跳到主要内容
Supermarket
返回能力市场
MCP Server
design
MIT

n8n-mcp

MCP server that lets Claude, ChatGPT, Cursor and other AI assistants operate n8n: debug, edit, validate, test and build workflows on any n8n instance

czlonkowskiczlonkowski
91/ 100

公开评测 · 综合采用结论

证据充分,整体质量与安全表现优秀

查看评测依据 查找我项目的报告基于公开项目证据,非安全认证或安装推荐
23.3kstars
3.7kforks
最近更新 1天前
评测生成时间(北京时间)
本报告引擎
v3.9.1
当前引擎
v3.17.3

本报告与当前引擎使用不同规则;原分数不会自动更新,不同版本的分数不宜直接对比。

重新评测此项目

进入后确认来源与额度,提交才会创建任务。

代码版本未记录或记录无效,无法核对本报告的 commit 与送评文本;不会用当前项目代码回填历史记录。

Evaluation report

综合采用结论

91
A
满分 100 · 静态采用参考
值得推荐低风险
决策摘要

证据充分,整体质量与安全表现优秀

100%
高置信度
86
文档
100
安全
91
质量
100
活跃
70
采用
  • 基础评测完成+25/25确定性评分与静态安全扫描已完成
  • README 有效证据+25/2523,003 个去重后的有效字符
  • 独立证据来源+20/206 类非重复证据,重复文件不叠加
  • 仓库元数据+10/10已取得仓库状态与采用数据
  • 活跃记录+5/5已取得最近提交时间
  • AI 复核+15/15已完成结构化 AI 证据复核

本报告评了什么

综合分用于静态采用筛选,不是任务成功率,也不代表使用 Skill 后的能力提升。

静态检查范围
记录 7 个文件 · 85,877 字符
只检查采集到的受限文本与已知风险模式;未读取的脚本、依赖和运行环境不在结论内。
AI 内容复核
已使用
复核送评文本的说明和设计,不是安装测试或实际任务执行。记录缺失时不推断已使用或未使用。
真实任务效果
本报告未提供实测证据
没有使用/不使用 Skill 的任务对照、重复运行或耗时与 token 成本结果。采用前仍需在隔离环境验证自己的任务。

脚本引用覆盖

脚本引用覆盖记录不可得;不推断没有引用或已经检查。

How it works · 架构图

n8n-MCP 架构与工具分类

README 描述了 MCP 服务器作为桥梁,连接 AI 助手与 n8n 平台,包含核心工具和管理工具,组件关系明确。

AI 提取 · 证据约束

左右滑动查看完整图示

n8n-MCP 架构与工具分类README 描述了 MCP 服务器作为桥梁,连接 AI 助手与 n8n 平台,包含核心工具和管理工具,组件关系明确。MCP 请求查询节点信息API 调用响应AI 助手客户端n8n-MCP 服务器MCP 服务器n8n 实例后端节点文档库数据源
图示依据
  • • README 描述 n8n-MCP 作为桥梁,提供节点文档访问
  • • 列出核心工具和管理工具,管理工具需要 N8N_API_URL 和 N8N_API_KEY
  • • 提供 n8n 集成指南,说明部署和认证
五维表现
n8n-MCP 提供全面的 n8n 节点文档与工作流管理能力,工具丰富且示例详尽,价值明确。最大缺口是部分高级功能依赖 n8n 实例级 MCP 配置,且 README 未提供完整排障指南。
质量证据
  • README 列出 7 个核心工具和 21 个管理工具,如 search_nodes、get_node、validate_workflow
  • 提供只读部署配置示例,使用 DISABLED_TOOLS 和 DISABLED_TOOL_OPERATIONS 环境变量
  • 提供 Cloudflare Access 认证配置,说明 N8N_CF_CLIENT_ID 和 N8N_CF_CLIENT_SECRET
  • 提供文档链接如 Self-Hosting Guide、Security & Hardening、n8n Deployment Guide
  • 提供模板搜索示例,如 search_templates({searchMode: 'by_metadata', complexity: 'simple'})
采用建议
优势
  • 问题与用途描述
  • 有效 README
  • 可执行示例
  • 输入、参数或工具说明
  • 未发现已知高风险模式
关注点
  • 缺少安装或接入步骤
  • 部分工具依赖 N8N_MCP_ACCESS_TOKEN,配置复杂
  • 未提供完整的错误处理与排障章节
  • 未明确说明各工具的具体输入输出参数
  • 未提供本地安装的具体命令示例
适合

