@camunda/e2e-test-suite 0.0.248 → 0.0.249

package/README.md

@@ -70,7 +70,6 @@ For C8Run, a test run can be manually triggered for specific minor versions, for
 - [C8Run On-Demand with macOS GitHub Action](https://github.com/camunda/c8-cross-component-e2e-tests/actions/workflows/playwright_c8Run_tests_manual_mac.yml)
 - [C8Run On-Demand with Windows GitHub Action](https://github.com/camunda/c8-cross-component-e2e-tests/actions/workflows/playwright_c8Run_tests_manual_windows.yml)
 
-
 ## Prerequisites
 
 1. Before running the test suite, make sure you have the following prerequisites installed:
@@ -133,6 +132,7 @@ Request the creation of an organisation identical to `Automation QA`
 **Environment:** `INT`
 
 **Requirements:**
+
 - You must be set as the **Owner**
 - Name the organisation: `Automation QA - Your_Name`
 
@@ -160,22 +160,21 @@ Within the `Automation QA - Your_Name` organisation, invite the following users:
    - `C8_PASSWORD`
    - `C8_PASSWORD_TEST`
 
-
 ## Running Smoke Tests in CI
 
 ### SaaS On-Demand Generation Creation Smoke
 
-To trigger the SaaS smoke tests in CI, use the workflow:  
+To trigger the SaaS smoke tests in CI, use the workflow:
 
 [SaaS On-Demand Generation Creation Smoke](https://github.com/camunda/c8-cross-component-e2e-tests/actions/workflows/playwright_saas_manual_generation_creation_smoke.yml)
 
 - You’ll be prompted to input the version you want to test for each component.
 - The **minor version** is simply the minor version your PR is targeting.
-  - *Example:* If your PR targets `8.9`, then the minor version is `8.9`.
+  - _Example:_ If your PR targets `8.9`, then the minor version is `8.9`.
 
 ### Helm Chart Smoke Tests (Self-Managed)
 
-For Helm chart tests (if you want to run only the smoke test), use the workflow:  
+For Helm chart tests (if you want to run only the smoke test), use the workflow:
 
 [SM On-Demand Smoke 8.8+](https://github.com/camunda/c8-cross-component-e2e-tests/actions/workflows/playwright_sm_manual_smoke_8_8_plus.yml)
 
@@ -208,6 +207,7 @@ npx playwright test --ui
 ### Additional Notes for Local Testing
 
 #### C8 SaaS
+
 For a full test run, no additional configuration is required. However, if you are running a single test, ensure you have a healthy and active cluster.
 
 If you want to run tests against the production environment, add the following to your local .env file:
@@ -217,6 +217,7 @@ IS_PROD=true
 ```
 
 #### C8 Self-Managed (SM) with Helm
+
 To test locally on the SM Helm environment, you need an active SM instance. You can create one through any of the following GitHub Actions workflows:
 
 - [Basic Helm Chart for versions ≤ 8.7.X](https://github.com/camunda/c8-cross-component-e2e-tests/actions/workflows/playwright_sm_manual_no_tests.yml)
@@ -237,17 +238,20 @@ If you want to run tests with Resource-Based Access (RBA) enabled, add the follo
 ```bash
 IS_RBA=true
 ```
+
 If you want to run tests with Multi-Tenancy (MT) enabled, add the following to your .env file, otherwise SM tests default to MT disabled:
 
