@vellumai/assistant 0.5.1 → 0.5.3

package/src/__tests__/conversation-tool-setup-app-refresh.test.ts

@@ -2,11 +2,10 @@
  * Regression tests for app surface refresh and eventing side effects in
  * createToolExecutor (conversation-tool-setup.ts).
  *
- * After the App Builder tools source swap (PR 8), app tools like app_update,
- * app_file_edit, and app_file_write are provided by the app-builder skill
- * rather than core registration. The afterExecute hooks in createToolExecutor
- * fire based on tool *name* matching, so they must continue to work for
- * skill-origin tools. These tests verify that contract.
+ * The app_refresh hook is the sole hook for app change refresh. These tests
+ * verify that app_refresh, app_create, and app_delete hooks fire correctly,
+ * and that removed hooks (app_update, app_file_edit, app_file_write) no
+ * longer trigger side effects.
  */
 
 import { beforeEach, describe, expect, mock, test } from "bun:test";
@@ -69,6 +68,7 @@ function makeCtx(overrides: Partial<ToolSetupContext> = {}): ToolSetupContext {
       { surfaceType: SurfaceType; data: SurfaceData; title?: string }
     >(),
     surfaceUndoStacks: new Map(),
+    accumulatedSurfaceState: new Map(),
     surfaceActionRequestIds: new Set<string>(),
     currentTurnSurfaces: [],
     isProcessing: () => false,
@@ -109,9 +109,9 @@ describe("session-tool-setup app refresh side effects", () => {
     updatePublishedSpy.mockClear();
   });
 
-  // ── app_update ──────────────────────────────────────────────────────
+  // ── app_refresh ─────────────────────────────────────────────────────
 
-  describe("app_update", () => {
+  describe("app_refresh", () => {
     test("triggers refreshSurfacesForApp when result is not an error", async () => {
       const ctx = makeCtx();
       const executor = makeFakeExecutor({
@@ -129,7 +129,7 @@ describe("session-tool-setup app refresh side effects", () => {
         broadcastSpy,
       );
 
-      await toolFn("app_update", { app_id: "app-1", name: "New Name" });
+      await toolFn("app_refresh", { app_id: "app-1" });
 
       expect(refreshSpy).toHaveBeenCalledTimes(1);
       expect((refreshSpy.mock.calls as unknown[][])[0][0]).toBe(ctx);
@@ -150,7 +150,7 @@ describe("session-tool-setup app refresh side effects", () => {
         broadcastSpy,
       );
 
-      await toolFn("app_update", { app_id: "app-42" });
+      await toolFn("app_refresh", { app_id: "app-42" });
 
       expect(broadcastSpy).toHaveBeenCalledTimes(1);
       expect((broadcastSpy.mock.calls as unknown[][])[0][0]).toEqual({
@@ -172,7 +172,7 @@ describe("session-tool-setup app refresh side effects", () => {
         mock(() => {}),
       );
 
-      await toolFn("app_update", { app_id: "app-publish" });
+      await toolFn("app_refresh", { app_id: "app-publish" });
 
       // updatePublishedAppDeployment is called with void (fire-and-forget),
       // so just verify it was invoked.
@@ -199,7 +199,7 @@ describe("session-tool-setup app refresh side effects", () => {
         broadcastSpy,
       );
 
-      await toolFn("app_update", { app_id: "app-err" });
+      await toolFn("app_refresh", { app_id: "app-err" });
 
       expect(refreshSpy).not.toHaveBeenCalled();
       expect(broadcastSpy).not.toHaveBeenCalled();
@@ -220,303 +220,13 @@ describe("session-tool-setup app refresh side effects", () => {
         broadcastSpy,
       );
 
-      await toolFn("app_update", {});
+      await toolFn("app_refresh", {});
 
       expect(refreshSpy).not.toHaveBeenCalled();
       expect(broadcastSpy).not.toHaveBeenCalled();
     });
   });
 
