Verification & Testing Guide¶

Comprehensive testing guide to verify your n8n MCP Workflow Builder installation is working correctly.

Quick Health Check¶

Run these commands in Claude Desktop to verify basic functionality:

Test 1: MCP Server Connection¶

What MCP tools do you have for n8n?

Expected: List of 17+ n8n workflow management tools

If it fails: See Claude Desktop Integration troubleshooting

Test 2: API Connectivity¶

List my n8n workflows

Expected: List of your workflows (or "No workflows found")

If it fails: Check Configuration

Test 3: Read Operations¶

Show all tags

Expected: List of tags or "No tags found"

If it fails: Verify API key permissions

Comprehensive Testing¶

Test Suite Overview¶

Total Time: 10-17 minutes

Connection Tests¶

Test 1.1: MCP Server Discovery¶

Command:

What n8n tools can you use?

Expected Output: - List of tool categories (Workflow, Execution, Tag, Credential) - Total of 17+ tools - Brief description of each category

Troubleshooting: - If no tools appear → Check Claude Desktop config - If wrong tools appear → Verify server name is unique

Test 1.2: API Connection¶

Command:

List workflows

Expected Output: - Successful API call - Workflow list (may be empty) - No connection errors

Troubleshooting: - Connection refused → Check n8n is running - 401 Unauthorized → Verify API key - 404 Not Found → Check N8N_HOST URL format

Test 1.3: Multi-Instance Routing¶

(Skip if using single instance)

Command:

List workflows from production

Expected Output: - Workflows from production environment - Correct instance identified in response

Troubleshooting: - Wrong instance → Check defaultEnv in .config.json - Instance not found → Verify environment names

Read Operation Tests¶

Test 2.1: List Workflows¶

Command:

List all my workflows

Expected Output:

You have [N] workflows:

1. Workflow Name (ID: 1)
   - Status: Active/Inactive
   - Created: YYYY-MM-DD
   - Nodes: N

[... more workflows ...]

Pass Criteria: - ✅ Correct count - ✅ All workflow names visible - ✅ Status indicators accurate

Test 2.2: Get Workflow Details¶

Command:

Get details for workflow 1

(Replace 1 with an actual workflow ID)

Expected Output:

Workflow: [Name] (ID: 1)
Status: Active/Inactive
Created: YYYY-MM-DD
Updated: YYYY-MM-DD

Nodes (N total):
1. [Node Name] - [Node Type]
2. [Node Name] - [Node Type]
...

Connections:
- [Node A] → [Node B]
...

Pass Criteria: - ✅ Complete workflow information - ✅ All nodes listed - ✅ Connections shown correctly

Test 2.3: List Executions¶

Command:

Show recent executions

Expected Output:

Recent Executions:

1. Workflow: [Name] (ID: X)
   - Execution ID: Y
   - Status: Success/Error/Running
   - Started: YYYY-MM-DD HH:MM:SS
   - Finished: YYYY-MM-DD HH:MM:SS
   - Mode: trigger/manual

[... more executions ...]

Pass Criteria: - ✅ Executions displayed - ✅ Status indicators correct - ✅ Timestamps accurate

Test 2.4: Get Execution Details¶

Command:

Get execution [ID] details

(Replace [ID] with actual execution ID)

Expected Output:

Execution ID: [ID]
Workflow: [Name] (ID: X)
Status: Success/Error
Started: YYYY-MM-DD HH:MM:SS
Finished: YYYY-MM-DD HH:MM:SS
Duration: X.XX seconds

Node Results:
1. [Node Name]: Success
2. [Node Name]: Success
...

Pass Criteria: - ✅ Detailed execution information - ✅ Node-level results shown - ✅ Error details if applicable

Test 2.5: List Tags¶

Command:

List all tags

Expected Output:

Available Tags:

1. production (ID: 1)
   - Workflows: 5
   - Created: YYYY-MM-DD

2. development (ID: 2)
   - Workflows: 3
   - Created: YYYY-MM-DD

[... more tags ...]