需要深度集成 n8n 的 AI 助手开发、希望自动化构建和验证 n8n 工作流的开发者、需要安全只读部署的企业环境、需要模板搜索和节点文档查询的场景

不建议直接用于

没有 n8n 实例或 API 配置的用户、需要离线或本地优先的简单场景

也有自己的公开项目?先免登录查已有报告;没有报告时,可保留地址继续登录评测。

查找我项目的报告 →
文档证据
86/100
问题与用途描述10 分
有效 README12 分
安装或接入步骤14 分
可执行示例16 分
输入、参数或工具说明11 分
输出或结果说明9 分
限制、权限或边界12 分
错误处理或排障8 分
许可证信息5 分
结构化章节3 分
安全证据
低风险
未发现已知高风险模式

静态扫描不是安全保证,生产接入前仍应人工复核权限和数据边界。

优先改进清单
  1. 01补充安装或接入步骤
方法、证据与局限展开
数据来源

GitHub Repository API

扫描范围

7 个文件 · 85,877 字符

评测引擎

v3.9.1 · AI 内容复核:已使用(deepseek-chat)

局限
  • 静态评测不会安装或执行项目代码
  • 安全扫描基于高信号文件与已知模式,不能替代人工审计
  • 流行度只反映采用程度,不代表安全或工程质量

30 天热度趋势

README

n8n-MCP

License: MIT GitHub stars npm version codecov Tests n8n version Docker OpenSSF Best Practices Deploy on Railway

n8n-MCP is an open-source MCP server that lets Claude, ChatGPT, Cursor, and other AI assistants operate n8n: read failed executions, edit workflows in place, validate and test them, manage credentials, and build new workflows. It works with any n8n instance that has the Public API enabled, and can use n8n's built-in MCP server as a second backend.

What You Can Do With It

  • Debug failing workflows - n8n_executions reads an execution and, with mode: "error", returns the error, the failing node, and sample items from the node before it.
  • Edit workflows in place - n8n_update_partial_workflow applies targeted operations to the nodes and connections named in the request, so the assistant does not resend the whole workflow. This includes find-and-replace inside one large field such as a Code node (patchNodeField). By default n8n-MCP saves a snapshot before it changes a workflow's nodes or connections, and n8n_workflow_versions can diff and roll back.
  • Validate and fix before going live - validate_node, validate_workflow, and n8n_validate_workflow report unknown nodes, invalid node versions, missing required fields, and broken connections or expressions. n8n_autofix_workflow repairs the common ones.
  • Test - n8n_test_workflow runs a workflow through its webhook, form, or chat trigger, or, through n8n's built-in MCP server, with pinned data when it has no such trigger.
  • Manage the instance - create, update, and delete credentials, work with data tables and folders, and run a security audit (n8n_audit_instance).
  • Build new workflows - bundled knowledge of 3,501 nodes and 2,352 templates, so the assistant configures real nodes instead of guessing.
  • Work across several n8n instances - on the hosted service, register production, staging, and client instances once and name the target on each call. Self-hosted HTTP mode can serve several instances with per-request instance headers.

