Claude Code Streams Subagent Activity Live to Remote Control
Claude CodeClaude Code 2.1.251 lets Remote Control clients (the browser and mobile companion for a running local session) watch a foreground subagent's tool calls and results as they happen, rather than only a status indicator. Background subagents still show status only, keeping the busiest sessions from flooding a phone screen. The change makes Remote Control a genuine window into what an agent is doing while it runs, not just whether it is still working.
Key Takeaways
- Foreground subagent tool calls and results now stream live to Remote Control clients on the browser or mobile app, not just a status badge.
- Background subagents deliberately keep the lighter status-only view, preventing sessions with many parallel agents from overwhelming a phone screen.
- The feature closes a long-standing gap between the terminal experience and the remote/mobile experience for anyone checking in on a session away from their desk.
- Remote Control keeps execution on the original machine, so the streamed detail is a live mirror rather than a separate cloud run.
- It ships alongside a related change letting connected devices see a session's working-tree diff, part of a broader push to make remote monitoring as informative as sitting at the terminal.
- No new setup is required: existing Remote Control sessions (
/remote-control) pick up the richer subagent view automatically once updated to 2.1.251 or later.
Remote Control lets a Claude Code session running on a laptop stay connected to claude.ai or the Claude mobile app, so a developer can check in from a phone or another browser while execution stays on their own machine. Until 2.1.251, that view showed subagent activity only as a status indicator: working, idle, or stopped, with no detail on what a subagent was actually doing.
What Changed
As of Claude Code 2.1.251, released August 28, 2026, a foreground subagent's tool calls and their results now stream live to any connected Remote Control client. Someone watching from a phone can see the same file reads, edits, and command output that would scroll past in the terminal, in close to real time.
Anthropic drew a deliberate line here: background subagents, which are the default for most delegated work, still show status only rather than a full stream. That keeps a session running many parallel background agents from turning a phone screen into an unreadable wall of tool calls, while a subagent someone is actively watching in the foreground gets full visibility.
Why It Matters
Remote Control has increasingly become how people check on long-running Claude Code sessions away from their desk. Full subagent streaming closes the biggest gap between the terminal experience and the remote one: previously, stepping away from the laptop meant losing the detail needed to judge whether a subagent was on track, and now that detail follows the session to whatever device is watching it.