fix: AI node connection validation in partial workflow updates (#357) (#358)

* fix: AI node connection validation in partial workflow updates (#357)

Fix critical validation issue where n8n_update_partial_workflow incorrectly
required 'main' connections for AI nodes that exclusively use AI-specific
connection types (ai_languageModel, ai_memory, ai_embedding, ai_vectorStore, ai_tool).

Problem:
- Workflows containing AI nodes could not be updated via n8n_update_partial_workflow
- Validation incorrectly expected ALL nodes to have 'main' connections
- AI nodes only have AI-specific connection types, never 'main'

Root Cause:
- Zod schema in src/services/n8n-validation.ts defined 'main' as required field
- Schema didn't support AI-specific connection types

Fixed:
- Made 'main' connection optional in Zod schema
- Added support for all AI connection types: ai_tool, ai_languageModel, ai_memory,
  ai_embedding, ai_vectorStore
- Created comprehensive test suite (13 tests) covering all AI connection scenarios
- Updated documentation to clarify AI nodes don't require 'main' connections

Testing:
- All 13 new integration tests passing
- Tested with actual workflow 019Vrw56aROeEzVj from issue #357
- Zero breaking changes (making required fields optional is always safe)

Files Changed:
- src/services/n8n-validation.ts - Fixed Zod schema
- tests/integration/workflow-diff/ai-node-connection-validation.test.ts - New test suite
- src/mcp/tool-docs/workflow_management/n8n-update-partial-workflow.ts - Updated docs
- package.json - Version bump to 2.21.1
- CHANGELOG.md - Comprehensive release notes

Closes #357

🤖 Generated with Claude Code (https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

Conceived by Romuald Członkowski - www.aiadvisors.pl/en

* fix: Add missing id parameter in test file and JSDoc comment

Address code review feedback from PR #358:
- Add 'id' field to all applyDiff calls in test file (fixes TypeScript errors)
- Add JSDoc comment explaining why 'main' is optional in schema
- Ensures TypeScript compilation succeeds

Changes:
- tests/integration/workflow-diff/ai-node-connection-validation.test.ts:
  Added id parameter to all 13 test cases
- src/services/n8n-validation.ts:
  Added JSDoc explaining optional main connections

Testing:
- npm run typecheck: PASS ✅
- npm run build: PASS ✅
- All 13 tests: PASS ✅

🤖 Generated with Claude Code (https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>

