product updateAnthropic

Anthropic Adds Cross-Session Messaging to Claude Code v2.1.224

TL;DR

Claude Code v2.1.224 introduces cross-session messaging, letting separate Claude Code instances on macOS and Linux send each other summaries to coordinate work. The feature does not support approving permissions or executing commands remotely.

2 min read
0

Anthropic has released Claude Code v2.1.224 for macOS and Linux, adding a feature that lets separate Claude Code sessions send messages to one another. The update targets users who run multiple sessions simultaneously on different parts of the same project.

What changed

According to Anthropic's documentation, cross-session messaging lets Claude "deliver a message from one of your Claude Code sessions to another." The company gives a specific example: when a change in one session breaks work another session is building on, Claude can warn that session before the user notices. Similarly, if one session resolves a question another session is blocked on, the answer can be sent across automatically.

Users don't have to write the message text themselves. Anthropic says a prompt only needs to describe the content to convey, and Claude composes the actual message sent to the other session.

The feature was also detailed in a post from the ClaudeDevs account on X, which noted that instead of having to "re-explain yourself in another session," users can now tell Claude to do it. The account specified that Claude sends a summary — not full history or files — and the receiving session "picks it up mid-task."

Anthropic lists four primary use cases:

  • Handing off a finding from one session to another
  • Coordinating parallel worktrees on the same codebase
  • Getting a status update from a long-running session
  • Replying across different machines

Limitations

Cross-session messaging has explicit boundaries. It cannot be used to approve permission requests or change configurations. Commands like /compact arrive at the receiving session as plain text rather than executable instructions. If acting on a received message requires a permission grant, the receiving session will still prompt the user directly — the messaging system cannot bypass that approval step.

Anthropic has not disclosed pricing changes tied to this update, and the feature appears to be a capability addition to the existing Claude Code product rather than a new paid tier.

What this means

Developers running multiple Claude Code instances — a common pattern for those working across microservices, monorepos, or parallel git worktrees — previously had to manually copy context between sessions or re-explain state each time. Cross-session messaging automates that handoff, which matters most for teams treating Claude Code as a semi-autonomous collaborator across a larger codebase rather than a single-session assistant.

The restriction on permissions and command execution is a deliberate guardrail: Anthropic is limiting the feature to information transfer, not delegated authority. That keeps a human in the loop for any action with real consequences, even as sessions gain more autonomy to coordinate among themselves. It's a narrow but practical step toward multi-agent workflows within a single developer's toolchain, rather than a broader agent-to-agent framework.

Related Articles

product update

Anthropic Sets Claude Code Auto Mode as Default Starting August 14

Anthropic will switch Claude Code's default permission setting to auto mode on August 14 for Pro, Max, and Team users. The company says its safety classifier caught 89% of dangerous commands in testing, compared to 13.6% for human reviewers, and will no longer charge extra tokens for the classifier itself.

product update

GitHub Publishes Guide to Slash Commands in the Copilot App

GitHub has published a guide covering slash commands available in the GitHub Copilot app, designed to extend Copilot beyond simple chat into planning, team collaboration, task automation, and workflow customization. The guide targets developers looking to get more structured, repeatable value out of Copilot's interface.

benchmark

Composio Benchmark: Claude Code Fastest Agent Framework, But Costs Nearly 3x More Than OpenCode

Composio benchmarked DeepSeek V4 Flash across four agent frameworks—Claude Code, Codex, OpenCode, and Oh My Pi—on 30 real-world tasks. Claude Code finished fastest at 122 seconds per task but cost $0.195, nearly three times OpenCode's $0.073, while Oh My Pi had the highest success rate at 17/30 but took 272 seconds per task.

product update

AWS Details Two Paths for Single-Region Claude Code Deployments on Amazon Bedrock

AWS published a technical guide detailing two methods for keeping Claude Code inference confined to a single AWS Region: Anthropic's newer Mantle endpoint and the classic Bedrock Invoke API with application inference profiles. The right path depends entirely on which Region compliance teams require.

Comments

Loading...