-  // ── app_file_edit ───────────────────────────────────────────────────
-
-  describe("app_file_edit", () => {
-    test("triggers refreshSurfacesForApp with fileChange flag", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({
-        content: '{"ok":true}',
-        isError: false,
-      });
-      const broadcastSpy = mock(() => {});
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        broadcastSpy,
-      );
-
-      await toolFn("app_file_edit", {
-        app_id: "app-edit",
-        path: "index.html",
-        old_string: "old",
-        new_string: "new",
-      });
-
-      expect(refreshSpy).toHaveBeenCalledTimes(1);
-      expect((refreshSpy.mock.calls as unknown[][])[0][1]).toBe("app-edit");
-      // Verify opts include fileChange: true
-      expect((refreshSpy.mock.calls as unknown[][])[0][2]).toEqual({
-        fileChange: true,
-        status: undefined,
-      });
-    });
-
-    test("propagates status field through refresh opts", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({
-        content: '{"ok":true}',
-        isError: false,
-      });
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        mock(() => {}),
-      );
-
-      await toolFn("app_file_edit", {
-        app_id: "app-status",
-        path: "styles.css",
-        old_string: "x",
-        new_string: "y",
-        status: "updating styles",
-      });
-
-      expect(refreshSpy).toHaveBeenCalledTimes(1);
-      expect((refreshSpy.mock.calls as unknown[][])[0][2]).toEqual({
-        fileChange: true,
-        status: "updating styles",
-      });
-    });
-
-    test("broadcasts app_files_changed for file edit", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({ content: "{}", isError: false });
-      const broadcastSpy = mock(() => {});
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        broadcastSpy,
-      );
-
-      await toolFn("app_file_edit", {
-        app_id: "app-edit-bc",
-        path: "f",
-        old_string: "a",
-        new_string: "b",
-      });
-
-      expect(broadcastSpy).toHaveBeenCalledTimes(1);
-      expect((broadcastSpy.mock.calls as unknown[][])[0][0]).toEqual({
-        type: "app_files_changed",
-        appId: "app-edit-bc",
-      });
-    });
-
-    test("calls updatePublishedAppDeployment for file edit", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({ content: "{}", isError: false });
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        mock(() => {}),
-      );
-
-      await toolFn("app_file_edit", {
-        app_id: "app-pub-edit",
-        path: "f",
-        old_string: "a",
-        new_string: "b",
-      });
-
-      expect(updatePublishedSpy).toHaveBeenCalledTimes(1);
-      expect((updatePublishedSpy.mock.calls as unknown[][])[0][0]).toBe(
-        "app-pub-edit",
-      );
-    });
-
-    test("skips side effects when result is an error", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({ content: "Error", isError: true });
-      const broadcastSpy = mock(() => {});
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        broadcastSpy,
-      );
-
-      await toolFn("app_file_edit", {
-        app_id: "app-err",
-        path: "f",
-        old_string: "a",
-        new_string: "b",
-      });
-
-      expect(refreshSpy).not.toHaveBeenCalled();
-      expect(broadcastSpy).not.toHaveBeenCalled();
-      expect(updatePublishedSpy).not.toHaveBeenCalled();
-    });
-  });
-
-  // ── app_file_write ──────────────────────────────────────────────────
-
-  describe("app_file_write", () => {
-    test("triggers refreshSurfacesForApp with fileChange flag", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({
-        content: '{"written":true}',
-        isError: false,
-      });
-      const broadcastSpy = mock(() => {});
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        broadcastSpy,
-      );
-
-      await toolFn("app_file_write", {
-        app_id: "app-write",
-        path: "new.html",
-        content: "<div/>",
-      });
-
-      expect(refreshSpy).toHaveBeenCalledTimes(1);
-      expect((refreshSpy.mock.calls as unknown[][])[0][1]).toBe("app-write");
-      expect((refreshSpy.mock.calls as unknown[][])[0][2]).toEqual({
-        fileChange: true,
-        status: undefined,
-      });
-    });
-
-    test("propagates status field through refresh opts", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({
-        content: '{"written":true}',
-        isError: false,
-      });
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        mock(() => {}),
-      );
-
-      await toolFn("app_file_write", {
-        app_id: "app-ws",
-        path: "f.txt",
-        content: "hi",
-        status: "adding dark mode",
-      });
-
-      expect(refreshSpy).toHaveBeenCalledTimes(1);
-      expect((refreshSpy.mock.calls as unknown[][])[0][2]).toEqual({
-        fileChange: true,
-        status: "adding dark mode",
-      });
-    });
-
-    test("broadcasts app_files_changed for file write", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({ content: "{}", isError: false });
-      const broadcastSpy = mock(() => {});
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        broadcastSpy,
-      );
-
-      await toolFn("app_file_write", {
-        app_id: "app-write-bc",
-        path: "f",
-        content: "x",
-      });
-
-      expect(broadcastSpy).toHaveBeenCalledTimes(1);
-      expect((broadcastSpy.mock.calls as unknown[][])[0][0]).toEqual({
-        type: "app_files_changed",
-        appId: "app-write-bc",
-      });
-    });
-
-    test("calls updatePublishedAppDeployment for file write", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({ content: "{}", isError: false });
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        mock(() => {}),
-      );
-
-      await toolFn("app_file_write", {
-        app_id: "app-pub-write",
-        path: "f",
-        content: "x",
-      });
-
-      expect(updatePublishedSpy).toHaveBeenCalledTimes(1);
-      expect((updatePublishedSpy.mock.calls as unknown[][])[0][0]).toBe(
-        "app-pub-write",
-      );
-    });
-
-    test("skips side effects when result is an error", async () => {
-      const ctx = makeCtx();
-      const executor = makeFakeExecutor({ content: "Error", isError: true });
-      const broadcastSpy = mock(() => {});
-
-      const toolFn = createToolExecutor(
-        executor as unknown as ToolExecutor,
-        noopPrompter,
-        noopSecretPrompter,
-        ctx,
-        noopLifecycleHandler,
-        broadcastSpy,
-      );
-
-      await toolFn("app_file_write", {
-        app_id: "app-err",
-        path: "f",
-        content: "x",
-      });
-
-      expect(refreshSpy).not.toHaveBeenCalled();
-      expect(broadcastSpy).not.toHaveBeenCalled();
-      expect(updatePublishedSpy).not.toHaveBeenCalled();
-    });
-  });
-
   // ── app_create side effects ─────────────────────────────────────────
 