CHANGELOG.md

@@ -7,6 +7,139 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.21.1] - 2025-10-23
+
+### 🐛 Bug Fixes
+
+**Issue #357: Fix AI Node Connection Validation in Partial Workflow Updates**
+
+Fixed critical validation issue where `n8n_update_partial_workflow` incorrectly required `main` connections for AI nodes that exclusively use AI-specific connection types (`ai_languageModel`, `ai_memory`, `ai_embedding`, `ai_vectorStore`, `ai_tool`).
+
+#### Problem
+
+Workflows containing AI nodes (OpenAI Chat Model, Postgres Chat Memory, Embeddings OpenAI, Supabase Vector Store) could not be updated via `n8n_update_partial_workflow`, even for trivial changes to unrelated nodes. The validation logic incorrectly expected ALL nodes to have `main` connections, causing false positive errors:
+
+```
+Invalid connections: [
+  {
+    "code": "invalid_type",
+    "expected": "array",
+    "received": "undefined",
+    "path": ["OpenAI Chat Model", "main"],
+    "message": "Required"
+  }
+]
+```
+
+**Impact**: Users could not update any workflows containing AI Agent nodes via MCP tools, forcing manual updates through the n8n UI.
+
+#### Root Cause
+
+The Zod schema in `src/services/n8n-validation.ts` (lines 27-39) defined `main` connections as a **required field** for all nodes, without support for AI-specific connection types:
+
+```typescript
+// BEFORE (Broken):
+export const workflowConnectionSchema = z.record(
+  z.object({
+    main: z.array(...), // Required - WRONG for AI nodes!
+  })
+);
+```
+
+AI nodes use specialized connection types exclusively:
+- **ai_languageModel** - Language models (OpenAI, Anthropic, etc.)
+- **ai_memory** - Memory systems (Postgres Chat Memory, etc.)
+- **ai_embedding** - Embedding models (Embeddings OpenAI, etc.)
+- **ai_vectorStore** - Vector stores (Supabase Vector Store, etc.)
+- **ai_tool** - Tools for AI agents
+
+These nodes **never have `main` connections** - they only have their AI-specific connection types.
+
+#### Fixed
+
+**1. Updated Zod Schema** (`src/services/n8n-validation.ts` lines 27-49):
+```typescript
+// AFTER (Fixed):
+const connectionArraySchema = z.array(
+  z.array(
+    z.object({
+      node: z.string(),
+      type: z.string(),
+      index: z.number(),
+    })
+  )
+);
+
+export const workflowConnectionSchema = z.record(
+  z.object({
+    main: connectionArraySchema.optional(),              // Now optional
+    error: connectionArraySchema.optional(),              // Error connections
+    ai_tool: connectionArraySchema.optional(),            // AI tool connections
+    ai_languageModel: connectionArraySchema.optional(),   // Language model connections
+    ai_memory: connectionArraySchema.optional(),          // Memory connections
+    ai_embedding: connectionArraySchema.optional(),       // Embedding connections
+    ai_vectorStore: connectionArraySchema.optional(),     // Vector store connections
+  })
+);
+```
+
+**2. Comprehensive Test Suite** (New file: `tests/integration/workflow-diff/ai-node-connection-validation.test.ts`):
+- 13 test scenarios covering all AI connection types
+- Tests for AI nodes with ONLY AI-specific connections (no `main`)
+- Tests for mixed workflows (regular nodes + AI nodes)
+- Tests for the exact scenario from issue #357
+- All tests passing ✅
+
+**3. Updated Documentation** (`src/mcp/tool-docs/workflow_management/n8n-update-partial-workflow.ts`):
+- Added clarification that AI nodes do NOT require `main` connections
+- Documented fix for issue #357
+- Updated best practices for AI workflows
+
+#### Testing
+
+**Before Fix**:
+- ❌ `n8n_validate_workflow`: Returns `valid: true` (correct)
+- ❌ `n8n_update_partial_workflow`: FAILS with "main connections required" errors
+- ❌ Cannot update workflows containing AI nodes at all
+
+**After Fix**:
+- ✅ `n8n_validate_workflow`: Returns `valid: true` (still correct)
+- ✅ `n8n_update_partial_workflow`: SUCCEEDS without validation errors
+- ✅ AI nodes correctly recognized with AI-specific connection types only
+- ✅ All 13 new integration tests passing
+- ✅ Tested with actual workflow `019Vrw56aROeEzVj` from issue #357
+
+#### Impact
+
+**Zero Breaking Changes**:
+- Making required fields optional is always backward compatible
+- All existing workflows continue working
+- Validation now correctly matches n8n's actual connection model
+
+**Fixes**:
+- Users can now update AI workflows via `n8n_update_partial_workflow`
+- AI nodes no longer generate false positive validation errors
+- Consistent validation between `n8n_validate_workflow` and `n8n_update_partial_workflow`
+
+#### Files Changed
+
+**Modified (3 files)**:
+- `src/services/n8n-validation.ts` - Fixed Zod schema to support all connection types
+- `src/mcp/tool-docs/workflow_management/n8n-update-partial-workflow.ts` - Updated documentation
+- `package.json` - Version bump to 2.21.1
+
+**Added (1 file)**:
+- `tests/integration/workflow-diff/ai-node-connection-validation.test.ts` - Comprehensive test suite (13 tests)
+
+#### References
+
+- **Issue**: #357 - n8n_update_partial_workflow incorrectly validates AI nodes requiring 'main' connections
+- **Workflow**: `019Vrw56aROeEzVj` (WOO_Workflow_21_POST_Chat_Send_AI_Agent)
+- **Investigation**: Deep code analysis by Explore agent identified exact root cause in Zod schema
+- **Confirmation**: n8n-mcp-tester agent verified fix with real workflow
+
+Conceived by Romuald Członkowski - [www.aiadvisors.pl/en](https://www.aiadvisors.pl/en)
+
 ## [2.21.0] - 2025-10-23
 