 ```bash
 IS_MT=true
 ```
+
 #### C8 Run
+
 To test locally against C8 Run, you must set up a local instance of C8 Run. Please follow the instructions provided in the [C8 Run README.md](https://github.com/camunda/camunda/blob/main/c8run/README.md). By default, the BASE_URL is localhost:8080, however if this is configured as something differently, tests can be run against this by inputting the new BASE_URL into your `.env` file under the `PLAYWRIGHT_BASE_URL` variable.
 
 ## Running MCP Tests (c8Run Only)
 
-The MCP (Model Context Protocol) test suite requires the MCP test server to be running locally. These tests are designed for **c8Run environments only**, as they require localhost access to the MCP server.
+The MCP (Model Context Protocol) test suite requires the MCP test server to be running locally. These tests are designed for **c8Run environments only** (version 8.9+), as they require localhost access to the MCP server.
 
 The MCP test server is available as a Docker image from the [camunda/mcp](https://github.com/camunda/mcp) repository.
 
@@ -257,6 +261,15 @@ To run MCP tests locally with c8Run, you need to start both the MCP test server
 
 1. Start the MCP server:
 
+   **macOS/Windows:**
+   ```bash
+   docker run -d \
+     --name mcp-test-server \
+     -p 12001:12001 \
+     registry.camunda.cloud/mcp/mcp-test-server:latest
+   ```
+
+   **Linux:**
    ```bash
    docker run -d \
      --name mcp-test-server \
@@ -266,10 +279,10 @@ To run MCP tests locally with c8Run, you need to start both the MCP test server
 
 2. Start c8Run locally (follow [c8Run setup instructions](https://github.com/camunda/camunda/blob/main/c8run/README.md))
 
-3. Run the MCP tests:
+3. Run the MCP tests (the `MCP_SERVER_AVAILABLE` env var is required to enable them):
 
    ```bash
-   npx playwright test --project=chromium-v2 tests/c8Run-8.9/mcp-user-flows.spec.ts 
+   MCP_SERVER_AVAILABLE=true npx playwright test --project=chromium-v2 tests/c8Run-8.9/mcp-user-flows.spec.ts
    ```
 
 4. Stop the MCP server when done:
@@ -291,6 +304,32 @@ You can override the MCP server URL by setting the `MCP_SERVER_URL` environment
 MCP_SERVER_URL=http://localhost:12001
 ```
 
+### Available MCP Operations
+
+The comprehensive MCP test suite validates all 7 operations of the MCP Remote Client connector:
+
+#### Tools Operations
+
+- **tools/list**: Lists all available MCP tools (e.g., `greet`, `add`, `echo`)
+- **tools/call**: Executes a specific tool with arguments (tests `greet` tool with `name: "John"`)
+
+#### Resources Operations
+
+- **resources/list**: Lists available MCP resources (tests `file://jokes-guide.md`)
+- **resources/templates/list**: Lists resource templates (validates empty array response)
+- **resources/read**: Reads content from a specific resource URI (`file://jokes-guide.md`)
+
+#### Prompts Operations
+
+- **prompts/list**: Lists available MCP prompts (tests `get-greeting` prompt)
+- **prompts/get**: Retrieves a specific prompt with arguments (tests `get-greeting` with `name: "John"`)
+
+Each operation stores its result in a dedicated process variable:
+
+- `listToolsResult`, `callToolResult`, `listResourcesResult`, `listResourcesTemplateResult`, `readResourceResult`, `listPromptsResult`, `getPromptResult`
+
+The test suite validates the JSON structure and content of each operation's response to ensure complete connector functionality.
+
 ### Available MCP Tools
 
 The MCP test server provides the following tools for testing:
@@ -303,7 +342,20 @@ For more details about the MCP server and its capabilities, see the [MCP reposit
 
 ### CI/CD Integration
 
-In c8Run CI/CD pipelines (Linux, macOS, Windows), the MCP server is automatically started before running tests and stopped after tests complete. No additional configuration is required.
+In CI/CD pipelines, the MCP server runs **only on Linux (`ubuntu-latest`)** and only under the following conditions:
+
+- c8Run version **>= 8.9**
+- Tasklist **v2** mode (not v1)
+- **Detached mode** only (not Docker Compose mode)
+
+The `MCP_SERVER_AVAILABLE` environment variable is set to `true` when the server starts successfully. The test suite checks this variable and skips all MCP tests if it is not set, so runs on macOS, Windows, or unsupported versions will skip gracefully rather than fail.
+
+**Why macOS and Windows are excluded from CI:**
+
+- **macOS runners** (`macos-latest`, `macos-14`): GitHub Actions macOS runners do not have Docker pre-installed, so `docker run` fails with `docker: command not found`.
+- **Windows runners**: The MCP test server image (`registry.camunda.cloud/mcp/mcp-test-server:latest`) has no `windows/amd64` manifest, so `docker run` fails with "no matching manifest for windows/amd64".
+
+MCP tests can still be run locally on macOS and Windows using the instructions above.
 
 ## Automated Nightly Test Runs
 
@@ -331,34 +383,37 @@ login.spec.ts - Tests related to the login functionality using the LoginPage obj
 
 Feel free to add more test files or page objects to expand the test coverage based on your specific requirements.
 
