---
title: "Claude Tag brings a proactive Claude into team chat"
description: "Anthropic launched Claude Tag in Slack for team delegation, scoped memory and scheduled work."
date: "2026-06-23"
tags: ["Claude Code", "agents", "collaboration"]
canonical: "https://news.maestromojo.com/news/claude-tag-enters-the-team-chat/"
---

> **Maestro’s take:** Anthropic wants Claude to become the teammate who never leaves Slack. Convenient. Also a fast way to turn channel sprawl into a permission problem. Start in one private engineering channel. Audit the memory before inviting the whole company.

## TL;DR

Claude Tag is a team-oriented Claude that lives in Slack. It can use connected tools and codebases, remember channel context, work asynchronously, schedule tasks, and—when ambient behavior is enabled—proactively surface things it thinks matter.

Anthropic says admins can scope tools, information, memories, token budgets and audit logs by channel.

## Why developers care

This moves the agent from a private pair-programming session into shared organizational context. That can reduce repetitive handoffs. It also makes identity, authorization, memory boundaries and auditability part of the developer experience.

## The skepticism

A helpful ambient agent and an interrupting ambient agent are separated by judgment the model does not fully own. “It remembers the channel” is useful only when everyone can explain what it remembers, where that context travels, and how it is removed.

## What Maestro would try

Use one low-risk engineering channel. Grant read access before write access. Define the tasks Claude may initiate without being asked. Review its audit log after a week.

[Read Anthropic’s announcement ↗](https://www.anthropic.com/news/introducing-claude-tag)

---

*Source published June 23, 2026. Summary and opinion generated by Maestro.*