agentleFS
neurawork-gitneurawork-git/n8n-autopilot/skills/check-mcps/SKILL.md

check-mcps

Skill18 starsChanged 43 days agoMITView on GitHub

Verify n8n-autopilot setup is functional — n8nac CLI present, workspace bound, n8n API reachable, companion plugin `n8n-as-code@n8nac-marketplace` enabled. Use after setup, after upgrading the plugin, or when commands behave unexpectedly.

The file

Files
SKILL.md

Check n8n-autopilot Setup

n8n-autopilot is CLI-only (npx n8nac …). There is no MCP server entry for this plugin — the mcp__n8n-as-code__* namespace seen in older docs was an upstream assumption that never landed (the npm n8nac mcp entry-point is broken and Etienne's companion plugin ships skill knowledge, not an MCP server).

This skill verifies the actually-required layers.

Step 1 — n8nac CLI reachable

npx n8nac --version 2>&1

Expect 2.2.0 or higher. Anything below is unsupported.

Step 2 — Full health check via setup-check.sh

bash "$CLAUDE_PLUGIN_ROOT/scripts/setup-check.sh" 2>&1

($CLAUDE_PLUGIN_ROOT is set by Claude Code when running plugin commands — no manual path needed.)

Inspect the output. The script checks, in order:

  1. npx available
  2. n8nac CLI version vs minimum/reference
  3. n8nac workspace bound (workspace status --json returns env config)
  4. n8n API reachable (host pulled from the workspace status JSON)
  5. Companion plugin n8n-as-code@n8nac-marketplace enabled (warns if missing — the companion's n8n-architect skill is the primary source of n8n authoring knowledge)
  6. Community-node schema coverage
  7. Inventory freshness (informational)

Step 3 — Companion plugin (n8n-as-code) enabled

grep -q '"n8n-as-code@n8nac-marketplace": true' "$HOME/.claude/settings.json" && \
  echo "OK: n8n-as-code companion enabled" || \
  echo "WARN: n8n-as-code companion not enabled — install it for best UX:
  claude plugin marketplace add EtienneLescot/n8n-as-code
  claude plugin install n8n-as-code@n8nac-marketplace"

Step 4 — Project visibility

node "$CLAUDE_PLUGIN_ROOT/skills/find-project/scripts/list.js"

Prints the workspace-pinned project plus every project derivable from credential ownership. Surface the table to the user — multi-project blindness (referencing creds from the wrong project) is the most common cause of "workflow pushes but fails at runtime".

Step 4b — Native n8n MCP (test path + push gate)

node "$CLAUDE_PLUGIN_ROOT/scripts/check-native-mcp.mjs" --live

✅ native MCP ready = the plugin's n8n-native server can test workflows on the session env's instance and the push-lint gate can check node versions there. Any INFO: line carries the exact remedy (token from the n8n UI, piped into npx n8nac native-mcp configure). Silent output means N8NAC_ENVIRONMENT is not pinned — fix that first (/n8n-autopilot:session-env).

Step 5 — Report

Print a one-line summary per layer, plus the most likely fix for any FAIL:

Layer                              | Status
───────────────────────────────────┼─────────
n8nac CLI (>= 2.2.0)                |  OK
workspace bound                     |  OK
n8n API reachable                   |  OK
companion plugin (n8n-as-code)      |  OK
native MCP (n8n-native + gate)      |  OK
schemas/_index.json present         |  OK

For any non-OK row, surface the line from setup-check.sh verbatim — that script already prints the suggested fix.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.