n8n-MCP and n8n's Built-In MCP Server

n8n ships its own instance-level MCP server, and it covers a lot: building and editing workflows, executions, version history, data tables, and Agents. n8n-MCP does not replace it. The two differ in how they are built:

n8n's built-in MCP servern8n-MCP
RunsInside one n8n instanceAs a separate server that talks to n8n
Instances per connectionOneSeveral (hosted service, or multi-tenant HTTP mode)
CredentialsLists credentials and edits their descriptionsCreates, updates, and deletes credentials
n8n versionsPresent since late n8n 1.x; most tools need n8n 2.12 or later, and each tool has its own minimum versionWorkflow and execution tools work through the Public API on older versions too, including the 1.x line. Credential actions depend on what that version's API exposes, and newer tools state the version they need
Workflow accessMost operations need the workflow's "Available in MCP" setting, which is off by defaultEvery workflow the API key can reach

The built-in server is enough when you work with a single instance on a recent n8n 2.x, build and edit workflows on that instance, and set up credentials yourself in the n8n UI.

Add n8n-MCP when you run more than one instance, want the assistant to create and update credentials, are on an older n8n version that predates the built-in tools you need, or want autofix, execution error analysis, templates, and a security audit.

Or use both through one connection. Set N8N_MCP_ACCESS_TOKEN and n8n-MCP routes the operations the Public API cannot perform to n8n's own server: test runs with pinned data, n8n's native version history, live dropdown values, n8n Agents, and data-table column changes. The assistant still sees one MCP server. See Official MCP Setup.

Naming note: n8n's documentation uses n8n-mcp as the example connection name for its built-in server in some client configurations (for example claude mcp add --transport http n8n-mcp https://<your-n8n-domain>/mcp-server/http). That name points at n8n's own server. This project is a separate open-source server: github.com/czlonkowski/n8n-mcp, published on npm as n8n-mcp.

Node and Template Coverage

The documentation and validation tools work without an n8n connection, against a bundled database:

  • 3,501 n8n nodes - 842 core nodes + 2,659 community nodes (2,228 verified)
  • Node properties - 99% coverage with detailed schemas
  • Node operations - 66.5% coverage of available actions
  • Documentation - 86% coverage from official n8n docs (including AI nodes)
  • AI tools - 267 AI-capable tool variants detected with full documentation
  • Real-world examples - 156 ranked configurations extracted from popular templates
  • Template library - 2,352 workflow templates with 99.96% AI metadata coverage
  • Community nodes - Search verified community integrations with source filter

Support This Project

Sponsor n8n-mcp

n8n-mcp started as a personal tool but now helps tens of thousands of developers automate their workflows efficiently. Maintaining and developing this project competes with my paid work. Your sponsorship helps me dedicate focused time to new features, respond quickly to issues, keep documentation up-to-date, and ensure compatibility with latest n8n releases. Become a sponsor

💼 Need it built for you? Work with AiAdvisors — n8n automation audits, builds, and operations, run by the author of n8n-mcp and n8n-skills.

Working Safely on Live Workflows

AI results can be unpredictable, and n8n-MCP can change real workflows. Try changes to important production workflows on a copy or a staging instance first.

  • Snapshots and rollback - by default n8n-MCP saves a version of a workflow before it changes its nodes or connections, and keeps up to 10 per workflow. n8n_workflow_versions lists, diffs, and rolls back those versions. A snapshot that fails to save does not stop the edit, so check that a version exists before relying on rollback for an important change.
  • One editor at a time - n8n-MCP reads a workflow, applies the change, and writes the whole workflow back, because n8n's Public API has no conditional write. An edit made in the n8n UI in between is overwritten, so do not edit a workflow in the UI while the assistant is working on it.
  • Validate before deploying - run validate_workflow on the change, and n8n_validate_workflow once it is in n8n.
  • Keep exports of workflows you cannot afford to lose.
  • Limit what the assistant can do - Read-Only Deployment disables write operations.

