Member Junction
    Preparing search index...

    Type Alias AgentContextData

    Context data provided to agent prompts during execution.

    This data structure is passed to the prompt templates and contains information about the agent, its sub-agents, and available actions. The sub-agent and action details are JSON stringified to work with the template engine.

    type AgentContextData = {
        actionCount: number;
        actionDetails: string;
        agentDescription: string | null;
        agentName: string | null;
        appContext?: string;
        clientToolDetails?: string;
        parentAgentName?: string | null;
        planApproved?: boolean;
        planModeActive?: boolean;
        skillCount?: number;
        skillsCatalog?: string;
        subAgentCount: number;
        subAgentDetails: string;
    }
    Index

    Properties

    actionCount actionDetails agentDescription agentName appContext? clientToolDetails? parentAgentName? planApproved? planModeActive? skillCount? skillsCatalog? subAgentCount subAgentDetails

    Properties

    actionCount: number

    Number of actions available to this agent

    actionDetails: string

    Markdown formatted details of available actions (name, params, result codes)

    agentDescription: string | null

    Description of the agent's purpose and capabilities

    agentName: string | null

    The name of the agent being executed

    appContext?: string

    Markdown formatted snapshot of the user's current application context

    clientToolDetails?: string

    Markdown formatted details of available client tools (name, category, description, input schema)

    parentAgentName?: string | null

    Optional parent agent name for sub-agent context

    planApproved?: boolean

    Whether Plan Mode's gate has already been satisfied (a prior plan was approved on this run chain)

    planModeActive?: boolean

    Whether Plan Mode is active for this run (SupportsPlanMode + per-request planMode + root agent)

    skillCount?: number

    Number of skills available to this agent (per AIAgent.AcceptsSkills — 0 when 'None')

    skillsCatalog?: string

    Markdown formatted CATALOG of available skills — name + description ONLY (progressive disclosure). Full Instructions are never injected here; they're appended to context only on activation (see executeSkillStep in @memberjunction/ai-agents).

    subAgentCount: number

    Number of sub-agents available to this agent

    subAgentDetails: string

    Markdown formatted list of sub-agent names and descriptions

    Settings

    Member Visibility

    On This Page

    Properties