Agent run outcomes and human corrections from observability tools (connector planned).
Agent framework and evaluation teams. Demand varies; we do not promise a sale for any dataset.