Variables and References - Superwire

References use dot paths.

input.project_id
secrets.api_key
agent.summarize.summary
dynamic.context
schema.shared.status

Root namespaces

Root	Meaning
`input`	Values from the request body.
`secrets`	Sensitive values from the request body.
`agent`	Structured outputs from agents.
`dynamic`	Values computed by dynamic blocks.
`tool`	Imported MCP tools.
`resource`	Imported MCP resources.
`prompt`	Imported MCP prompts.
`schema`	Schema declarations and nested reusable fields.
`model`	Named model profiles.

Loop variables

Loop variables are available only inside the looped agent.

agent summarize_each for task in input.tasks {
    model: model.fast
    instruction: "Summarize task {{ task.title }}."

    output {
        title: string
        summary: string
    }
}

Object-shaped items may be destructured in the loop header.

agent summarize_each for { a, b } in input.something {
    model: model.fast
    instruction: "Summarize A={{ a }} and B={{ b }}."

    output {
        summary: string
    }
}