9b3a5c891c
## What This PR adds the AIR client capability `backgroundSubagents`. A client declares it in `initialize.clientCapabilities._meta.jetbrains.air.capabilities`. With the capability, an ACP v1 prompt turn ends at its result. This is also true while background subagents of the turn still run, including a subagent that SendMessage resumed. Subagent updates, permission requests and elicitations then go out after the prompt response. Their routing does not change. The agent advertises the capability only to ACP v1 AIR clients. ## Why The hold of #864/#866 keeps `session/prompt` open until the subagents finish. The Claude CLI ends the model turn at `result`, and only `idle` waits for the background agents. In AIR the hold hides the end of the answer, and the next user message steers into an old turn. One SendMessage-resumed subagent kept a turn open for about two hours. ## Unchanged - A client without the capability keeps the hold exactly. So #866 stays fixed for clients that stop listening at the prompt response. - ACP v2 keeps the hold too. ## Related changes - Out of a turn, the ExitPlanMode request of such a client does not offer the clear-context option. The fresh query must continue an ACP turn, and there is none. - An ExitPlanMode interruption of an autonomous cycle with no held turn now clears its markers. ## Docs `docs/air-extensions.md` has a new "Background subagents" section. ## Tests New capability cases: - a fresh subagent; - a SendMessage-resumed subagent; - a permission request after the prompt response, with the `parentToolUseId` attribution; - an autonomous followup that does not end the queued next prompt; - ExitPlanMode out of a turn, with and without a reset; - the exclusion of a running subagent from the incomplete-tool check. The existing tests cover the path without the capability. When the gate is disabled, 7 of the 8 new cases fail. The eighth is the in-turn control case and passes in both configurations. `npm run build`, `npm run check` and `npm run test:run` pass. The exception is 4 `providers.test.ts` cases. They depend on the environment (a local proxy `baseUrl`) and fail on `main` too.
ACP adapter for the Claude Agent SDK
Use Claude Agent SDK from ACP-compatible clients!
This tool implements an ACP agent by using the official Claude Agent SDK, supporting:
- Context @-mentions
- Images
- Tool calls (with permission requests)
- Compact file changes through the negotiated AIR diff patch extension
- Following
- Edit review
- TODO lists
- Nested subagent transcripts
- Interactive (and background) terminals
- Custom Slash commands
- Client MCP servers
/mcpin the chat: the MCP server status as a list. The adapter runs/mcp reconnect,/mcp enable, and/mcp disablethrough the SDK control API, because Claude Code refuses them in SDK mode. A reconnect of an ACP server that needs authentication starts MCP OAuth through URL elicitation- Session-scoped long-running goals for AIR through the goal extension under
_meta.jetbrains.air.goal - Structured errors, recovery, and warnings through the opt-in session failure extension
- Concrete model and effort defaults through the opt-in recommended config value extension
- Tool permission presentation, editable choices, and durable effects through the permission extension
- One fact per field in every tool call report, as the ACP tool call contract defines
- All AIR extensions, capabilities, and
_metakeys: AIR extensions
Learn more about the Agent Client Protocol.
To try changes that have landed on main but are not released yet, install from the
preview channel — every push to main publishes one. See
docs/RELEASES.md.
