Let an AI agent build and change your n8n workflows with a guaranteed way back: every write is backed up, version-checked, and verified by readback.
Let an AI agent build and change your n8n workflows — with a guaranteed way back.
Agents are good at writing workflow logic and bad at knowing when they have broken something. n8n-mcp-guard sits in between. Every change is backed up before it happens, checked against the exact version it was designed for, and verified against the live instance after saving. If anything does not line up, the write is refused rather than attempted and hoped for.
Set it up once, use it from any MCP-capable agent, in every repository.
You're building something new. Describe the automation you want, let the agent write it, and create it as an inactive workflow you can inspect before anything runs. Iterate with a reviewable diff at every step. You never have to read raw JSON to know what changed, and you can always get back to the last version that worked.
You have something already running. The workflow that quietly emails your customers or syncs your orders is the one you least want an agent improvising on. Here, a change to it is a two-phase commit: plan it, review the diff, rehearse it on a disposable clone with every trigger and outbound call disabled, then commit — and only if nothing moved underneath you in the meantime.
- 68 unit tests covering every path where a write is supposed to be refused, plus a live integration suite that drives the write tools through the real MCP surface against a real n8n instance: create, plan, apply, stale-plan refusal, clone neutralization, the delete guard, and verified cleanup. Run it yourself with npm run test:integration. - Validated against a real
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/FarazHayder--n8n-mcp-guard.html)
AI image and video generation skill for Claude Code and other coding agents — Gemini, OpenAI and Kie AI behind one CLI and MCP server, with EU AI Act content marking.
howznguyen/pi-delegate-mcpMCP server that turns the pi coding agent into a steerable background worker: delegate a task, redirect it mid-run, and keep its context out of your own.
zalaso/shareboxPublish and share the small web tools your AI agents build, as easily as a Google Doc. Google sign-in, permissions, storage and gVisor sandboxing built in. Self-hosted.
rashedInt32/jev-mcpMCP server exposing TypeSafe Jev as typed, calibrated judgment tools: classify, score, check, batched ask. Ships as a Claude Code plugin.
luckeyfaraday/frontier-orchestratorMulti-agent orchestration for Claude Code: MCP server + skill that keeps Claude as lead engineer while delegating backend work to OpenAI Codex and frontend/design work to Kimi
noteflowai/dsh-skills-anywhereReuse Agent Skills across DeepSeek Harness and MCP. Check local SKILL.md files in the browser or CI, with parser diagnostics, file hashes and clear exit codes. Try the Hugging Face playground.
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.