Member Junction
    Preparing search index...

    Type Alias AgentSkillActivationRequest

    A skill the agent's response requested be activated (by catalog name — the agent only ever sees name + description in its prompt, per progressive disclosure).

    type AgentSkillActivationRequest = {
        name: string;
        reason?: string;
    }
    Index

    Properties

    Properties

    name: string

    Name of the skill (MJ: AI Skills.Name) to activate

    reason?: string

    Brief agent-stated rationale for why this skill is being activated. Optional — supplied by the LLM when it self-activates a skill mid-run, and carried through to the run step's AgentSkillInvocation record for observability. User-requested activations (/skill mentions) don't carry a reason.