 ### ✨ Features

package.json

@@ -1,6 +1,6 @@
 {
   "name": "n8n-mcp",
-  "version": "2.21.0",
+  "version": "2.21.1",
   "description": "Integration between n8n workflow automation and Model Context Protocol (MCP)",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",

src/mcp/tool-docs/workflow_management/n8n-update-partial-workflow.ts

@@ -81,6 +81,10 @@ Full support for all 8 AI connection types used in n8n AI workflows:
 - Multiple tools: Batch multiple \`sourceOutput: "ai_tool"\` connections to one AI Agent
 - Vector retrieval: Chain ai_embedding → ai_vectorStore → ai_tool → AI Agent
 
+**Important Notes**:
+- **AI nodes do NOT require main connections**: Nodes like OpenAI Chat Model, Postgres Chat Memory, Embeddings OpenAI, and Supabase Vector Store use AI-specific connection types exclusively. They should ONLY have connections like \`ai_languageModel\`, \`ai_memory\`, \`ai_embedding\`, or \`ai_tool\` - NOT \`main\` connections.
+- **Fixed in v2.21.1**: Validation now correctly recognizes AI nodes that only have AI-specific connections without requiring \`main\` connections (resolves issue #357).
+
 **Best Practices**:
 - Always specify \`sourceOutput\` for AI connections (defaults to "main" if omitted)
 - Connect language model BEFORE creating/enabling AI Agent (validation requirement)

src/services/n8n-validation.ts

@@ -24,9 +24,8 @@ export const workflowNodeSchema = z.object({
   executeOnce: z.boolean().optional(),
 });
 
-export const workflowConnectionSchema = z.record(
+// Connection array schema used by all connection types
-  z.object({
+const connectionArraySchema = z.array(
-    main: z.array(
   z.array(
     z.object({
       node: z.string(),
@@ -34,7 +33,22 @@ export const workflowConnectionSchema = z.record(
       index: z.number(),
     })
   )
-    ),
+);
+
+/**
+ * Workflow connection schema supporting all connection types.
+ * Note: 'main' is optional because AI nodes exclusively use AI-specific
+ * connection types (ai_languageModel, ai_memory, etc.) without main connections.
+ */
+export const workflowConnectionSchema = z.record(
+  z.object({
+    main: connectionArraySchema.optional(),
+    error: connectionArraySchema.optional(),
+    ai_tool: connectionArraySchema.optional(),
+    ai_languageModel: connectionArraySchema.optional(),
+    ai_memory: connectionArraySchema.optional(),
+    ai_embedding: connectionArraySchema.optional(),
+    ai_vectorStore: connectionArraySchema.optional(),
   })
 );
 

