List the tool approval queue
List the human-in-the-loop tool approval queue — one row per tool invocation the agent paused on, each with the conversation and agent that requested it, the tool id, the (redaction-safe) tool arguments with the operator-facing cost estimate extracted, and the current decision state. Filters: status (pending default; approved/rejected for history) and limit (1–50, default 50), newest first. On a transient read failure the endpoint degrades to an empty queue instead of erroring on a polled view. Owner or admin role required.
Authorizations
Dashboard JWT token from Clerk
Headers
Sandbox opt-in for Clerk-session-authenticated requests. Set to true to route the call through the test-mode pipeline: no real provider delivery, no credits deducted, response meta.test_mode: true. Ignored for live API keys (dv_live_sk_*) — server-to-server clients must use a test-prefixed key (dv_test_sk_*) to exercise sandbox. Test-prefixed keys unconditionally enable sandbox regardless of this header.
true, false Response
The approval slice — approvals (each with id, conversation_id, agent_id, tool_id, tool_args, cost_estimate, requested_at, status, decided fields), the resolved status, and count — in the standard { data, meta } envelope.