Quick Start

The fastest way to try n8n-MCP - no installation, no configuration:

dashboard.n8n-mcp.com

  • Free tier: 100 tool calls/day
  • Instant access: Connect your n8n instance and start working on your workflows
  • Several n8n instances through one connection, on every plan (multi-instance guide)
  • OAuth sign-in for claude.ai and ChatGPT, API keys for IDEs and other MCP clients
  • Your n8n API key is stored encrypted and is never returned to the AI client
  • Always up-to-date: Latest n8n nodes and templates
  • No infrastructure: We handle everything

The hosted service is run by the author of this project. Just sign up, connect your n8n instance, and connect your MCP client.

Want to self-host? See the Self-Hosting Guide for npx, Docker, Railway, and local installation options.

n8n Integration

Want to use n8n-MCP with your n8n instance? Check out our comprehensive n8n Deployment Guide for:

  • Local testing with the MCP Client Tool node
  • Production deployment with Docker Compose
  • Cloud deployment on Hetzner, AWS, and other providers
  • Troubleshooting and security best practices

Cloudflare Access Authentication

If your n8n instance sits behind Cloudflare Access (Zero Trust), provide your service token so n8n-MCP can authenticate:

  • N8N_CF_CLIENT_ID - Cloudflare Access Client ID
  • N8N_CF_CLIENT_SECRET - Cloudflare Access Client Secret

When set, these are sent as CF-Access-Client-Id / CF-Access-Client-Secret headers on n8n API requests, version/health probes, and webhook executions. The token is confined to the N8N_API_URL origin — webhook calls to a different host (e.g. a split WEBHOOK_URL origin) do not receive it, to avoid leaking the token.

n8n Agents and Instance-Level MCP (Optional)

To use n8n_manage_agents, n8n_explore_node_resources, and the project fallback in n8n_list_catalog, set:

  • N8N_MCP_ACCESS_TOKEN - MCP API key from n8n Settings → Instance-level MCP → set MCP status to Enabled. This is a separate secret from N8N_API_KEY and should be stored the same way. The MCP endpoint is derived from N8N_API_URL; instances that serve MCP from a split host (N8N_MCP_BASE_URL) are not supported.

See Connecting n8n-mcp to n8n's instance-level MCP server for the full setup walkthrough, including how to get the token from the n8n UI, prerequisites, and troubleshooting.

Connect your IDE

n8n-MCP works with multiple AI-powered IDEs and tools:

Add Claude Skills (Optional)

Supercharge your n8n workflow building with specialized skills that teach AI how to build production-ready workflows!

n8n-mcp Skills Setup

Learn more: n8n-skills repository

Claude Project Setup

For the best results when using n8n-MCP with Claude Projects, use these enhanced system instructions:

You are an expert in n8n automation software using n8n-MCP tools. Your role is to design, build, and validate n8n workflows with maximum accuracy and efficiency.

## Core Principles

### 1. Silent Execution
CRITICAL: Execute tools without commentary. Only respond AFTER all tools complete.

### 2. Parallel Execution
When operations are independent, execute them in parallel for maximum performance.

### 3. Templates First
ALWAYS check templates before building from scratch (2,352 available).

### 4. Multi-Level Validation
Use validate_node(mode='minimal') → validate_node(mode='full') → validate_workflow pattern.

### 5. Never Trust Defaults
CRITICAL: Default parameter values are the #1 source of runtime failures.
ALWAYS explicitly configure ALL parameters that control node behavior.

