# n8n Template

> Search and deploy pre-built n8n workflow templates

- Skill: `cs-workx/n8n-template` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add cs-workx/n8n-template`
- Raw SKILL.md: https://api.skillmd.com/api/skills/cs-workx/n8n-template/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: DevOps & Infra
- Author: CS-Workx (https://skillmd.com/u/cs-workx)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/cs-workx/n8n-template

---


# Deploy an n8n Template

Search 2,700+ pre-built workflow templates and deploy them to your n8n instance at `your-n8n-instance.example.com`.

## Process

### 1. Search Templates
Use the user's description to find matching templates:
```
n8n-mcp → search_templates("send slack notification on new github issue")
```

### 2. Present Options
Show the user a summary of matching templates:
- Template name and description
- Nodes used
- Complexity (number of nodes)

### 3. Get Full Template
Once the user picks one:
```
n8n-mcp → get_template(templateId)
```

### 4. Customize
Help the user adapt the template:
- **Credentials**: Identify which credentials are needed and note them
- **Parameters**: Adjust URLs, channels, filters, etc. to match user's needs
- **Nodes**: Add, remove, or modify nodes if needed

### 5. Verify Credentials and Data Paths
**Verify credentials exist.** Run `GET /credentials` and confirm every credential type referenced in the template exists on the instance. Flag any missing credentials to the user — they must configure them in the n8n UI before the workflow can run.

**Check empty-data paths.** For each node that fetches external data (HTTP Request, Google Sheets, API calls, etc.), consider what happens if it returns 0 items. If downstream nodes would fail on empty input, add an IF node to check for data first.

### 6. Validate
Templates can be outdated or break after customization. Always validate:
```
n8n-mcp → validate_workflow({ customized template JSON })
```
Also manually check expressions against the expression pitfalls in CLAUDE.md — especially if you modified any data paths or field references during customization.

### 7. Deploy
```
n8n API → POST /workflows  (body: customized template JSON)
```

### 8. Smoke Test
**NEVER skip this step.** After deploying, trigger a test execution:

- **Manual trigger** → Run from the n8n UI or use `n8n_test_workflow`
- **Webhook trigger** → Fire a test HTTP request with realistic sample data
- **Schedule trigger** → Trigger a manual run via the n8n UI

After the test run:
1. Check execution status: `GET /executions?workflowId={id}&limit=1`
2. Get execution details: `GET /executions/{executionId}`
3. Verify output data is correct — not just that status is "success"
4. If the test fails, debug and fix before reporting success

### 9. Report Back
Only after a successful smoke test, show the user:
- Workflow name and ID
- URL: `https://your-n8n-instance.example.com/workflow/{id}`
- Template used (name and ID)
- Smoke test result: what was tested and what the output looked like
- List of credentials they need to configure in the n8n UI
- Any parameters they should review

## API Gotchas for Templates

Templates designed in the n8n UI may be missing properties required by the API:

- **Webhook nodes:** Ensure every webhook node has a `webhookId` property at node level (not inside `parameters`). The UI auto-generates this but templates deployed via API may lack it — causing 404 "not registered" at runtime. The validator does NOT catch this.
- **AI sub-node connections:** Every AI sub-node connection entry (model, tool, memory, parser) MUST include `"main": []` alongside the AI connection type. Without it, the API rejects with "Invalid connections." The validator does NOT catch this.

## Tips
- Templates often include placeholder credentials — always flag these
- Some templates use community nodes that may not be installed
- If no template matches exactly, suggest combining elements from multiple templates or use `/n8n-build` instead
- Even trusted templates need validation after customization — don't skip steps 5-6