   describe("app_create side effects", () => {
@@ -616,7 +326,7 @@ describe("session-tool-setup app refresh side effects", () => {
 
   describe("name-based hooks fire for skill-origin tools", () => {
     test("hooks fire purely on tool name, regardless of tool origin", async () => {
-      // The key invariant: createToolExecutor uses `name === 'app_update'`
+      // The key invariant: createToolExecutor uses `name === 'app_refresh'`
       // string comparison, not tool metadata or origin. This means skill-
       // projected tools with the same name trigger the same afterExecute
       // hooks as core tools.
@@ -633,22 +343,14 @@ describe("session-tool-setup app refresh side effects", () => {
         broadcastSpy,
       );
 
-      // Simulate calling each app tool by name (as the agent loop does)
-      for (const toolName of [
-        "app_update",
-        "app_file_edit",
-        "app_file_write",
-      ]) {
+      // Simulate calling app_refresh by name (as the agent loop does)
+      for (const toolName of ["app_refresh"]) {
         refreshSpy.mockClear();
         broadcastSpy.mockClear();
         updatePublishedSpy.mockClear();
 
         await toolFn(toolName, {
           app_id: "skill-app",
-          path: "f",
-          old_string: "a",
-          new_string: "b",
-          content: "x",
         });
 
         expect(refreshSpy).toHaveBeenCalledTimes(1);
@@ -675,7 +377,15 @@ describe("session-tool-setup app refresh side effects", () => {
         broadcastSpy,
       );
 
-      for (const toolName of ["read_file", "write_file", "shell", "app_list"]) {
+      for (const toolName of [
+        "read_file",
+        "write_file",
+        "shell",
+        "app_list",
+        "app_update",
+        "app_file_edit",
+        "app_file_write",
+      ]) {
         refreshSpy.mockClear();
         broadcastSpy.mockClear();
         updatePublishedSpy.mockClear();
@@ -706,7 +416,7 @@ describe("session-tool-setup app refresh side effects", () => {
       );
 
       // Should not throw even though broadcastToAllClients is undefined
-      const result = await toolFn("app_update", { app_id: "app-no-bc" });
+      const result = await toolFn("app_refresh", { app_id: "app-no-bc" });
 
       expect(result.isError).toBe(false);
       expect(refreshSpy).toHaveBeenCalledTimes(1);

package/src/__tests__/conversation-tool-setup-memory-scope.test.ts

@@ -56,6 +56,7 @@ function makeCtx(overrides: Partial<ToolSetupContext> = {}): ToolSetupContext {
       { surfaceType: SurfaceType; data: SurfaceData; title?: string }
     >(),
     surfaceUndoStacks: new Map(),
+    accumulatedSurfaceState: new Map(),
     surfaceActionRequestIds: new Set<string>(),
     currentTurnSurfaces: [],
     isProcessing: () => false,

package/src/__tests__/conversation-tool-setup-side-effect-flag.test.ts

@@ -56,6 +56,7 @@ function makeCtx(overrides: Partial<ToolSetupContext> = {}): ToolSetupContext {
       { surfaceType: SurfaceType; data: SurfaceData; title?: string }
     >(),
     surfaceUndoStacks: new Map(),
+    accumulatedSurfaceState: new Map(),
     surfaceActionRequestIds: new Set<string>(),
     currentTurnSurfaces: [],
     isProcessing: () => false,

