Add permission rejection test and use codegen'd types for result assertions

Add E2E test verifying one client can reject a permission and both clients
see the denial in the permission.completed event. Regenerate SDK types from
updated runtime schemas so permission.completed includes result.kind enum.
Use type-narrowing filters instead of casts in test assertions.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Author: SteveSandersonMS

nodejs/src/generated/session-events.ts

@@ -2394,6 +2394,20 @@ export type SessionEvent =
          * Request ID of the resolved permission request; clients should dismiss any UI for this request
          */
         requestId: string;
+        /**
+         * The result of the permission request
+         */
+        result: {
+          /**
+           * The outcome of the permission request
+           */
+          kind:
+            | "approved"
+            | "denied-by-rules"
+            | "denied-no-approval-rule-and-could-not-request-from-user"
+            | "denied-interactively-by-user"
+            | "denied-by-content-exclusion-policy";
+        };
       };
     }
   | {

nodejs/test/e2e/multi-client.test.ts

@@ -127,15 +127,70 @@ describe("Multi-client broadcast", async () => {
         expect(client1PermRequested.length).toBeGreaterThan(0);
         expect(client2PermRequested.length).toBeGreaterThan(0);
 
-        // Both clients should have seen permission.completed events
+        // Both clients should have seen permission.completed events with approved result
         const client1PermCompleted = client1Events.filter(
-            (e) => e.type === "permission.completed"
+            (e): e is SessionEvent & { type: "permission.completed" } => e.type === "permission.completed"
         );
         const client2PermCompleted = client2Events.filter(
-            (e) => e.type === "permission.completed"
+            (e): e is SessionEvent & { type: "permission.completed" } => e.type === "permission.completed"
         );
         expect(client1PermCompleted.length).toBeGreaterThan(0);
         expect(client2PermCompleted.length).toBeGreaterThan(0);
+        for (const event of [...client1PermCompleted, ...client2PermCompleted]) {
+            expect(event.data.result.kind).toBe("approved");
+        }
+
+        await session2.destroy();
+    });
+
+    it("one client rejects permission and both see the result", async () => {
+        // Client 1 creates a session and denies all permission requests
+        const session1 = await client1.createSession({
+            onPermissionRequest: () => ({ kind: "denied-interactively-by-user" as const }),
+        });
+
+        // Client 2 resumes — its handler never resolves so only client 1's denial takes effect
+        const session2 = await client2.resumeSession(session1.sessionId, {
+            onPermissionRequest: () => new Promise(() => {}),
+        });
+
+        const client1Events: SessionEvent[] = [];
+        const client2Events: SessionEvent[] = [];
+
+        session1.on((event) => client1Events.push(event));
+        session2.on((event) => client2Events.push(event));
+
+        // Ask the agent to write a file (requires permission)
+        const { writeFile } = await import("fs/promises");
+        const { join } = await import("path");
+        const testFile = join(ctx.workDir, "protected.txt");
+        await writeFile(testFile, "protected content");
+
+        await session1.sendAndWait({
+            prompt: "Edit protected.txt and replace 'protected' with 'hacked'.",
+        });
+
+        // Verify the file was NOT modified (permission was denied)
+        const { readFile } = await import("fs/promises");
+        const content = await readFile(testFile, "utf-8");
+        expect(content).toBe("protected content");
+
+        // Both clients should have seen permission.requested and permission.completed
+        expect(client1Events.filter((e) => e.type === "permission.requested").length).toBeGreaterThan(0);
+        expect(client2Events.filter((e) => e.type === "permission.requested").length).toBeGreaterThan(0);
+
+        // Both clients should see the denial in the completed event
+        const client1PermCompleted = client1Events.filter(
+            (e): e is SessionEvent & { type: "permission.completed" } => e.type === "permission.completed"
+        );
+        const client2PermCompleted = client2Events.filter(
+            (e): e is SessionEvent & { type: "permission.completed" } => e.type === "permission.completed"
+        );
+        expect(client1PermCompleted.length).toBeGreaterThan(0);
+        expect(client2PermCompleted.length).toBeGreaterThan(0);
+        for (const event of [...client1PermCompleted, ...client2PermCompleted]) {
+            expect(event.data.result.kind).toBe("denied-interactively-by-user");
+        }
 
         await session2.destroy();
     });

test/snapshots/multi_client/both_clients_see_tool_request_and_completion_events.yaml

@@ -47,4 +47,4 @@ conversations:
         tool_call_id: toolcall_1
         content: MAGIC_hello_42
       - role: assistant
-        content: "The magic_number tool with seed 'hello' returned: **MAGIC_hello_42**"
+        content: "The magic number tool returned: **MAGIC_hello_42**"

test/snapshots/multi_client/one_client_approves_permission_and_both_see_the_result.yaml

@@ -7,7 +7,7 @@ conversations:
       - role: user
         content: Create a file called hello.txt containing the text 'hello world'
       - role: assistant
-        content: I'll create the hello.txt file with "hello world" content for you.
+        content: I'll create the hello.txt file with the text 'hello world' for you.
       - role: assistant
         tool_calls:
           - id: toolcall_0
@@ -28,7 +28,7 @@ conversations:
       - role: user
         content: Create a file called hello.txt containing the text 'hello world'
       - role: assistant
-        content: I'll create the hello.txt file with "hello world" content for you.
+        content: I'll create the hello.txt file with the text 'hello world' for you.
         tool_calls:
           - id: toolcall_0
             type: function
@@ -47,4 +47,4 @@ conversations:
         tool_call_id: toolcall_1
         content: Created file ${workdir}/hello.txt with 11 characters
       - role: assistant
-        content: Done! I've created hello.txt containing "hello world" in your current directory.
+        content: Done! I've created the file `hello.txt` containing the text 'hello world' in your current directory.

test/snapshots/multi_client/one_client_rejects_permission_and_both_see_the_result.yaml

@@ -0,0 +1,24 @@
+models:
+  - claude-sonnet-4.5
+conversations:
+  - messages:
+      - role: system
+        content: ${system}
+      - role: user
+        content: Edit protected.txt and replace 'protected' with 'hacked'.
+      - role: assistant
+        content: I'll help you edit the file. Let me first view its current contents, then make the change.
+      - role: assistant
+        tool_calls:
+          - id: toolcall_0
+            type: function
+            function:
+              name: report_intent
+              arguments: '{"intent":"Editing protected.txt file"}'
+      - role: assistant
+        tool_calls:
+          - id: toolcall_1
+            type: function
+            function:
+              name: view
+              arguments: '{"path":"${workdir}/protected.txt"}'