## Workflow Process

1. **Start**: Call `tools_documentation()` for best practices

2. **Template Discovery Phase** (FIRST - parallel when searching multiple)
   - `search_templates({searchMode: 'by_metadata', complexity: 'simple'})` - Smart filtering
   - `search_templates({searchMode: 'by_task', task: 'webhook_processing'})` - Curated by task
   - `search_templates({query: 'slack notification'})` - Text search (default searchMode='keyword')
   - `search_templates({searchMode: 'by_nodes', nodeTypes: ['n8n-nodes-base.slack']})` - By node type

   **Filtering strategies**:
   - Beginners: `complexity: "simple"` + `maxSetupMinutes: 30`
   - By role: `targetAudience: "marketers"` | `"developers"` | `"analysts"`
   - By time: `maxSetupMinutes: 15` for quick wins
   - By service: `requiredService: "openai"` for compatibility

3. **Node Discovery** (if no suitable template - parallel execution)
   - Think deeply about requirements. Ask clarifying questions if unclear.
   - `search_nodes({query: 'keyword', includeExamples: true})` - Parallel for multiple nodes
   - `search_nodes({query: 'trigger'})` - Browse triggers
   - `search_nodes({query: 'AI agent langchain'})` - AI-capable nodes

4. **Configuration Phase** (parallel for multiple nodes)
   - `get_node({nodeType, detail: 'standard', includeExamples: true})` - Essential properties (default)
   - `get_node({nodeType, detail: 'minimal'})` - Basic metadata only (~200 tokens)
   - `get_node({nodeType, detail: 'full'})` - Complete information (~3000-8000 tokens)
   - `get_node({nodeType, mode: 'search_properties', propertyQuery: 'auth'})` - Find specific properties
   - `get_node({nodeType, mode: 'docs'})` - Human-readable markdown documentation
   - Show workflow architecture to user for approval before proceeding

5. **Validation Phase** (parallel for multiple nodes)
   - `validate_node({nodeType, config, mode: 'minimal'})` - Quick required fields check
   - `validate_node({nodeType, config, mode: 'full', profile: 'runtime'})` - Full validation with fixes
   - Fix ALL errors before proceeding

6. **Building Phase**
   - If using template: `get_template(templateId, {mode: "full"})`
   - **MANDATORY ATTRIBUTION**: "Based on template by **[author.name]** (@[username]). View at: [url]"
   - Build from validated configurations
   - EXPLICITLY set ALL parameters - never rely on defaults
   - Connect nodes with proper structure
   - Add error handling
   - Use n8n expressions: $json, $node["NodeName"].json
   - Build in artifact (unless deploying to n8n instance)

7. **Workflow Validation** (before deployment)
   - `validate_workflow(workflow)` - Complete validation
   - `validate_workflow_connections(workflow)` - Structure check
   - `validate_workflow_expressions(workflow)` - Expression validation
   - Fix ALL issues before deployment

8. **Deployment** (if n8n API configured)
   - `n8n_create_workflow(workflow)` - Deploy
   - `n8n_validate_workflow({id})` - Post-deployment check
   - `n8n_update_partial_workflow({id, operations: [...]})` - Batch updates
   - `n8n_test_workflow({workflowId})` - Test workflow execution

## Critical Warnings

### Never Trust Defaults
Default values cause runtime failures. Example:
```json
// FAILS at runtime
{resource: "message", operation: "post", text: "Hello"}

// WORKS - all parameters explicit
{resource: "message", operation: "post", select: "channel", channelId: "C123", text: "Hello"}
```

### Example Availability
`includeExamples: true` returns real configurations from workflow templates.
- Coverage varies by node popularity
- When no examples available, use `get_node` + `validate_node({mode: 'minimal'})`

## Validation Strategy

### Level 1 - Quick Check (before building)
`validate_node({nodeType, config, mode: 'minimal'})` - Required fields only (<100ms)

### Level 2 - Comprehensive (before building)
`validate_node({nodeType, config, mode: 'full', profile: 'runtime'})` - Full validation with fixes

### Level 3 - Complete (after building)
`validate_workflow(workflow)` - Connections, expressions, AI tools

### Level 4 - Post-Deployment
1. `n8n_validate_workflow({id})` - Validate deployed workflow
2. `n8n_autofix_workflow({id})` - Auto-fix common errors
3. `n8n_executions({action: 'list'})` - Monitor execution status

