ReleaseBytes
anthropic Claude Platform Release Notes ·

Anthropic Enhances Claude Managed Agents with `auto` Policy, CLI Connect

aipreviewengineer
feature

Anthropic has introduced new capabilities for Claude Managed Agents, enhancing control over tool execution and real-time session management. A new `auto` permission policy allows the server to evaluate and run, deny, or pause agent tool calls, providing granular control over agent behavior. Additionally, the `ant` CLI now includes `ant beta:sessions connect`, enabling developers to connect to and interact with agent sessions live, including approving or denying pending tool calls. These updates improve operational oversight and developer interaction with managed agents.

  • →New `auto` permission policy for Claude Managed Agents
  • →`ant` CLI adds live session connection for Managed Agents
Features (2) ›
  • New `auto` permission policy for Claude Managed Agents

    Claude Managed Agents permission policies now include `auto`, which allows the server to evaluate each agent or MCP tool call and either run it, deny it, or pause for user approval. The `agent.tool_use` and `agent.mcp_tool_use` events now report how each call was evaluated in an `evaluation` field alongside `evaluated_permission`.

  • `ant` CLI adds live session connection for Managed Agents

    The `ant` CLI introduces `ant beta:sessions connect`, a new command that attaches your terminal to a Claude Managed Agents session. This enables users to follow the session live, send messages, and allow or deny tool calls that are awaiting approval, with an option to serve the Claude Console's session viewer locally using `--web`.

Read the original announcement →

https://platform.claude.com/docs/en/release-notes/overview

Related releases