-## GitHub Actions Workflow Naming Convention  
+## GitHub Actions Workflow Naming Convention
+
+To maintain consistency and clarity in workflow names displayed in the GitHub Actions UI, we use the following naming pattern:
 
-To maintain consistency and clarity in workflow names displayed in the GitHub Actions UI, we use the following naming pattern:  
+### **Naming Pattern:**
 
-### **Naming Pattern:**  
-`<Distribution> <Trigger> <Scope (if needed)>`  
+`<Distribution> <Trigger> <Scope (if needed)>`
 
-- **Distribution:**  
-  - `SaaS` → Cloud-based environment  
-  - `SM` → Self-Managed using the Helm Chart 
-  - `C8Run` → C8Run environment  
+- **Distribution:**
 
-- **Trigger:**  
-  - `Nightly` → Runs automatically every night  
-  - `PR` → Runs when a pull request is opened/updated  
-  - `On-Demand` → Runs when triggered via **workflow dispatch** 
+  - `SaaS` → Cloud-based environment
+  - `SM` → Self-Managed using the Helm Chart
+  - `C8Run` → C8Run environment
 
-- **Scope** (optional):  
-  - `Linux`, `Connectors`,  `OpenSearch`,  `No Tests` etc.  
+- **Trigger:**
 
-### **Examples**  
+  - `Nightly` → Runs automatically every night
+  - `PR` → Runs when a pull request is opened/updated
+  - `On-Demand` → Runs when triggered via **workflow dispatch**
 
-| Workflow Name | Meaning |
-|--------------|---------|
-| `SaaS Nightly 8.8` | Runs SaaS tests nightly against the latest 8.8 SNAPSHOT |
-| `SM Nightly Chrome 8.7 OpenSearch` | Runs SM tests nightly against the latest 8.7 SNAPSHOT in Chrome with OpenSearch |
-| `SaaS/SM/C8Run PR` | Runs SaaS or Self-Manged or C8Run tests on PR updates |
-| `SM On-Demand MT No Tests` | Creates a Self-Managed environment with Multi-Tenancy Enabled via workflow dispatch with no test run|
+- **Scope** (optional):
+  - `Linux`, `Connectors`, `OpenSearch`, `No Tests` etc.
+
+### **Examples**
+
+| Workflow Name                      | Meaning                                                                                              |
+| ---------------------------------- | ---------------------------------------------------------------------------------------------------- |
+| `SaaS Nightly 8.8`                 | Runs SaaS tests nightly against the latest 8.8 SNAPSHOT                                              |
+| `SM Nightly Chrome 8.7 OpenSearch` | Runs SM tests nightly against the latest 8.7 SNAPSHOT in Chrome with OpenSearch                      |
+| `SaaS/SM/C8Run PR`                 | Runs SaaS or Self-Manged or C8Run tests on PR updates                                                |
+| `SM On-Demand MT No Tests`         | Creates a Self-Managed environment with Multi-Tenancy Enabled via workflow dispatch with no test run |
 
 ---
 
@@ -366,18 +421,17 @@ By following this convention, we ensure **clarity, consistency, and alignment wi
 
 ### SM Possible Directory Names
 
-| Directory Name                  | Meaning                                                                                                      |
-|---------------------------------|--------------------------------------------------------------------------------------------------------------|
-| `camunda-platform-8.9-latest`  | Runs SM tests against the latest 8.9 SNAPSHOT                                                               |
-| `camunda-platform-8.8-latest`  | Runs SM tests against the latest 8.8 SNAPSHOT                                                               |
-| `camunda-platform-8.7-latest`  | Runs SM tests against the latest 8.7 SNAPSHOT                                                               |
-| `camunda-platform-8.6-latest`  | Runs SM tests against the latest 8.6 SNAPSHOT                                                               |
-| `camunda-platform-8.9`         | Runs SM tests against the 8.9 directory. This is where the `release-candidate-8.9` branch is located       |
-| `camunda-platform-8.8`         | Runs SM tests against the 8.8 directory. This is where the `release-candidate-8.8` branch is located       |
-| `camunda-platform-8.7`         | Runs SM tests against the 8.7 directory. This is where the `release-candidate-8.7` branch is located       |
-| `camunda-platform-8.6`         | Runs SM tests against the 8.6 directory. This is where the `release-candidate-8.6` branch is located       |
-| `all-latest`                    | Runs SM tests against all latest versions                                                                     |
-
+| Directory Name                | Meaning                                                                                              |
+| ----------------------------- | ---------------------------------------------------------------------------------------------------- |
+| `camunda-platform-8.9-latest` | Runs SM tests against the latest 8.9 SNAPSHOT                                                        |
+| `camunda-platform-8.8-latest` | Runs SM tests against the latest 8.8 SNAPSHOT                                                        |
+| `camunda-platform-8.7-latest` | Runs SM tests against the latest 8.7 SNAPSHOT                                                        |
+| `camunda-platform-8.6-latest` | Runs SM tests against the latest 8.6 SNAPSHOT                                                        |
+| `camunda-platform-8.9`        | Runs SM tests against the 8.9 directory. This is where the `release-candidate-8.9` branch is located |
+| `camunda-platform-8.8`        | Runs SM tests against the 8.8 directory. This is where the `release-candidate-8.8` branch is located |
+| `camunda-platform-8.7`        | Runs SM tests against the 8.7 directory. This is where the `release-candidate-8.7` branch is located |
+| `camunda-platform-8.6`        | Runs SM tests against the 8.6 directory. This is where the `release-candidate-8.6` branch is located |
+| `all-latest`                  | Runs SM tests against all latest versions                                                            |
 