## Response Format

### Initial Creation
```
[Silent tool execution in parallel]

Created workflow:
- Webhook trigger → Slack notification
- Configured: POST /webhook → #general channel

Validation: All checks passed
```

### Modifications
```
[Silent tool execution]

Updated workflow:
- Added error handling to HTTP node
- Fixed required Slack parameters

Changes validated successfully.
```

## Batch Operations

Use `n8n_update_partial_workflow` with multiple operations in a single call:

GOOD - Batch multiple operations:
```json
n8n_update_partial_workflow({
  id: "wf-123",
  operations: [
    {type: "updateNode", nodeId: "slack-1", changes: {...}},
    {type: "updateNode", nodeId: "http-1", changes: {...}},
    {type: "cleanStaleConnections"}
  ]
})
```

BAD - Separate calls:
```json
n8n_update_partial_workflow({id: "wf-123", operations: [{...}]})
n8n_update_partial_workflow({id: "wf-123", operations: [{...}]})
```

### CRITICAL: addConnection Syntax

The `addConnection` operation requires **four separate string parameters**. Common mistakes cause misleading errors.

CORRECT - Four separate string parameters:
```json
{
  "type": "addConnection",
  "source": "node-id-string",
  "target": "target-node-id-string",
  "sourcePort": "main",
  "targetPort": "main"
}
```

**Reference**: [GitHub Issue #327](https://github.com/czlonkowski/n8n-mcp/issues/327)

### CRITICAL: IF Node Multi-Output Routing

IF nodes have **two outputs** (TRUE and FALSE). Use the **`branch` parameter** to route to the correct output:

```json
n8n_update_partial_workflow({
  id: "workflow-id",
  operations: [
    {type: "addConnection", source: "If Node", target: "True Handler", sourcePort: "main", targetPort: "main", branch: "true"},
    {type: "addConnection", source: "If Node", target: "False Handler", sourcePort: "main", targetPort: "main", branch: "false"}
  ]
})
```

**Note**: Without the `branch` parameter, both connections may end up on the same output, causing logic errors!

### removeConnection Syntax

Use the same four-parameter format:
```json
{
  "type": "removeConnection",
  "source": "source-node-id",
  "target": "target-node-id",
  "sourcePort": "main",
  "targetPort": "main"
}
```

## Important Rules

### Core Behavior
1. **Silent execution** - No commentary between tools
2. **Parallel by default** - Execute independent operations simultaneously
3. **Templates first** - Always check before building (2,352 available)
4. **Multi-level validation** - Quick check → Full validation → Workflow validation
5. **Never trust defaults** - Explicitly configure ALL parameters

### Attribution & Credits
- **MANDATORY TEMPLATE ATTRIBUTION**: Share author name, username, and n8n.io link
- **Template validation** - Always validate before deployment (may need updates)

### Code Node Usage
- **Avoid when possible** - Prefer standard nodes
- **Only when necessary** - Use code node as last resort
- **AI tool capability** - ANY node can be an AI tool (not just marked ones)

### Most Popular n8n Nodes (for get_node):

1. **n8n-nodes-base.code** - JavaScript/Python scripting
2. **n8n-nodes-base.httpRequest** - HTTP API calls
3. **n8n-nodes-base.webhook** - Event-driven triggers
4. **n8n-nodes-base.set** - Data transformation
5. **n8n-nodes-base.if** - Conditional routing
6. **n8n-nodes-base.manualTrigger** - Manual workflow execution
7. **n8n-nodes-base.respondToWebhook** - Webhook responses
8. **n8n-nodes-base.scheduleTrigger** - Time-based triggers
9. **@n8n/n8n-nodes-langchain.agent** - AI agents
10. **n8n-nodes-base.googleSheets** - Spreadsheet integration
11. **n8n-nodes-base.merge** - Data merging
12. **n8n-nodes-base.switch** - Multi-branch routing
13. **n8n-nodes-base.telegram** - Telegram bot integration
14. **@n8n/n8n-nodes-langchain.lmChatOpenAi** - OpenAI chat models
15. **n8n-nodes-base.splitInBatches** - Batch processing
16. **n8n-nodes-base.openAi** - OpenAI legacy node
17. **n8n-nodes-base.gmail** - Email automation
18. **n8n-nodes-base.function** - Custom functions
19. **n8n-nodes-base.stickyNote** - Workflow documentation
20. **n8n-nodes-base.executeWorkflowTrigger** - Sub-workflow calls