npm install @agentclientprotocol/claude-agent-acp@preview
Subagent sessions
A client that declares clientCapabilities.subagents, other than JetBrains AIR, gets each Agent or
Task subagent as its own child session, following the subagents RFD:
subagent_updateon the parent announces the child before any of its traffic, with the parent'stitleanddescriptionfor it, and reports its work state:running,requires_actionwhile one of its requests is open (andrunningagain when the last one ends, unless a cancel withdrew it: the child is stopping then), andidlewhen its work ends, withend_turnorcancelledwhen the SDK gives the reason. A failure isidlewith theerrorstop reason and the SDK's error text as itserror. The child keeps its session for every delegation to it.- The prompt of the Agent or Task call that launches the child, and the message of each
SendMessage call that resumes it, is a
session_messagein the child's transcript. The tool call stays a tool call of the session that made it. session/cancelon the parent interrupts Claude Code, which stops the foreground children that the turn waits on. Background children keep running, as the RFD prefers; a client stops one with its owncancel. Each stopped child reportsidlewithcancelledfrom Claude Code's own events, and its open requests are withdrawn. The adapter ends no child at a cancel itself, so a child that keeps running keeps its open requests and its open tool calls. To spare background children, the adapter declares the SDK optionperTaskStopAffordance, which a client can turn off in_meta.claudeCode.options. That option spares background workflows too, which a client cannot stop, so the adapter stops them after the interrupt, as Claude Code did without it. Background commands keep running after a cancel either way. A tool call of a stopped child that Claude Code reports no result for, such as the command it was running, ends in the child's session before the cancelled prompt resolves. When Claude Code stops a background child after that, or while no prompt runs, the tool call ends right after the child reportsidle. A child that still runs when Claude Code exits isidlewith no stop reason.- A child whose parent is not yet known when it ends is not exposed, since the RFD forbids guessing its parent.
- A replay restores each child under its live session id, from its own transcript: its prompt and
each SendMessage message are
session_messages, and its end is anidlebefore the response. A launch without Claude Code's metadata file for the child stays an ordinary tool call of the parent, and a notice, for a client that takes notices, says that subagent history is missing. - After a load or resume, with or without replay, a SendMessage that resumes an agent from before it runs that agent in its session from before. Before the agent's traffic, each of its ancestors that the client was not told about in this connection is announced, parent before child and with no work state, as the RFD requires. An agent whose launch the session's history does not record stays unexposed.
- Each running child is announced with the
cancelcapability.session/cancelwith the child's session id stops that child's current work, through the SDK'sQuery.stopTask, and not its parent's: Claude Code also stops the foreground children it waits on, and the parent's turn goes on. Each stopped child reportsidlewithcancelledfrom Claude Code's own events, and its requests that the stop withdraws report norunningfirst. A replayed child, which no longer runs, gets no capabilities. A client cannot prompt or otherwise change a child.
An ACP v2 client gets the same without declaring anything, since subagents
are part of v2's baseline. Each child also reports its work state as its own state_update, right
after the parent's copy.
JetBrains AIR keeps the earlier draft of the RFD (subagent_spawned, subagent_state_update),
which it enables with nativeSubagentSessions in _meta.jetbrains.air.capabilities. See
AIR extensions.
Without either signal, Agent and Task stay ordinary tool calls and child interactions stay on the
root session. Clients that use the historical _meta["subagent-transcript"] capability or the
forwardSubagentText session option retain the flattened child transcript.
Experimental ACP v2
The adapter also serves the draft ACP v2:
a client that asks for protocol version 2 or later in initialize gets v2, and every other client
gets v1. The draft can still change, and this surface changes with it.
CLAUDE_AGENT_ACP_PROTOCOL limits the adapter to one version. With v1, a client that asks
for v2 is negotiated down to v1. With v2, the adapter rejects the initialize of a v1 client.
auto, the default, serves both.
On v2 the adapter serves the session baseline (session/new, session/list, session/resume
with replay, session/close, session/prompt, session/cancel), session/delete,
session/fork, additional directories, config options (the mode is one), plans, notices,
compaction, display terminals for commands, permission requests, elicitation, providers, auth, and
subagents. Steering and the AIR extensions are v1 only for now.
Contribution Policy
This project does not require a Contributor License Agreement (CLA). Instead, contributions are accepted under the following terms:
By contributing to this project, you agree that your contributions will be licensed under the Apache License, Version 2.0. You affirm that you have the legal right to submit your work, that you are not including code you do not have rights to, and that you understand contributions are made without requiring a Contributor License Agreement (CLA).