tests/integration/workflow-diff/ai-node-connection-validation.test.ts

@@ -0,0 +1,722 @@
+/**
+ * Integration tests for AI node connection validation in workflow diff operations
+ * Tests that AI nodes with AI-specific connection types (ai_languageModel, ai_memory, etc.)
+ * are properly validated without requiring main connections
+ *
+ * Related to issue #357
+ */
+
+import { describe, test, expect } from 'vitest';
+import { WorkflowDiffEngine } from '../../../src/services/workflow-diff-engine';
+
+describe('AI Node Connection Validation', () => {
+  describe('AI-specific connection types', () => {
+    test('should accept workflow with ai_languageModel connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'AI Language Model Test',
+        nodes: [
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'llm-node',
+            name: 'OpenAI Chat Model',
+            type: '@n8n/n8n-nodes-langchain.lmChatOpenAi',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          'OpenAI Chat Model': {
+            ai_languageModel: [
+              [{ node: 'AI Agent', type: 'ai_languageModel', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+    });
+
+    test('should accept workflow with ai_memory connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'AI Memory Test',
+        nodes: [
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'memory-node',
+            name: 'Postgres Chat Memory',
+            type: '@n8n/n8n-nodes-langchain.memoryPostgresChat',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          'Postgres Chat Memory': {
+            ai_memory: [
+              [{ node: 'AI Agent', type: 'ai_memory', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+    });
+
+    test('should accept workflow with ai_embedding connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'AI Embedding Test',
+        nodes: [
+          {
+            id: 'vectorstore-node',
+            name: 'Vector Store',
+            type: '@n8n/n8n-nodes-langchain.vectorStoreSupabase',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'embedding-node',
+            name: 'Embeddings OpenAI',
+            type: '@n8n/n8n-nodes-langchain.embeddingsOpenAi',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          'Embeddings OpenAI': {
+            ai_embedding: [
+              [{ node: 'Vector Store', type: 'ai_embedding', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+    });
+
+    test('should accept workflow with ai_tool connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'AI Tool Test',
+        nodes: [
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'vectorstore-node',
+            name: 'Vector Store Tool',
+            type: '@n8n/n8n-nodes-langchain.vectorStoreSupabase',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          'Vector Store Tool': {
+            ai_tool: [
+              [{ node: 'AI Agent', type: 'ai_tool', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+    });
+
+    test('should accept workflow with ai_vectorStore connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'AI Vector Store Test',
+        nodes: [
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'vectorstore-node',
+            name: 'Supabase Vector Store',
+            type: '@n8n/n8n-nodes-langchain.vectorStoreSupabase',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          'Supabase Vector Store': {
+            ai_vectorStore: [
+              [{ node: 'AI Agent', type: 'ai_vectorStore', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+    });
+  });
+
+  describe('Mixed connection types', () => {
+    test('should accept workflow mixing main and AI connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'Mixed Connections Test',
+        nodes: [
+          {
+            id: 'webhook-node',
+            name: 'Webhook',
+            type: 'n8n-nodes-base.webhook',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          },
+          {
+            id: 'llm-node',
+            name: 'OpenAI Chat Model',
+            type: '@n8n/n8n-nodes-langchain.lmChatOpenAi',
+            typeVersion: 1,
+            position: [200, 200],
+            parameters: {}
+          },
+          {
+            id: 'respond-node',
+            name: 'Respond to Webhook',
+            type: 'n8n-nodes-base.respondToWebhook',
+            typeVersion: 1,
+            position: [400, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          'Webhook': {
+            main: [
+              [{ node: 'AI Agent', type: 'main', index: 0 }]
+            ]
+          },
+          'AI Agent': {
+            main: [
+              [{ node: 'Respond to Webhook', type: 'main', index: 0 }]
+            ]
+          },
+          'OpenAI Chat Model': {
+            ai_languageModel: [
+              [{ node: 'AI Agent', type: 'ai_languageModel', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+    });
+
+    test('should accept workflow with error connections alongside AI connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'Error + AI Connections Test',