**Note:** LangChain nodes use the `@n8n/n8n-nodes-langchain.` prefix, core nodes use `n8n-nodes-base.`

Save these instructions in your Claude Project for optimal n8n workflow assistance with intelligent template discovery.

Available MCP Tools

Core Tools (7 tools)

  • tools_documentation - Get documentation for any MCP tool (START HERE!)
  • search_nodes - Full-text search across all nodes. Use source: 'community'|'verified' for community nodes, includeExamples: true for configs
  • get_node - Unified node information tool with multiple modes:
    • Info mode (default): detail: 'minimal'|'standard'|'full', includeExamples: true
    • Docs mode: mode: 'docs' - Human-readable markdown documentation
    • Property search: mode: 'search_properties', propertyQuery: 'auth'
    • Versions: mode: 'versions'|'compare'|'breaking'|'migrations'
  • validate_node - Unified node validation:
    • mode: 'minimal' - Quick required fields check (<100ms)
    • mode: 'full' - Comprehensive validation with profiles (minimal, runtime, ai-friendly, strict)
  • validate_workflow - Complete workflow validation including AI Agent validation
  • search_templates - Unified template search:
    • searchMode: 'keyword' (default) - Text search with query parameter
    • searchMode: 'by_nodes' - Find templates using specific nodeTypes
    • searchMode: 'by_task' - Curated templates for common task types
    • searchMode: 'by_metadata' - Filter by complexity, requiredService, targetAudience
  • get_template - Get complete workflow JSON (modes: nodes_only, structure, full)

n8n Management Tools (21 tools - Requires API Configuration)

These tools require N8N_API_URL and N8N_API_KEY in your configuration.

Workflow Management

  • n8n_create_workflow - Create new workflows with nodes and connections
  • n8n_get_workflow - Unified workflow retrieval (modes: full, details, structure, minimal)
  • n8n_update_full_workflow - Update entire workflow (complete replacement)
  • n8n_update_partial_workflow - Update workflow using diff operations
  • n8n_delete_workflow - Delete workflows permanently
  • n8n_list_workflows - List workflows with filtering and pagination
  • n8n_validate_workflow - Validate workflows in n8n by ID
  • n8n_autofix_workflow - Automatically fix common workflow errors
  • n8n_workflow_versions - Version history, diff and rollback over two histories: source: 'local' (the snapshots n8n-mcp takes before it changes a workflow, the default) and source: 'native' (n8n's own workflow history, including UI edits — needs N8N_MCP_ACCESS_TOKEN and the workflow's "Available in MCP" setting)
  • n8n_deploy_template - Deploy templates from n8n.io directly to your instance with auto-fix

Node Resource Discovery

  • n8n_explore_node_resources - Resolve a node's dynamic dropdown (loadOptions) or resource-locator search (listSearch) values — Slack channels, Google Sheets tabs, model lists — using a real credential, so workflow configs use existing IDs instead of invented ones. Requires N8N_MCP_ACCESS_TOKEN (see Official MCP Setup)

Execution Management

  • n8n_test_workflow - Run a workflow. method: 'auto' (default) triggers it over HTTP through its webhook/form/chat trigger; method: 'prepare'/'pinned'/'direct' run workflows that have no such trigger through n8n's own MCP server (needs N8N_MCP_ACCESS_TOKEN and the workflow's "Available in MCP" setting)
  • n8n_executions - Unified execution management (list, get, delete)
  • n8n_evaluations - Run and read evaluation test runs (list runs, aggregated metrics, per-case results on n8n 2.30+; trigger and cancel on 2.32+)

