Agents

Slack launches Slack Code for team-based AI development

Slack has introduced Slack Code, a new feature that lets developers and non-technical employees collaborate with AI coding agents directly inside dedicated project channels.

Computerworld AI4 days agoAgents
Image: Computerworld AI

Slack has launched Slack Code, a new feature designed to turn software development into a multiplayer experience by embedding AI coding agents directly into shared, project-based channels. Available at no additional cost across all Slack subscription tiers, the tool integrates with four major AI coding agents at launch: Claude Code, Devin, GitHub Copilot, and Vercel. When a user tags one of these agents in a channel or direct message, Slack automatically generates a dedicated code channel, links the necessary files, and invites the relevant team members to participate.

Inside these specialized channels, both technical and non-technical team members can interact with the AI agent using natural language. The workspace displays session artifacts to help teams track progress, including code diffs that highlight modifications, Slack canvas documents, and live HTML previews of prototypes. Once a task, such as fixing a bug or updating a web page, is completed, the channel automatically archives itself. Slack has also introduced agent DMs for one-on-one interactions and a dedicated agents tab to organize these conversations.

For development teams, this shift expands who can participate in the software lifecycle. Marketers and designers can directly suggest edits or approve prototypes within a secure environment. To prevent non-technical staff from accidentally deploying broken code, administrators can establish guardrails requiring engineering sign-off before any code is shipped. The system operates under Slack's existing security and permissions framework, meaning agents only access data based on established workspace controls. However, IT administrators will need to carefully manage permissions, which analysts note remains a persistent challenge in multiplayer environments.

This is our own summary of reporting by Computerworld AI

More in Agents