Skills Overview

Skills are JavaScript functions that extend your AI agent's capabilities. Once deployed, the AI can invoke them as tools during conversations.

Based on the official Skills documentation: Skills Overview - Cloodot

What skills can do

Custom JavaScript/TypeScript functions let AI agents:

• Fetch and present data (informational)
• Perform operations that change state (action)
• Present interactive UI for continued interaction (interactive)

Skill types

Informational Skills

Fetch and return data without side effects.

Examples:

• Weather lookup
• Stock prices
• Knowledge base search
• Inventory check

Characteristics:

• Read-only operations
• No external state changes
• Fast execution
• Return data to display

Action Skills

Perform operations that change external state.

Examples:

• Create calendar events
• Send emails
• Update CRM records
• Process orders

Characteristics:

• Modify external systems
• Require configuration/credentials
• May have side effects
• Return confirmation/results

Interactive Skills

Present UI elements for continued user interaction.

Examples:

• Multi-step forms
• Approval workflows
• Guided tutorials
• Customer surveys

Characteristics:

• Collect user input
• Enable ongoing conversation
• Dynamic response generation
• Maintain interaction state

Tool call flow

When a user requests an action, skills execute through this flow:

1. Tool discovery

   • The AI agent analyzes available skills.
   • It examines skill parameters and descriptions.

2. Parameter extraction

   • The agent extracts relevant data from conversation context.
   • It maps the user request to skill parameters.

3. Skill invocation

   • The platform executes the skill with the extracted parameters.
   • It runs inside an isolated sandbox.

4. Response processing

   • Cloodot validates and formats the output.
   • Errors are captured and returned.

5. Context integration

   • Results flow back into the conversation.
   • The AI generates a response to the user.

When skills are called

Skills trigger in these scenarios:

• User Request: User explicitly asks for an action ("book a meeting", "check inventory")
• Contextual AI Decision: AI determines that invoking a skill would help resolve the conversation
• UI Interaction: User clicks skill-related buttons or carousels
• Future: Scheduled or event-driven executions (planned feature)

Skill context

When a skill executes, it receives rich context about the conversation:

{
  "conversationMetadata": {
    "id": "conv_12345",
    "participantInfo": "customer details",
    "channelType": "whatsapp"
  },
  "organizationContext": {
    "settings": "org settings",
    "timezone": "America/New_York",
    "customFields": {}
  },
  "messageHistory": "recent messages for context",
  "userProfile": {
    "contactInfo": "customer info",
    "preferences": "user preferences"
  }
}

This context enables skills to provide personalized, relevant responses.

Skill definition fields

A skill definition includes:

• slug: Unique identifier (alphanumeric + underscores)
• name: Display name
• description: What the skill does
• prompt: How to describe it to the AI
• definition: JavaScript/TypeScript code with handler function
• parameters: Input schema expected from AI
• response: Output schema returned to AI
• buttons: Optional quick action buttons

Handler function

All skills require a handler function:

async function handler(input) {
  const { config, parameters, context } = input
  
  // Your skill logic here
  
  return {
    message: "Response to AI",
    // ... other fields
  }
}

Input structure

{
  config: Record<string, any>           // Configuration values
  parameters: Record<string, any>       // Input from AI tool call
  context: {
    conversationId?: string
    organizationId?: string
    // ... organization context
  }
}

Further reading

Build and deploy

• Creating Skills — Step-by-step guide to building your first skill
• Skill Schema — Skill definition format and validation rules
• Skill Sets — Organize skills into reusable Skill Sets
• AI Agent Configuration — Configure your agent to use custom skills
• MCP Servers — Connect skills to external services via MCP

Related features

• Knowledge Base — Enhance skills with knowledge base integration
• Custom Fields — Use custom fields in skill parameters and responses
• Integrations — Connect skills to external APIs and services

Next steps