Pass Criteria: - ✅ All tags listed - ✅ Workflow counts accurate - ✅ Metadata displayed

Write Operation Tests¶

Test 3.1: Create Tag¶

Command:

Create a tag called "test-tag-verification"

Expected Output:

✅ Successfully created tag!

Tag Details:
- ID: [new ID]
- Name: test-tag-verification
- Created: YYYY-MM-DD

You can now use this tag to organize your workflows.

Pass Criteria: - ✅ Tag created successfully - ✅ New ID assigned - ✅ Name stored correctly

Cleanup:

Delete tag "test-tag-verification"

Test 3.2: Create Simple Workflow¶

Command:

Create a test workflow called "Verification Test Workflow" with:
- A manual trigger
- A Set node that creates a message "Hello World"

Expected Output:

✅ Successfully created workflow!

Workflow Details:
- ID: [new ID]
- Name: Verification Test Workflow
- Status: Inactive
- Nodes: 2
  1. Manual Trigger
  2. Set Node (message)

The workflow has been created but is not active.

Pass Criteria: - ✅ Workflow created - ✅ Both nodes present - ✅ Initially inactive

Save the workflow ID for subsequent tests

Test 3.3: Update Workflow¶

Command:

Update workflow [ID] to add description "This is a test workflow"

(Use ID from previous test)

Expected Output:

✅ Successfully updated workflow!

Updated Fields:
- Description: "This is a test workflow"

Workflow: Verification Test Workflow (ID: [ID])

Pass Criteria: - ✅ Update successful - ✅ Description saved

Test 3.4: Activate Workflow¶

Command:

Activate workflow [ID]

Expected Output:

✅ Workflow "Verification Test Workflow" (ID: [ID]) has been activated!

Note: This workflow has a manual trigger and cannot be executed
automatically. You'll need to trigger it manually in the n8n interface.

Pass Criteria: - ✅ Activation successful - ✅ Status changed to active - ✅ Appropriate trigger warning shown

Test 3.5: Deactivate Workflow¶

Command:

Deactivate workflow [ID]

Expected Output:

✅ Workflow "Verification Test Workflow" (ID: [ID]) has been deactivated!

The workflow will no longer run automatically.

Pass Criteria: - ✅ Deactivation successful - ✅ Status changed to inactive

Test 3.6: Delete Workflow¶

Command:

Delete workflow [ID]

Expected Output:

✅ Successfully deleted workflow "Verification Test Workflow" (ID: [ID])

The workflow and all its execution history have been removed.

Pass Criteria: - ✅ Deletion successful - ✅ Workflow no longer appears in list

Execution Tests¶

Test 4.1: Execute Workflow¶

Command:

Execute workflow [ID] manually

(Use a workflow with non-manual trigger)

Expected Behavior: - Manual execution attempted - Guidance provided if not supported

Note: Most trigger types (except webhook/manual) cannot be executed via API.

Test 4.2: Retry Failed Execution¶

(Skip if no failed executions)

Command:

Retry execution [failed-execution-ID]

Expected Output:

✅ Successfully queued execution retry!

Original Execution: [old-ID]
New Execution: [new-ID]
Status: Running

The workflow will rerun with the same data.

Pass Criteria: - ✅ Retry initiated - ✅ New execution ID created - ✅ Links to original execution

Test 4.3: Delete Execution¶

Command:

Delete execution [old-execution-ID]

Expected Output:

✅ Successfully deleted execution [ID]

The execution record has been removed from history.

Pass Criteria: - ✅ Deletion successful - ✅ Record no longer in list

Multi-Instance Tests¶

(Skip if using single instance)

Test 5.1: List Instances¶

Command:

What n8n instances do you have access to?

