Member Junction
    Preparing search index...

    Interface ConversationToolSummaryHost

    The seam through which summarizeRange runs its recursive LLM sub-call. Implemented by BaseAgent (which owns prompt execution and run-step observability) so this manager stays free of prompt-runner dependencies and trivially testable.

    interface ConversationToolSummaryHost {
        RunSummaryPrompt(
            rangeText: string,
            lens: string,
        ): Promise<{ promptRunId?: string; text: string }>;
    }
    Index

    Methods

    RunSummaryPrompt

    Methods

    • Runs the range-summary prompt over the rendered slice.

      Parameters

      • rangeText: string

        The [seq N] Role: text rendering of the requested range

      • lens: string

        The task-specific focus the summary must serve

      Returns Promise<{ promptRunId?: string; text: string }>

      The summary text plus the AIPromptRun ID of the sub-call (for lineage)

    Settings

    Member Visibility

    On This Page

    Methods