Super Agent Block YAML Schema

Schema Definition

type: object
required:
  - type
  - name
properties:
  type:
    type: string
    enum: [ask_agent]
    description: Block type identifier
  name:
    type: string
    description: Display name for this Super Agent block
  inputs:
    type: object
    properties:
      systemPrompt:
        type: string
        description: Optional system instructions
      userPrompt:
        type: string
        description: Task or question (can reference other blocks)
      skillIds:
        type: array
        description: User skill IDs appended to the system prompt at runtime
        items:
          type: string
      model:
        type: string
        description: Lite model override (e.g. gpt-5.4, claude-opus-4-8)
      depth:
        type: integer
        minimum: 0
        maximum: 3
        description: Copilot depth when model is not explicitly set
      allowMutations:
        type: boolean
        description: Allow mutating tool operations (default false)
      maxToolRounds:
        type: integer
        description: Maximum tool call rounds (default 20)
      memories:
        type: string
        description: Memory context from memory blocks
    required:
      - userPrompt
  connections:
    type: object
    properties:
      success:
        type: string
        description: Target block ID for successful execution
      error:
        type: string
        description: Target block ID for error handling

Connection Configuration

connections:
  success: <string>   # Target block ID for successful execution
  error: <string>     # Target block ID for error handling (optional)

Examples

Read-only research

research-agent:
  type: ask_agent
  name: "News Research"
  inputs:
    systemPrompt: "You are a concise research assistant. Cite sources when possible."
    userPrompt: "Summarize the latest news about <start.input> in 3 bullet points."
    depth: 1
    allowMutations: false
    maxToolRounds: 15
  connections:
    success: response-block

response-block:
  type: response
  name: "Final Response"
  inputs:
    content: <research-agent.content>

With Skills and explicit model

support-agent:
  type: ask_agent
  name: "Support Triage"
  inputs:
    userPrompt: "Analyze this customer message and suggest next steps: <start.input>"
    skillIds:
      - "skill_support_playbook"
    model: gpt-5.4
    allowMutations: false
  connections:
    success: route-block

Production automation (mutations enabled)

notify-agent:
  type: ask_agent
  name: "Daily Digest"
  inputs:
    userPrompt: "Search for today's top AI headlines and post a summary to Slack #general."
    allowMutations: true
    maxToolRounds: 25
  connections:
    success: done-response

Output References

Reference Super Agent outputs from other blocks:

content: <research-agent.content>
model: <research-agent.model>
tokens: <research-agent.tokens>
toolCalls: <research-agent.toolCalls>

Notes

Super Agent uses copilot/lite billing, not provider API keys on the block.
OAuth integrations must be connected in workspace Settings before mutating or private tools succeed.
See the Super Agent block guide for tool scope, mutation policy, and troubleshooting.

Super Agent Block YAML Schema

On this page