Expected Output:

I have access to the following n8n environments:

1. production (default)
   - Host: https://prod.n8n.example.com
   - Status: Connected

2. staging
   - Host: https://staging.n8n.example.com
   - Status: Connected

3. development
   - Host: http://localhost:5678
   - Status: Connected

Pass Criteria: - ✅ All instances listed - ✅ Default marked correctly - ✅ Connection status shown

Test 5.2: Query Specific Instance¶

Command:

List workflows from staging

Expected Output: - Workflows from staging environment - Correct instance identification

Test 5.3: Create in Specific Instance¶

Command:

Create a tag called "multi-instance-test" in development

Expected Output:

✅ Successfully created tag in development instance!

Tag Details:
- ID: [ID]
- Name: multi-instance-test
- Instance: development

Cleanup:

Delete tag "multi-instance-test" from development

Test 5.4: Switch Default Instance¶

(Test configuration only - don't actually change)

Verification:

cat .config.json | grep defaultEnv

Expected: Shows current default environment

Performance Tests¶

Test 6.1: Large Workflow List¶

(If you have 10+ workflows)

Command:

List all workflows with their details

Expected: - Response within 5 seconds - All workflows displayed - No timeout errors

Test 6.2: Execution History¶

Command:

Show last 50 executions

Expected: - Response within 10 seconds - Pagination working correctly - No memory issues

Security Tests¶

Test 7.1: API Key Validation¶

Test with invalid key:

1. Temporarily change API key in configuration
2. Restart Claude Desktop
3. Try: List workflows

Expected: Authorization error (401)

Restore correct API key after test

Test 7.2: Permission Scope¶

Command:

Get credential schema for httpBasicAuth

Expected: - Schema returned successfully - Security note about credential protection

Note: Actual credential data should never be returned by API

Test Results Summary¶

Use this checklist to track your test results:

Connection Tests¶

• 1.1 MCP Server Discovery
• 1.2 API Connection
• 1.3 Multi-Instance Routing (optional)

Read Operations¶

• 2.1 List Workflows
• 2.2 Get Workflow Details
• 2.3 List Executions
• 2.4 Get Execution Details
• 2.5 List Tags

Write Operations¶

• 3.1 Create Tag
• 3.2 Create Workflow
• 3.3 Update Workflow
• 3.4 Activate Workflow
• 3.5 Deactivate Workflow
• 3.6 Delete Workflow

Execution Tests¶

• 4.1 Execute Workflow
• 4.2 Retry Execution (if applicable)
• 4.3 Delete Execution

Multi-Instance (if applicable)¶

• 5.1 List Instances
• 5.2 Query Specific Instance
• 5.3 Create in Specific Instance
• 5.4 Verify Default Instance

Performance¶

• 6.1 Large Workflow List
• 6.2 Execution History

Security¶

• 7.1 API Key Validation
• 7.2 Permission Scope

Automated Testing¶

For automated testing, use the test scripts:

# Run comprehensive MCP tool tests
node test-mcp-tools.js

# Expected output:
# ✅ Workflow Tests: 8/8 passed
# ✅ Tag Tests: 5/5 passed
# ✅ Execution Tests: 4/4 passed
# ✅ Total: 17/17 tests passed

Common Issues¶

Issue: "No tools available"¶

Cause: MCP server not configured correctly

Solution: Reconfigure Claude Desktop Integration

Issue: "Connection refused"¶

Cause: n8n instance not running or unreachable

Solution: 1. Check n8n is running: curl https://your-n8n-host.com 2. Verify firewall settings 3. Test API manually

Issue: "401 Unauthorized"¶

Cause: Invalid or expired API key

Solution: 1. Regenerate API key in n8n 2. Update configuration 3. Restart Claude Desktop

Issue: "Slow responses"¶

Cause: Large workflow/execution datasets or network latency

Solution: 1. Use filters to limit results 2. Paginate large lists 3. Check network connection 4. Consider local n8n instance for development

Next Steps¶

After completing verification:

1. Workflow Examples - Build real workflows
2. API Reference - Deep dive into all tools
3. Features Guide - Explore all capabilities