 ### c8Run Supported Versions, Modes and Projects
 | Version | Docker Allowed? | OS When Docker | OS When Non-Docker | Projects (Non-Docker) | Projects (Docker) |
@@ -409,7 +463,7 @@ To record video, you need to set the following config in the `use` block in the
 
 ```typescript
 use: {
-  video: 'retain-on-failure'
+  video: 'retain-on-failure';
 }
 ```
 
@@ -429,9 +483,8 @@ To contribute, please follow these guidelines:
 
 ### TestRail Integration
 
-To ensure our test case documentation on TestRail remains up-to-date, if a PR contains updates to any test file or page file, the associated TestRail test case must be linked in the PR description. 
+To ensure our test case documentation on TestRail remains up-to-date, if a PR contains updates to any test file or page file, the associated TestRail test case must be linked in the PR description.
 
 If you do not have access to TestRail, please contact the Test Automation Team to request access.
 
-
 Thank you for using the C8 Cross-Component End-to-End Test Suite for Camunda C8 full product stack e2e testing. Happy testing! If you have any questions or need assistance, feel free to reach out to the maintainers for support.

package/dist/pages/c8Run-8.9/OperateProcessInstancePage.d.ts

@@ -15,5 +15,13 @@ declare class OperateProcessInstancePage {
     completedIconAssertion(): Promise<void>;
     activeIconAssertion(): Promise<void>;
     assertProcessVariableContainsText(variableName: string, text: string): Promise<void>;
+    /**
+     * Gets the JSON value of a variable by name
+     */
+    getVariableJsonValue(variableName: string): Promise<unknown>;
+    /**
+     * Verifies that a variable exists and contains expected JSON content
+     */
+    verifyVariableJsonContent(variableName: string, assertions: (value: unknown) => void | Promise<void>): Promise<void>;
 }
 export { OperateProcessInstancePage };

package/dist/pages/c8Run-8.9/OperateProcessInstancePage.js

@@ -94,5 +94,27 @@ class OperateProcessInstancePage {
         }
         throw new Error(`Failed to assert variable ${variableName} contains ${text} after ${maxRetries} attempts.`);
     }
+    /**
+     * Gets the JSON value of a variable by name
+     */
+    async getVariableJsonValue(variableName) {
+        const variable = await this.getProcessVariableValue(variableName);
+        await (0, test_1.expect)(variable).toBeVisible();
+        const valueText = (await variable.textContent()) || '{}';
+        try {
+            return JSON.parse(valueText);
+        }
+        catch (error) {
+            throw new Error(`Failed to parse JSON for variable "${variableName}". ` +
+                `Raw value: ${valueText}. Original error: ${String(error)}`);
+        }
+    }
+    /**
+     * Verifies that a variable exists and contains expected JSON content
+     */
+    async verifyVariableJsonContent(variableName, assertions) {
+        const value = await this.getVariableJsonValue(variableName);
+        await assertions(value);
+    }
 }
 exports.OperateProcessInstancePage = OperateProcessInstancePage;

