new SessionActivityHub({
       url: 'http://127.0.0.1:58627',
@@ -132,25 +159,22 @@ describe('SessionActivityHub', () => {
     });
     instances[0]!.emit('open');
 
-    instances[0]!.emitFrame({
-      type: 'event.session.work_changed',
-      session_id: 's1',
-      payload: {
-        type: 'event.session.work_changed',
+    instances[0]!.emitFrame(
+      sessionMessage('updated', {
+        id: 's1',
         busy: true,
         main_turn_active: true,
         pending_interaction: 'question',
-        last_turn_reason: null,
-      },
-    });
+      }),
+    );
 
     expect(hub.store.get('s1')).toEqual(
       facts({ busy: true, mainTurnActive: true, pendingInteraction: 'question' }),
     );
     hub.close();
   });
 
-  it('forwards created and meta updates as list-level signals', () => {
+  it('forwards created and updated messages as list-level signals', () => {
     const { ctor, instances } = makeFakeWsCtor();
     const onListChanged = vi.fn();
     const hub = new SessionActivityHub({
@@ -161,17 +185,17 @@ describe('SessionActivityHub', () => {
     });
     instances[0]!.emit('open');
 
-    instances[0]!.emitFrame({ type: 'event.session.created', session_id: 's1', payload: {} });
-    instances[0]!.emitFrame({ type: 'session.meta.updated', session_id: 's1', payload: {} });
-    // Agent-grained frames are ignored even if they somehow arrive.
+    instances[0]!.emitFrame(sessionMessage('created', { id: 's1' }));
+    instances[0]!.emitFrame(sessionMessage('updated', { id: 's1' }));
+    // Unknown future message types are ignored silently.
     instances[0]!.emitFrame({ type: 'turn.started', session_id: 's1', payload: {} });
 
     expect(onListChanged).toHaveBeenCalledTimes(2);
-    expect(hub.store.get('s1')).toBeUndefined();
+    expect(hub.store.get('s1')).toEqual(facts());
     hub.close();
   });
 
-  it('forwards archived and workspace frames as list-level signals and drops archived facts', () => {
+  it('forwards archived/deleted and workspace messages as list-level signals and drops gone facts', () => {
     const { ctor, instances } = makeFakeWsCtor();
     const onListChanged = vi.fn();
     const hub = new SessionActivityHub({
@@ -182,46 +206,36 @@ describe('SessionActivityHub', () => {
     });
     instances[0]!.emit('open');
 
-    instances[0]!.emitFrame({
-      type: 'event.session.work_changed',
-      session_id: 's1',
-      payload: { type: 'event.session.work_changed', busy: true },
-    });
+    instances[0]!.emitFrame(sessionMessage('updated', { id: 's1', busy: true }));
     expect(hub.store.get('s1')).toBeDefined();
 
-    // Global-dispatched frames carry the __global__ watermark; the real
-    // session id rides in the payload.
-    instances[0]!.emitFrame({
-      type: 'event.session.archived',
-      session_id: '__global__',
-      payload: { type: 'event.session.archived', sessionId: 's1', workspace_id: 'wd_1' },
-    });
+    instances[0]!.emitFrame(sessionMessage('archived', { id: 's1', archived: true }));
     expect(hub.store.get('s1')).toBeUndefined();
-    expect(onListChanged).toHaveBeenCalledTimes(1);
+    expect(onListChanged).toHaveBeenCalledTimes(2);
 
-    instances[0]!.emitFrame({
-      type: 'event.session.work_changed',
-      session_id: 's2',
-      payload: { type: 'event.session.work_changed', busy: true },
-    });
+    instances[0]!.emitFrame(sessionMessage('updated', { id: 's2', busy: true }));
     expect(hub.store.get('s2')).toBeDefined();
 
-    instances[0]!.emitFrame({
-      type: 'event.session.deleted',
-      session_id: '__global__',
-      payload: { type: 'event.session.deleted', sessionId: 's2', workspace_id: 'wd_1' },
-    });
+    instances[0]!.emitFrame(sessionMessage('deleted', { id: 's2' }));
     expect(hub.store.get('s2')).toBeUndefined();
-    expect(onListChanged).toHaveBeenCalledTimes(2);
-
-    for (const type of [
-      'event.workspace.created',
-      'event.workspace.updated',
-      'event.workspace.deleted',
-    ]) {
-      instances[0]!.emitFrame({ type, session_id: '__global__', payload: {} });
+    expect(onListChanged).toHaveBeenCalledTimes(4);
+
+    for (const subtype of ['created', 'updated', 'deleted']) {
+      instances[0]!.emitFrame({
+        type: 'workspace',
+        timestamp: Date.now(),
+        subtype,
+        workspace: {
+          id: 'wd_example_0123456789ab',
+          root: '/tmp/example',
+          name: 'example',
+          created_at: new Date().toISOString(),
+          last_opened_at: new Date().toISOString(),
+          session_count: 0,
+        },
+      });
     }
-    expect(onListChanged).toHaveBeenCalledTimes(5);
+    expect(onListChanged).toHaveBeenCalledTimes(7);
     hub.close();
   });
 });