package/src/__tests__/conversation-wipe.test.ts

@@ -26,6 +26,7 @@ mock.module("../util/logger.js", () => ({
 import {
   addMessage,
   createConversation,
+  deleteConversation,
   getConversation,
   getMessages,
   wipeConversation,
@@ -436,3 +437,228 @@ describe("wipeConversation", () => {
     expect(itemBRow).not.toBeNull();
   });
 });
+
+describe("deleteConversation — private scope cleanup", () => {
+  beforeEach(() => {
+    const db = getDb();
+    db.run(`DELETE FROM conversation_starters`);
+    db.run(`DELETE FROM memory_item_sources`);
+    db.run(`DELETE FROM memory_segments`);
+    db.run(`DELETE FROM memory_items`);
+    db.run(`DELETE FROM memory_summaries`);
+    db.run(`DELETE FROM memory_embeddings`);
+    db.run(`DELETE FROM memory_jobs`);
+    db.run(`DELETE FROM tool_invocations`);
+    db.run(`DELETE FROM llm_request_logs`);
+    db.run(`DELETE FROM messages`);
+    db.run(`DELETE FROM conversations`);
+  });
+
+  test("sourceless items cleaned up", () => {
+    const conv = createConversation({ conversationType: "private" });
+    const scopeId = conv.memoryScopeId;
+    const now = Date.now();
+
+    const raw = (
+      getDb() as unknown as {
+        $client: import("bun:sqlite").Database;
+      }
+    ).$client;
+
+    // Insert a memory item with matching scopeId but no memory_item_sources
+    raw
+      .query(
+        `INSERT INTO memory_items (id, status, kind, subject, statement, confidence, fingerprint, scope_id, first_seen_at, last_seen_at)
+         VALUES ('priv-item-1', 'active', 'fact', 'test', 'test fact', 0.8, 'fp-priv-1', ?, ?, ?)`,
+      )
+      .run(scopeId, now, now);
+
+    const result = deleteConversation(conv.id);
+
+    // Item should be gone
+    const itemRow = raw
+      .query("SELECT * FROM memory_items WHERE id = 'priv-item-1'")
+      .get();
+    expect(itemRow).toBeNull();
+
+    // Its ID should be in orphanedItemIds
+    expect(result.orphanedItemIds).toContain("priv-item-1");
+  });
+
+  test("summaries cleaned up", () => {
+    const conv = createConversation({ conversationType: "private" });
+    const scopeId = conv.memoryScopeId;
+    const now = Date.now();
+
+    const raw = (
+      getDb() as unknown as {
+        $client: import("bun:sqlite").Database;
+      }
+    ).$client;
+
+    // Insert a memory summary with matching scopeId
+    raw
+      .query(
+        `INSERT INTO memory_summaries (id, scope, scope_key, summary, token_estimate, version, scope_id, start_at, end_at, created_at, updated_at)
+         VALUES ('priv-sum-1', 'global', 'all', 'private summary', 100, 1, ?, ?, ?, ?, ?)`,
+      )
+      .run(scopeId, now, now, now, now);
+
+    const result = deleteConversation(conv.id);
+
+    // Summary should be gone
+    const summaryRow = raw
+      .query("SELECT * FROM memory_summaries WHERE id = 'priv-sum-1'")
+      .get();
+    expect(summaryRow).toBeNull();
+
+    // Its ID should be in deletedSummaryIds
+    expect(result.deletedSummaryIds).toContain("priv-sum-1");
+  });
+
+  test("standard conversations unaffected", async () => {
+    const conv = createConversation("standard test");
+    const now = Date.now();
+
+    const raw = (
+      getDb() as unknown as {
+        $client: import("bun:sqlite").Database;
+      }
+    ).$client;
+
+    // Insert items with scopeId = "default"
+    raw
+      .query(
+        `INSERT INTO memory_items (id, status, kind, subject, statement, confidence, fingerprint, scope_id, first_seen_at, last_seen_at)
+         VALUES ('default-item-1', 'active', 'fact', 'test', 'test fact', 0.8, 'fp-default', 'default', ?, ?)`,
+      )
+      .run(now, now);
+
+    deleteConversation(conv.id);
+
+    // Default-scope items should still exist
+    const itemRow = raw
+      .query("SELECT * FROM memory_items WHERE id = 'default-item-1'")
+      .get();
+    expect(itemRow).not.toBeNull();
+  });
+
+  test("embeddings cleaned up", () => {
+    const conv = createConversation({ conversationType: "private" });
