Skip to content
NewsProduct launch

Cursor's Slack integration now previews its plan before starting and spans multiple repos

· by Pondero Newsdesk

The short version

Three Slack improvements shipped July 17: plan previews before agents start, named multi-repo environments so teams target their full stack from one mention, and cross-channel context gathering during tasks.

Cursor's Slack integration now previews its plan before starting and spans multiple repos

Slack-triggered AI agents have defaulted to one behavior: start immediately. Cursor's July 17 changelog changed that for its own integration. Per Anysphere's release notes, Cursor now shares a plan before writing any code, waits for redirection, then posts live status updates as it works. Three additional workflow changes shipped alongside it.

What changed

Plan preview and live status. When a team member mentions @Cursor in Slack with a task, the agent responds with a proposed plan before executing. Teams can redirect at that point rather than after it has burned compute on the wrong approach. Once Cursor begins, it posts a status update at each step so engineers can follow the work without interrupting the thread.

Named multi-repo environments. Cursor can now start a Slack task in a named environment spanning multiple repositories, not just a single default repo. A team with frontend, backend, and shared library code in separate repos defines one environment covering all three; Cursor reads the message content and routes automatically to the right environment. Mid-task, if Cursor needs a repository outside the current environment, it surfaces a Switch Repository button. Selecting a target lets the agent resume rather than restart.

Cross-channel workflows. The integration can now read from and post to other Slack channels and threads during a task, not only the channel where it was first mentioned. An agent can pull context from a design channel or product thread and post updates back to the original thread or any relevant channel without manual copying between windows.

The update also reworked message layout. In-message action buttons were removed and replaced by compact footer links. Tables, pull request references, and build artifacts render more cleanly in thread view, per the changelog.

Why it matters

The plan-preview step addresses the most expensive failure mode in Slack-driven agent workflows: compute spent on misunderstood tasks. On complex, multi-step tasks, a short review of the agent's intended approach before it runs prevents unnecessary token spend and avoids the friction of stopping and restarting mid-execution.

Named environments close a structural gap for teams with distributed codebases. Before this update, Slack-triggered Cursor agents defaulted to a single repository. Cross-repo work required repeated manual configuration or additional prompts to redirect the agent. One @Cursor mention can now reach the full relevant codebase from a single Slack message.

The Cloud Agents capability powering the Slack integration requires a paid plan. Teams evaluating the feature can review current tiers at cursor.com/pricing or start an account at cursor.com.

What to watch next

Cursor has not specified which paid tiers include multi-repo environments or cross-channel posting, two of the four capabilities from the July 17 update. Routing-rule configuration (keyword-to-environment mappings for automatic repo targeting) is described in Cursor's Slack docs but Cursor has not announced when that configuration surface will appear in the Slack settings interface. The next changelog post will likely clarify scope.

Sources