Registry

The Registry category has 2 tools:

• agent_list — List the agents the authenticated caller owns: each one's id, the name the owner gave it, its description, and the MCP client id it registered with. Takes no arguments, costs nothing, and reads the caller's own records only. id is opaque and never meaningful — quote it, do not parse it. mcpClientId is reported for attribution and is NOT a unique key, because two agents on one host may send the same clientInfo.name; key on id. A deleted agent is absent from this list, but its past spend still appears in agent_usage under the name it had, because a month-old charge belongs to the agent that made it. This tool does NOT run an agent and does NOT call a tool on an agent's behalf: a metered agent call spends the caller's credits, and that is a charge the run approval path exists to gate.
• agent_usage — Report what the caller's agents spent over a window: one row per agent that spent, most expensive first, with its call count, its credits, and the first and last charge in the window — plus the totals across every agent. Free, and it reads the credit ledger rather than writing to it. Two fields to read carefully: totalCredits is the caller's spend across EVERY agent, so it does not change when you filter, and an agent that spent nothing in the window is NOT a row — its absence means zero, not missing. A soft-deleted agent still reports here under the name it had, because a charge belongs to the agent that made it whether or not it is still registered. Refunds do not subtract from a row: only spends count against an agent, and a refunded call that happened still counts as a call.

The full request and response shape for each, and a live playground, live on the generated reference at GET /mcp/tools.