n8n-project-init

n8n Project Initializer

Scaffold a new n8n automation project for the Pattern AI Automations team. This is Step 2 of 3 in the build pipeline:

[Step 1] PRD Generator  →  [Step 2] Project Init  →  [Step 3] n8n Builder

---

Step 1: Verify Pattern Credentials

Run this command to check for the team credentials file:

cat ~/.n8n.env 2>/dev/null

If the file exists and contains values:

• Confirm with the user: "Using Pattern n8n instance at https://pattern.app.n8n.cloud — correct?"
• Load N8N_API_URL, N8N_API_KEY, and N8N_CREDENTIALS_TEMPLATE_URL from the file
• Proceed to Step 2

If the file is missing or empty:

• Stop immediately and tell the user:

"Your Pattern credentials file (~/.n8n.env) is missing or empty. Before initializing a project, you need to set this up.

Ask Jon for the three credential values (N8N_API_URL, N8N_API_KEY, N8N_CREDENTIALS_TEMPLATE_URL), then run:

cat > ~/.n8n.env << 'EOF'
N8N_API_URL=https://pattern.app.n8n.cloud
N8N_API_KEY=PASTE_YOUR_KEY_HERE
N8N_CREDENTIALS_TEMPLATE_URL=PASTE_TEMPLATE_URL_HERE
EOF

Once that's done, come back and run the project initializer again."

Do not proceed until credentials are confirmed.

---

Step 2: Check for Blueprint

Before collecting project details, check whether a blueprint was already generated.

Run:

ls *.md 2>/dev/null

If a .md file is found in the current directory:

• Ask: "I see a markdown file here — is this your PRD blueprint? If so, I'll pull it into the project automatically."
• If yes, read it and use it in Step 4c

If no .md file is found:

• Ask using AskUserQuestion:

Question: "Do you have a blueprint from the PRD generator?"
Header: "Blueprint"
Options:
  - "Yes — I'll paste it now"
  - "No — skip for now, I'll add it later"
  - "No — take me back to the PRD generator first"

• If they choose "take me back to PRD generator": stop and say "Run the PRD generator first (Generate an automation blueprint) and come back when your blueprint is ready."
• If they paste a blueprint: use it in Step 4c
• If they skip: proceed without it, leave the Blueprint Reference section blank in CLAUDE.md

---

Step 3: Collect Project Details

Use AskUserQuestion tool to gather:

First call (2 questions):

• "What is the project name?" — kebab-case (e.g., lead-enrichment-workflow, slack-approval-bot) Header: "Project Name" Options: free text — remind them to use kebab-case
• "One sentence: what does this automation do?" Header: "Description" Options: free text

Second call (2 questions):

• "Who is building this?" (team member name) Header: "Builder" Options: free text
• "Who requested this, and what team are they on?" (e.g., "Sarah Chen, RevOps") Header: "Stakeholder" Options: free text

---

Step 4: Create Project Files

4a: Create directory

mkdir -p <project-name>

4b: .gitignore

cat > <project-name>/.gitignore << 'EOF'
.env
.env.*
!.env.template
.DS_Store
Thumbs.db
.vscode/settings.json
.idea/
node_modules/
*.log
tmp/
temp/
EOF

4c: .env

Write the Pattern credentials into the project .env:

# ============================================
# Pattern AI Automations — n8n Configuration
# ============================================
# NEVER commit this file to git
# ============================================

N8N_API_URL=https://pattern.app.n8n.cloud
N8N_API_KEY=<value from ~/.n8n.env>
N8N_CREDENTIALS_TEMPLATE_URL=<value from ~/.n8n.env>

# ============================================
# Project-Specific Variables (add as needed)
# ============================================
# WEBHOOK_PATH=
# SLACK_CHANNEL=
# GOOGLE_SHEET_ID=

4d: CLAUDE.md

Create CLAUDE.md with all placeholders replaced:

Placeholder	Value
{{PROJECT_NAME}}	Project name from Step 3
{{PROJECT_DESCRIPTION}}	Description from Step 3
{{DATE}}	Today's date (YYYY-MM-DD)
{{TEAM_MEMBER_NAME}}	Builder name from Step 3
{{STAKEHOLDER_NAME}}	Stakeholder name from Step 3
{{STAKEHOLDER_TEAM}}	Stakeholder team from Step 3
{{N8N_CREDENTIALS_TEMPLATE_URL}}	From ~/.n8n.env

If a blueprint was provided in Step 2: paste it verbatim into the "Blueprint Reference" section. If not: leave the section with the placeholder note.

---

Step 5: Verify API Connection

export $(cat <project-name>/.env | grep -v '^#' | xargs) && \
curl -s "${N8N_API_URL}/api/v1/workflows?limit=1" \
  -H "X-N8N-API-KEY: ${N8N_API_KEY}" | jq '.data | length'

• Returns a number → connection confirmed
• Fails → help troubleshoot: confirm URL is https://pattern.app.n8n.cloud, verify API key matches ~/.n8n.env, check network

---

Step 6: Report and Hand Off to Builder

Project initialized: <project-name>/

Files created:
  ✓ CLAUDE.md       — project context (update as you build)
  ✓ .env            — Pattern credentials (never commit)
  ✓ .gitignore      — protects .env from git

n8n instance: Pattern (https://pattern.app.n8n.cloud)
API connection: verified
Blueprint: <included / not included>

Then output the next step callout:

─────────────────────────────────────────────
NEXT STEP: Build the workflow

Your project is scaffolded and ready. To start building, say:

  "Build the workflow for <project-name>"

The n8n builder will implement the workflow node by node using the
blueprint in CLAUDE.md, testing after each step before moving on.
─────────────────────────────────────────────

---

Important Notes

• Pattern instance only — this plugin is configured for pattern.app.n8n.cloud. Do not modify for other instances.
• Never put API keys in CLAUDE.md — credentials live in .env only
• Always run the PRD generator first — a blueprint makes the build phase significantly faster and more accurate
• Verify the API connection every time — catch auth issues before building starts
• CLAUDE.md is a living doc — encourage the builder to update it with decisions, issues, and test notes as they build