package/dist/tests/c8Run-8.9/mcp-user-flows.spec.js

@@ -1,42 +1,108 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 const c8Run_8_9_1 = require("../../fixtures/c8Run-8.9");
+const test_1 = require("@playwright/test");
 const UtilitiesPage_1 = require("../../pages/c8Run-8.9/UtilitiesPage");
 const _setup_1 = require("../../test-setup.js");
 const apiHelpers_1 = require("../../utils/apiHelpers");
 const zeebeClient_1 = require("../../utils/zeebeClient");
 const sleep_1 = require("../../utils/sleep");
-if (process.env.RUN_AS_DOCKER_COMPOSE != 'true') {
-    //MCP server is not configured with Docker
-    c8Run_8_9_1.test.beforeAll(async () => {
-        await (0, apiHelpers_1.validateMcpServerHealth)();
-        await (0, zeebeClient_1.deploy)('./resources/mcp_server/mcp_server_list_tools.bpmn');
-        await (0, zeebeClient_1.createInstances)('mcp_remote_client', 1, 1);
-        await (0, sleep_1.sleep)(5000);
-    });
+c8Run_8_9_1.test.beforeAll(async () => {
+    if (process.env.MCP_SERVER_AVAILABLE !== 'true') {
+        return;
+    }
+    await (0, apiHelpers_1.validateMcpServerHealth)();
+    await (0, zeebeClient_1.deploy)(['./resources/mcp_server/mcp_remote_client_operations.bpmn']);
+    await (0, zeebeClient_1.createInstances)('mcp_remote_client', 1, 1);
+    await (0, sleep_1.sleep)(2000);
+});
+c8Run_8_9_1.test.describe('@tasklistV2 MCP Client connector tests', () => {
+    c8Run_8_9_1.test.skip(process.env.MCP_SERVER_AVAILABLE !== 'true', 'MCP server is only available on Linux and macOS CI runners');
     c8Run_8_9_1.test.beforeEach(async ({ page, operateLoginPage, operateHomePage }) => {
         await (0, UtilitiesPage_1.navigateToApp)(page, 'operate');
         await operateLoginPage.login('demo', 'demo');
         await operateHomePage.operateBannerIsVisible();
     });
-    c8Run_8_9_1.test.describe('@tasklistV2 MCP Test Server User Flows', () => {
-        c8Run_8_9_1.test.afterEach(async ({ page }, testInfo) => {
-            await (0, _setup_1.captureScreenshot)(page, testInfo);
-            await (0, _setup_1.captureFailureVideo)(page, testInfo);
-        });
-        (0, c8Run_8_9_1.test)('MCP server returns list of tools from process', async ({ operateHomePage, operateProcessInstancePage, operateProcessesPage, }) => {
-            c8Run_8_9_1.test.slow();
-            await c8Run_8_9_1.test.step('User can access the completed process on operate', async () => {
-                await operateHomePage.clickProcessesTab();
-                await operateProcessesPage.clickProcessCompletedCheckbox();
-                await operateProcessesPage.clickProcessInstanceLink('mcp_remote_client');
-                await operateProcessInstancePage.completedIconAssertion();
+    c8Run_8_9_1.test.afterEach(async ({ page }, testInfo) => {
+        await (0, _setup_1.captureScreenshot)(page, testInfo);
+        await (0, _setup_1.captureFailureVideo)(page, testInfo);
+    });
+    (0, c8Run_8_9_1.test)('As an user I can invoke the MCP Client connector and use all its operations', async ({ operateHomePage, operateProcessesPage, operateProcessInstancePage, }) => {
+        c8Run_8_9_1.test.slow();
+        await c8Run_8_9_1.test.step('Navigate to completed process instance', async () => {
+            await operateHomePage.clickProcessesTab();
+            await operateProcessesPage.clickProcessCompletedCheckbox();
+            await operateProcessesPage.clickProcessInstanceLink('mcp_remote_client');
+            await operateProcessInstancePage.completedIconAssertion();
+        });
+        await c8Run_8_9_1.test.step('Verify listToolsResult variable contains tools', async () => {
+            await operateProcessInstancePage.verifyVariableJsonContent('listToolsResult', (value) => {
+                const data = value;
+                (0, test_1.expect)(data.toolDefinitions).toBeDefined();
+                (0, test_1.expect)(data.toolDefinitions.length).toBeGreaterThan(0);
+                const greetTool = data.toolDefinitions.find((tool) => tool.name === 'greet');
+                (0, test_1.expect)(greetTool).toBeDefined();
+                (0, test_1.expect)(greetTool?.description).toBe('Return a nice greeting message');
+            });
+        });
+        await c8Run_8_9_1.test.step('Verify callToolResult variable from greet tool call', async () => {
+            await operateProcessInstancePage.verifyVariableJsonContent('callToolResult', (value) => {
+                const data = value;