+        nodes: [
+          {
+            id: 'webhook-node',
+            name: 'Webhook',
+            type: 'n8n-nodes-base.webhook',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          },
+          {
+            id: 'llm-node',
+            name: 'OpenAI Chat Model',
+            type: '@n8n/n8n-nodes-langchain.lmChatOpenAi',
+            typeVersion: 1,
+            position: [200, 200],
+            parameters: {}
+          },
+          {
+            id: 'error-handler',
+            name: 'Error Handler',
+            type: 'n8n-nodes-base.set',
+            typeVersion: 1,
+            position: [200, -200],
+            parameters: {}
+          }
+        ],
+        connections: {
+          'Webhook': {
+            main: [
+              [{ node: 'AI Agent', type: 'main', index: 0 }]
+            ]
+          },
+          'AI Agent': {
+            error: [
+              [{ node: 'Error Handler', type: 'main', index: 0 }]
+            ]
+          },
+          'OpenAI Chat Model': {
+            ai_languageModel: [
+              [{ node: 'AI Agent', type: 'ai_languageModel', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+    });
+  });
+
+  describe('Complex AI workflow (Issue #357 scenario)', () => {
+    test('should accept full AI agent workflow with RAG components', async () => {
+      // Simplified version of the workflow from issue #357
+      const workflow = {
+        id: 'test-workflow',
+        name: 'AI Agent with RAG',
+        nodes: [
+          {
+            id: 'webhook-node',
+            name: 'Webhook',
+            type: 'n8n-nodes-base.webhook',
+            typeVersion: 2,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'code-node',
+            name: 'Prepare Inputs',
+            type: 'n8n-nodes-base.code',
+            typeVersion: 2,
+            position: [200, 0],
+            parameters: {}
+          },
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1.7,
+            position: [400, 0],
+            parameters: {}
+          },
+          {
+            id: 'llm-node',
+            name: 'OpenAI Chat Model',
+            type: '@n8n/n8n-nodes-langchain.lmChatOpenAi',
+            typeVersion: 1,
+            position: [400, 200],
+            parameters: {}
+          },
+          {
+            id: 'memory-node',
+            name: 'Postgres Chat Memory',
+            type: '@n8n/n8n-nodes-langchain.memoryPostgresChat',
+            typeVersion: 1.1,
+            position: [500, 200],
+            parameters: {}
+          },
+          {
+            id: 'embedding-node',
+            name: 'Embeddings OpenAI',
+            type: '@n8n/n8n-nodes-langchain.embeddingsOpenAi',
+            typeVersion: 1,
+            position: [600, 400],
+            parameters: {}
+          },
+          {
+            id: 'vectorstore-node',
+            name: 'Supabase Vector Store',
+            type: '@n8n/n8n-nodes-langchain.vectorStoreSupabase',
+            typeVersion: 1.3,
+            position: [600, 200],
+            parameters: {}
+          },
+          {
+            id: 'respond-node',
+            name: 'Respond to Webhook',
+            type: 'n8n-nodes-base.respondToWebhook',
+            typeVersion: 1.1,
+            position: [600, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          'Webhook': {
+            main: [
+              [{ node: 'Prepare Inputs', type: 'main', index: 0 }]
+            ]
+          },
+          'Prepare Inputs': {
+            main: [
+              [{ node: 'AI Agent', type: 'main', index: 0 }]
+            ]
+          },
+          'AI Agent': {
+            main: [
+              [{ node: 'Respond to Webhook', type: 'main', index: 0 }]
+            ]
+          },
+          'OpenAI Chat Model': {
+            ai_languageModel: [
+              [{ node: 'AI Agent', type: 'ai_languageModel', index: 0 }]
+            ]
+          },
+          'Postgres Chat Memory': {
+            ai_memory: [
+              [{ node: 'AI Agent', type: 'ai_memory', index: 0 }]
+            ]
+          },
+          'Embeddings OpenAI': {
+            ai_embedding: [
+              [{ node: 'Supabase Vector Store', type: 'ai_embedding', index: 0 }]
+            ]
+          },
+          'Supabase Vector Store': {
+            ai_tool: [
+              [{ node: 'AI Agent', type: 'ai_tool', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+      expect(result.errors || []).toHaveLength(0);
+    });
+
+    test('should successfully update AI workflow nodes without connection errors', async () => {
+      // Test that we can update nodes in an AI workflow without triggering validation errors
+      const workflow = {
+        id: 'test-workflow',
+        name: 'AI Workflow Update Test',
+        nodes: [
+          {
+            id: 'webhook-node',
+            name: 'Webhook',
+            type: 'n8n-nodes-base.webhook',
+            typeVersion: 2,