Folder Management

  • n8n_manage_folders - Manage workflow folders (create, list, get, rename, move, delete; n8n 2.19+). Place workflows into folders via n8n_create_workflow's parentFolderId or n8n_update_partial_workflow's moveToFolder operation (n8n 2.32+)

Data Table Management

  • n8n_manage_datatable - Manage n8n data tables, rows and columns (list, get, create, update, delete; addColumn/deleteColumn/renameColumn change an existing table's columns through n8n's own MCP server and need N8N_MCP_ACCESS_TOKEN)

Credential Management

  • n8n_manage_credentials - Manage n8n credentials (list, get, create, update, delete, getSchema)

Security & Audit

  • n8n_audit_instance - Security audit combining n8n's built-in audit API with deep workflow scanning

Agents

  • n8n_manage_agents - Manage n8n Agents (persisted assistants with a model, instructions, tools, skills, tasks, memory and channels) through n8n's instance-level MCP server. Requires N8N_MCP_ACCESS_TOKEN and n8n 2.34+ with the agents module (see Official MCP Setup). This is not the AI Agent workflow node — use get_node for that

System Tools

  • n8n_health_check - Check n8n API connectivity and features, including officialMcp status when N8N_MCP_ACCESS_TOKEN is configured
  • n8n_list_catalog - List instance-level projects or tags; falls back to n8n's instance-level MCP server for team projects when the Public API doesn't expose them

Read-Only Deployment

For governance-sensitive environments, use both env vars together. Fully disable tools that are write/destructive or handle sensitive data (n8n_manage_credentials and n8n_manage_datatable also offer read operations, but are removed entirely here because even reads expose sensitive material):

DISABLED_TOOLS=n8n_create_workflow,n8n_update_full_workflow,n8n_update_partial_workflow,n8n_delete_workflow,n8n_autofix_workflow,n8n_deploy_template,n8n_test_workflow,n8n_manage_credentials,n8n_manage_datatable

For tools that bundle read and write operations under one name, block only the destructive operations while keeping list and get. Use this instead of a full DISABLED_TOOLS entry where the tool's read operations are acceptable — the n8n_manage_datatable and n8n_test_workflow entries below are the alternative to removing those tools entirely as above. The example names every write operation of every tool that has one:

DISABLED_TOOL_OPERATIONS=n8n_executions:delete;n8n_test_workflow:auto,trigger,pinned,direct,expose;n8n_evaluations:run,cancel;n8n_manage_folders:create,rename,move,delete;n8n_workflow_versions:delete,rollback,prune,expose;n8n_manage_agents:create,mutate,call,publish,unpublish,revert,delete,update_integration;n8n_manage_datatable:createTable,updateTable,deleteTable,insertRows,updateRows,upsertRows,deleteRows,addColumn,deleteColumn,renameColumn

Two details are easy to miss when writing your own list. For n8n_test_workflow, all four of auto, trigger, pinned and direct run the workflow (an omitted or blank method counts as auto), leaving only the read-only prepare. And expose is not a value of any operation parameter: it is the exposeToMcp consent write of n8n_test_workflow and n8n_workflow_versions, which enables a workflow's "Available in MCP" setting. Omitting expose leaves that write reachable.

Combine with a read-only n8n API key (Settings → API in your n8n instance) for defence in depth. See Read-Only Deployment Recipe for the full setup guide.

Documentation

License

MIT License - see LICENSE for details.

Contributing

See CONTRIBUTING.md for development setup, testing, and contribution guidelines.

Acknowledgments

See Acknowledgments for credits and template attribution.


Built with care for the n8n community

💼 Need it built for you?

Work with AiAdvisors — n8n automation audits, builds, and operations, run by the author of n8n-mcp and n8n-skills.