Skip to Content
MCP ServerPrompts Reference

Prompts Reference

MCP prompts are pre-built templates that help AI agents tackle common VelaFlows tasks. They inject the right context (API schemas, node definitions, event data) so the AI gives accurate, platform-specific answers.

How Prompts Work

When you invoke a prompt, the AI receives a structured message with VelaFlows-specific context injected. This is more effective than asking a generic question because the AI has the exact API schemas and node configurations it needs.

Available Prompts (6)

build-integration

Generate a step-by-step integration plan.

ArgumentTypeRequiredDescription
descriptionstringYesWhat you want to integrate (e.g., “sync leads to Salesforce”)

What it does:

  1. Searches VelaFlows endpoints matching your description keywords
  2. Injects the relevant API schemas into context
  3. Returns a structured plan: authentication setup, endpoint calls, error handling, code example

Example: “Use the build-integration prompt with: sync new conversations to a Google Sheet”


build-workflow

Suggest a workflow automation from a plain English description.

ArgumentTypeRequiredDescription
descriptionstringYesWhat the workflow should do (e.g., “auto-reply to new leads”)

What it does:

  1. Searches the node catalog and event catalog for relevant nodes and triggers
  2. Injects matching node configurations into context
  3. Returns a suggested workflow DAG: trigger selection, node chain, config for each node, connection map

Example: “Use the build-workflow prompt with: when a conversation is idle for 30 minutes, escalate to the team lead and send a Slack notification”


debug-api-error

Diagnose API errors with endpoint-specific context.

ArgumentTypeRequiredDescription
statusCodenumberYesHTTP status code (e.g., 400, 401, 404, 422)
endpointstringYesThe endpoint that returned the error (e.g., POST /crm/leads)
responseBodystringNoThe error response body (if available)

What it does:

  1. Looks up the endpoint’s schema from the bundled OpenAPI specs
  2. Identifies common causes for that status code on that specific endpoint
  3. Returns a diagnostic checklist: required fields, auth requirements, validation rules, common mistakes

Example: “Use the debug-api-error prompt with statusCode 422, endpoint POST /crm/leads, and this response body: {error details}


generate-code

Generate working code for a VelaFlows API task.

ArgumentTypeRequiredDescription
languagestringYesjavascript, python, or curl
taskstringYesWhat the code should do (e.g., “list all open conversations”)

What it does:

  1. Searches for relevant endpoints based on the task description
  2. Injects the correct authentication pattern for the chosen language
  3. Returns complete, runnable code with proper headers, error handling, and response parsing

Example: “Use the generate-code prompt with language javascript and task: create a lead with name and email, then move it to the first pipeline stage”


explain-event

Get a comprehensive explanation of a platform event.

ArgumentTypeRequiredDescription
eventTypestringYesEvent type (e.g., inbox.conversation.created, crm.lead.stage_changed)

What it does:

  1. Looks up the event in the bundled catalog
  2. Returns: what triggers the event, expected payload structure, common use cases, related events, and workflow integration patterns

Example: “Use the explain-event prompt with: crm.lead.stage_changed”


migrate-webhook

Help migrate from another platform’s webhooks to VelaFlows.

ArgumentTypeRequiredDescription
fromPlatformstringYesSource platform (e.g., “HubSpot”, “Intercom”, “Zendesk”)
webhookPayloadstringYesExample webhook payload from the source platform

What it does:

  1. Analyzes the source webhook payload to identify fields and event type
  2. Searches VelaFlows events for equivalent matches
  3. Returns: field mapping (source → VelaFlows), equivalent event subscriptions, migration steps

Example: “Use the migrate-webhook prompt with fromPlatform HubSpot and this webhook payload: {paste HubSpot webhook JSON}