+    const scopeId = conv.memoryScopeId;
+    const now = Date.now();
+
+    const raw = (
+      getDb() as unknown as {
+        $client: import("bun:sqlite").Database;
+      }
+    ).$client;
+
+    // Insert a memory item with matching scopeId
+    raw
+      .query(
+        `INSERT INTO memory_items (id, status, kind, subject, statement, confidence, fingerprint, scope_id, first_seen_at, last_seen_at)
+         VALUES ('priv-item-emb', 'active', 'fact', 'test', 'test fact', 0.8, 'fp-priv-emb', ?, ?, ?)`,
+      )
+      .run(scopeId, now, now);
+
+    // Insert a corresponding embedding
+    raw
+      .query(
+        `INSERT INTO memory_embeddings (id, target_type, target_id, provider, model, dimensions, created_at, updated_at)
+         VALUES ('emb-priv-item', 'item', 'priv-item-emb', 'test', 'test', 384, ?, ?)`,
+      )
+      .run(now, now);
+
+    deleteConversation(conv.id);
+
+    // Both item and embedding should be deleted
+    const itemRow = raw
+      .query("SELECT * FROM memory_items WHERE id = 'priv-item-emb'")
+      .get();
+    expect(itemRow).toBeNull();
+
+    const embeddingRow = raw
+      .query("SELECT * FROM memory_embeddings WHERE id = 'emb-priv-item'")
+      .get();
+    expect(embeddingRow).toBeNull();
+  });
+
+  test("conversationStarters cleaned up", () => {
+    const conv = createConversation({ conversationType: "private" });
+    const scopeId = conv.memoryScopeId;
+    const now = Date.now();
+
+    const raw = (
+      getDb() as unknown as {
+        $client: import("bun:sqlite").Database;
+      }
+    ).$client;
+
+    // Insert a conversation_starters row with the private scopeId
+    raw
+      .query(
+        `INSERT INTO conversation_starters (id, label, prompt, generation_batch, scope_id, card_type, created_at)
+         VALUES ('starter-1', 'Test starter', 'Tell me about tests', 1, ?, 'chip', ?)`,
+      )
+      .run(scopeId, now);
+
+    // Also insert a default-scope starter that should NOT be deleted
+    raw
+      .query(
+        `INSERT INTO conversation_starters (id, label, prompt, generation_batch, scope_id, card_type, created_at)
+         VALUES ('starter-default', 'Default starter', 'Hello', 1, 'default', 'chip', ?)`,
+      )
+      .run(now);
+
+    deleteConversation(conv.id);
+
+    // Private-scope starter should be gone
+    const starterRow = raw
+      .query("SELECT * FROM conversation_starters WHERE id = 'starter-1'")
+      .get();
+    expect(starterRow).toBeNull();
+
+    // Default-scope starter should still exist
+    const defaultStarterRow = raw
+      .query("SELECT * FROM conversation_starters WHERE id = 'starter-default'")
+      .get();
+    expect(defaultStarterRow).not.toBeNull();
+  });
+
+  test("no duplicate IDs", async () => {
+    const conv = createConversation({ conversationType: "private" });
+    const scopeId = conv.memoryScopeId;
+    const msg = await addMessage(conv.id, "user", "hello");
+    const now = Date.now();
+
+    const raw = (
+      getDb() as unknown as {
+        $client: import("bun:sqlite").Database;
+      }
+    ).$client;
+
+    // Insert a memory item with the private scopeId AND a source linking to the message
+    raw
+      .query(
+        `INSERT INTO memory_items (id, status, kind, subject, statement, confidence, fingerprint, scope_id, first_seen_at, last_seen_at)
+         VALUES ('priv-item-dup', 'active', 'fact', 'test', 'test fact', 0.8, 'fp-priv-dup', ?, ?, ?)`,
+      )
+      .run(scopeId, now, now);
+
+    raw
+      .query(
+        `INSERT INTO memory_item_sources (memory_item_id, message_id, created_at) VALUES ('priv-item-dup', ?, ?)`,
+      )
+      .run(msg.id, now);
+
+    const result = deleteConversation(conv.id);
+
+    // The item ID should appear exactly once in orphanedItemIds (caught by
+    // source-based cleanup, not double-counted by scope sweep).
+    const count = result.orphanedItemIds.filter(
+      (id) => id === "priv-item-dup",
+    ).length;
+    expect(count).toBe(1);
+  });
+});