AgentGuidanceRunTest

What an agent answers about the input of one past run, with guidance that has not been saved.

type AgentGuidanceRunTest {
  ai_agent_run_id: ID!
  usable: Boolean!
  answer_summary: String
}

Fields

AgentGuidanceRunTest.ai_agent_run_id ● ID! non-null scalar

AgentGuidanceRunTest.usable ● Boolean! non-null scalar

False when the agent's answer could not be used; answer_summary is then null.

AgentGuidanceRunTest.answer_summary ● String scalar

In the wording of AgentPromptEvaluation.answer_summary, so the two can be compared.

Returned By

testAgentGuidanceOnRun mutation