+                (0, test_1.expect)(data.name).toBe('greet');
+                (0, test_1.expect)(data.isError).toBe(false);
+                (0, test_1.expect)(data.content).toBeDefined();
+                (0, test_1.expect)(data.content[0].type).toBe('text');
+                (0, test_1.expect)(data.content[0].text).toContain('Hello, John');
+            });
+        });
+        await c8Run_8_9_1.test.step('Verify listResourcesResult variable contains resources', async () => {
+            await operateProcessInstancePage.verifyVariableJsonContent('listResourcesResult', (value) => {
+                const data = value;
+                (0, test_1.expect)(data.resources).toBeDefined();
+                (0, test_1.expect)(data.resources.length).toBeGreaterThan(0);
+                const jokesResource = data.resources.find((resource) => resource.uri === 'file://jokes-guide.md');
+                (0, test_1.expect)(jokesResource).toBeDefined();
+                (0, test_1.expect)(jokesResource?.name).toBe('Jokes Guideline');
+                (0, test_1.expect)(jokesResource?.mimeType).toBe('text/markdown');
             });
-            await c8Run_8_9_1.test.step('Verify MCP tools are returned in the process result', async () => {
-                await operateProcessInstancePage.assertProcessVariableContainsText('toolCallResult', 'add');
-                await operateProcessInstancePage.assertProcessVariableContainsText('toolCallResult', 'greet');
-                await operateProcessInstancePage.assertProcessVariableContainsText('toolCallResult', 'echo');
+        });
+        await c8Run_8_9_1.test.step('Verify listResourcesTemplateResult variable contains resource templates', async () => {
+            await operateProcessInstancePage.verifyVariableJsonContent('listResourcesTemplateResult', (value) => {
+                const data = value;
+                (0, test_1.expect)(data.resourceTemplates).toBeDefined();
+                (0, test_1.expect)(data.resourceTemplates).toEqual([]);
+            });
+        });
+        await c8Run_8_9_1.test.step('Verify readResourceResult variable from jokes-guide.md', async () => {
+            await operateProcessInstancePage.verifyVariableJsonContent('readResourceResult', (value) => {
+                const data = value;
+                (0, test_1.expect)(data.contents).toBeDefined();
+                (0, test_1.expect)(data.contents.length).toBeGreaterThan(0);
+                (0, test_1.expect)(data.contents[0].uri).toBe('file://jokes-guide.md');
+                (0, test_1.expect)(data.contents[0].mimeType).toBe('text/markdown');
+                (0, test_1.expect)(data.contents[0].text).toContain('How to Write Jokes');
+            });
+        });
+        await c8Run_8_9_1.test.step('Verify listPromptsResult variable contains prompts', async () => {
+            await operateProcessInstancePage.verifyVariableJsonContent('listPromptsResult', (value) => {
+                const data = value;
+                (0, test_1.expect)(data.promptDescriptions).toBeDefined();
+                (0, test_1.expect)(data.promptDescriptions.length).toBeGreaterThan(0);
+                const greetingPrompt = data.promptDescriptions.find((prompt) => prompt.name === 'get-greeting');
+                (0, test_1.expect)(greetingPrompt).toBeDefined();
+                (0, test_1.expect)(greetingPrompt?.description).toContain('personalized greeting');
+            });
+        });
+        await c8Run_8_9_1.test.step('Verify getPromptResult variable from get-greeting prompt', async () => {
+            await operateProcessInstancePage.verifyVariableJsonContent('getPromptResult', (value) => {
+                const data = value;
+                (0, test_1.expect)(data.description).toBe('Greeting Prompt');
+                (0, test_1.expect)(data.messages).toBeDefined();
+                (0, test_1.expect)(data.messages.length).toBeGreaterThan(0);
+                (0, test_1.expect)(data.messages[0].role).toBe('user');
+                (0, test_1.expect)(data.messages[0].content.text).toContain('John');
+                (0, test_1.expect)(data.messages[0].content.text).toContain('greet tool');
             });
         });
     });
-}
+});

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@camunda/e2e-test-suite",
-  "version": "0.0.248",
+  "version": "0.0.249",
   "description": "End-to-end test helpers for Camunda 8",
   "repository": {
     "type": "git",