+            position: [0, 0],
+            parameters: { path: 'test' }
+          },
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          },
+          {
+            id: 'llm-node',
+            name: 'OpenAI Chat Model',
+            type: '@n8n/n8n-nodes-langchain.lmChatOpenAi',
+            typeVersion: 1,
+            position: [200, 200],
+            parameters: {}
+          }
+        ],
+        connections: {
+          'Webhook': {
+            main: [
+              [{ node: 'AI Agent', type: 'main', index: 0 }]
+            ]
+          },
+          'OpenAI Chat Model': {
+            ai_languageModel: [
+              [{ node: 'AI Agent', type: 'ai_languageModel', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+
+      // Update the webhook node (unrelated to AI nodes)
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: [
+          {
+            type: 'updateNode',
+            nodeId: 'webhook-node',
+            updates: {
+              notes: 'Updated webhook configuration'
+            }
+          }
+        ]
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+      expect(result.errors || []).toHaveLength(0);
+
+      // Verify the update was applied
+      const updatedNode = result.workflow.nodes.find((n: any) => n.id === 'webhook-node');
+      expect(updatedNode?.notes).toBe('Updated webhook configuration');
+    });
+  });
+
+  describe('Node-only AI nodes (no main connections)', () => {
+    test('should accept AI nodes with ONLY ai_languageModel connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'AI Node Without Main',
+        nodes: [
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'llm-node',
+            name: 'OpenAI Chat Model',
+            type: '@n8n/n8n-nodes-langchain.lmChatOpenAi',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          // OpenAI Chat Model has NO main connections, ONLY ai_languageModel
+          'OpenAI Chat Model': {
+            ai_languageModel: [
+              [{ node: 'AI Agent', type: 'ai_languageModel', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+      expect(result.errors || []).toHaveLength(0);
+    });
+
+    test('should accept AI nodes with ONLY ai_memory connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'Memory Node Without Main',
+        nodes: [
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'memory-node',
+            name: 'Postgres Chat Memory',
+            type: '@n8n/n8n-nodes-langchain.memoryPostgresChat',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          // Memory node has NO main connections, ONLY ai_memory
+          'Postgres Chat Memory': {
+            ai_memory: [
+              [{ node: 'AI Agent', type: 'ai_memory', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+      expect(result.errors || []).toHaveLength(0);
+    });
+
+    test('should accept embedding nodes with ONLY ai_embedding connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'Embedding Node Without Main',
+        nodes: [
+          {
+            id: 'vectorstore-node',
+            name: 'Vector Store',
+            type: '@n8n/n8n-nodes-langchain.vectorStoreSupabase',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'embedding-node',
+            name: 'Embeddings OpenAI',
+            type: '@n8n/n8n-nodes-langchain.embeddingsOpenAi',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          // Embedding node has NO main connections, ONLY ai_embedding
+          'Embeddings OpenAI': {
+            ai_embedding: [
+              [{ node: 'Vector Store', type: 'ai_embedding', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+      expect(result.errors || []).toHaveLength(0);
+    });
+
+    test('should accept vector store nodes with ONLY ai_tool connections', async () => {
+      const workflow = {
+        id: 'test-workflow',
+        name: 'Vector Store Node Without Main',
+        nodes: [
+          {
+            id: 'agent-node',
+            name: 'AI Agent',
+            type: '@n8n/n8n-nodes-langchain.agent',
+            typeVersion: 1,
+            position: [0, 0],
+            parameters: {}
+          },
+          {
+            id: 'vectorstore-node',
+            name: 'Supabase Vector Store',
+            type: '@n8n/n8n-nodes-langchain.vectorStoreSupabase',
+            typeVersion: 1,
+            position: [200, 0],
+            parameters: {}
+          }
+        ],
+        connections: {
+          // Vector store has NO main connections, ONLY ai_tool
+          'Supabase Vector Store': {
+            ai_tool: [
+              [{ node: 'AI Agent', type: 'ai_tool', index: 0 }]
+            ]
+          }
+        }
+      };
+
+      const engine = new WorkflowDiffEngine();
+      const result = await engine.applyDiff(workflow as any, {
+        id: workflow.id,
+        operations: []
+      });
+
+      expect(result.success).toBe(true);
+      expect(result.workflow).toBeDefined();
+      expect(result.errors || []).toHaveLength(0);
+    });
+  });
+});