Claude Guidelines​

​
​

n8n Workflow Prompt Framework &

Guidelines
🎯 Core Prompt Structure
1. Context Setting (Always Include)

Unset
"Create an n8n workflow that [primary objective]. This workflow
should be production-ready with proper error handling and
documentation."

2. Trigger Definition (Required)

Specify ONE of these trigger types clearly:

●​ Manual Trigger: "Manually triggered workflow for..."

●​ Webhook: "When receiving a webhook from [source]..."
●​ Schedule: "Run every [frequency] to..."
●​ App Trigger: "When [specific event] occurs in [app]..."
●​ File/Email: "When new [file/email] is received..."

3. Data Flow Specification (Critical)

Unset
"The workflow should:
1. [Input step] - describe data source and expected format
2. [Processing step] - describe transformations, conditions,
loops
3. [Output step] - describe destination and required format
4. [Error handling] - describe fallback behavior"

📋 Essential Components Checklist

Trigger Details

●​ [ ] Trigger Type: Clearly specify (webhook, schedule, manual, etc.)

●​ [ ] Trigger Configuration: Include specific settings (frequency, endpoints, etc.)
●​ [ ] Expected Input Format: Define data structure/schema

Data Processing

●​ [ ] Field Mapping: Specify which input fields map to which outputs

●​ [ ] Data Transformations: Detail any formatting, calculations, or modifications
●​ [ ] Conditional Logic: Define if/then scenarios with specific conditions
●​ [ ] Loops/Iterations: Specify when to process arrays or repeat actions

Integrations & Actions

●​ [ ] Target Applications: Name specific services (Discord, Airtable, etc.)

●​ [ ] Required Operations: Be specific (create record, send message, update row)
●​ [ ] Authentication Needs: Mention credential requirements
●​ [ ] Data Format Requirements: Specify expected formats for each integration

Error Handling & Quality

●​ [ ] Error Scenarios: Define what should happen when things fail

●​ [ ] Retry Logic: Specify retry attempts and conditions
●​ [ ] Logging Requirements: Define what should be logged and where
●​ [ ] Validation Rules: Specify data quality checks

🏗️ Template Structures
Simple Linear Workflow

Unset
"Create a workflow that:
- Triggers: [specific trigger]
- Processes: [data transformation]
- Outputs: [specific destination]
- Handles errors by: [fallback action]"

Complex Multi-Path Workflow

 Unset
"Create a workflow that:
- Triggers: [trigger details]
- Checks if: [condition]
- If true: [path A actions]
- If false: [path B actions]
- For each item: [loop logic]
- Finally: [completion action]
- Error handling: [fallback strategy]"

Integration-Heavy Workflow

Unset
"Create a workflow that connects:
- Source: [app] via [method] expecting [data format]
- Processing: [transformation rules]
- Destinations:
- [App 1]: [specific action and data format]
- [App 2]: [specific action and data format]
- Include proper authentication placeholders and error handling"

⚠️ Common Pitfalls to Avoid

Vague Descriptions

❌ Don't say: "Process some data and send it somewhere" ✅ Do say: "Extract 'name', 'email',
and 'phone' fields from webhook payload, validate email format, and create contact in HubSpot"

Missing Error Context

❌ Don't say: "Handle errors" ✅ Do say: "If API call fails, retry 3 times with 5-second delays,
then log error to Google Sheets and send alert to Slack #alerts channel"

Unclear Data Flow

❌ Don't say: "Get data from Google Sheets" ✅ Do say: "Monitor Google Sheet 'Leads' for
new rows, extract columns A-E (name, email, phone, company, status), filter for status='new'"
🔧 Advanced Prompt Modifiers
For Complex Logic

Unset
"Include branching logic that:
- Evaluates [specific condition]
- Routes to different paths based on [criteria]
- Merges results using [merge strategy]"

For Data Transformation

Unset
"Transform data by:
- Converting [field] from [format A] to [format B]
- Calculating [formula] using [input fields]
- Filtering records where [condition]
- Grouping by [field] and aggregating [metrics]"

For Enterprise Features

Unset
"Include production-ready features:
- Comprehensive error handling with retry logic
- Detailed logging for audit trails
- Data validation and sanitization
- Rate limiting considerations
- Credential management placeholders"

📝 Example High-Quality Prompts

Example 1: E-commerce Order Processing

Unset
"Create an n8n workflow that processes new e-commerce orders:
 TRIGGER: Webhook from Shopify when new order is placed
INPUT: Order JSON with customer info, items, totals

PROCESSING:
1. Validate order data (required fields present, valid email
format)
2. Check if customer exists in CRM (HubSpot)
3. If new customer: create contact record
4. Calculate shipping cost based on weight and location
5. Check inventory levels for each item

OUTPUTS:
- Create order record in Airtable with all details
- Send order confirmation email via SendGrid
- Post to Slack #orders channel with summary
- If low inventory: alert warehouse team via email

ERROR HANDLING:
- Invalid data: log error and send to dead letter queue
- API failures: retry 3x with exponential backoff
- Critical errors: alert operations team immediately

Include proper field mappings, authentication placeholders, and

documentation sticky notes."

Example 2: Social Media Monitoring

Unset
"Create an n8n workflow for social media brand monitoring:

TRIGGER: Schedule (every 15 minutes)

PROCESSING:
 1. Search Twitter API for mentions of brand keywords:
['@ourcompany', 'our product', '#ourhashtag']
2. Filter out retweets and replies older than 1 hour
3. Analyze sentiment using built-in sentiment analysis
4. Categorize mentions: support request, complaint, praise,
general

ROUTING LOGIC:
- Negative sentiment + support keywords → create Zendesk ticket
- High engagement (>100 likes/retweets) → notify marketing team
- Influencer mentions (>10k followers) → add to influencer CRM
- All mentions → log to Google Sheets dashboard

OUTPUTS:
- Zendesk: Create ticket with tweet content, user info, sentiment
score
- Slack: Post high-priority mentions to #social-alerts
- Google Sheets: Log all mentions with timestamp, sentiment,
category
- Email: Daily digest to marketing team

Include rate limiting, duplicate detection, and comprehensive

error handling."

🎖️ Quality Indicators
A well-crafted prompt should result in workflows that include:

●​ Specific node types mentioned by name

●​ Proper expression syntax for data references
●​ Error handling nodes (IF, Try/Catch, etc.)
●​ Sticky notes with clear documentation
●​ Realistic placeholder values
●​ Proper connection mapping
●​ Authentication considerations
●​ Production-ready structure
🚀 Pro Tips
1.​ Be Specific About Data Formats: Always specify expected JSON structure, field
names, and data types
2.​ Define Success Criteria: Clearly state what constitutes successful completion
3.​ Include Edge Cases: Mention scenarios like empty responses, API limits, or missing
data
4.​ Specify Timing: Include details about delays, timeouts, and scheduling requirements
5.​ Reference Real Tools: Use actual application names and their specific API capabilities
6.​ Think Production: Consider monitoring, logging, and maintenance requirements

This framework ensures your prompts generate robust, importable n8n workflows that work in
real-world scenarios.