Tools
AI Minute Newsroom
2026-08-20
Slack has given coding agents their own channels — the work is now something the whole team watches happen
Slack launched Slack Code on 20 August. Mention a coding agent in a project channel and Slack opens a dedicated code channel for that piece of work: the agent posts its diffs there, renders live previews of what it built, and everyone in the channel reads the same conversation the agent is working from, leaves feedback it picks up, and approves or rejects before anything ships. High-stakes changes route to a named person for sign-off inside the channel. The channel archives itself when the task is finished, and Slack says more than 70 percent close within a single day. The launch partners are Anthropic's Claude Code, Cognition's Devin, GitHub Copilot and Vercel's agent, with OpenAI's ChatGPT listed as coming soon.
Why it mattersUntil now an agent's work has mostly been private: one developer, one terminal, one transcript nobody else reads. What the rest of the team sees arrives at the end, as a pull request, with the reasoning already gone. Putting the session in a channel changes who gets to interrupt — a designer can object to a layout while it is being built rather than after — and it turns approval into a message from a person rather than a checkbox on a form. It also changes what an audit looks like, since the argument, the diff and the sign-off end up in the same searchable place. The risk is the familiar one for anything that lives in chat: a channel that scrolls is easy to stop reading, and an approval that arrives twenty times a day stops being a decision.
✓ Verified · 2 sources
▶ Related video: Live Demo: Bring All Your AI Agents into Slack
Read in the app — free, in 9 languages
Related stories
Meta's assistant keeps an hourly file on everyone in your life.
2026-10-05Two senators want prison time for bosses whose AI agents hack.
2026-10-05One command restores the Apple Intelligence off switch Apple deleted.
2026-10-05OpenAI will ship a Codex upgrade daily for 28 days or reset limits.
2026-10-05Untuned models solved agent tasks their polished versions could not.
2026-10-04