Skip to content

Run coding agents from Slack

Mention the bot in a channel to start a session. It replies in a thread, and every plain reply in that thread continues the same session. No re-mention is needed. The first word after the mention can name a skillset to start with.

While the agent works, a progress message updates on a 10-second cadence. Files you attach are fetched by the agent inside the session (size caps in limits).

Slack reserves /, so Nori commands use !. In a session thread:

Command What it does
!help list commands
!status session status card
!shell / !ide / !chat / !urls links into the running session
!interrupt <prompt> interrupt the current turn with a new instruction
!thoughts show the agent’s current reasoning
!restart restart the agent on the same machine
!skillsets [pick|list|info|switch <name>] switch skillset mid-thread
!model [name|pick] and !provider [id|pick] switch model or agent provider
!resume pick from your resumable sessions, started on any surface
!catchup replay the thread’s history
!admin request admin escalation (single-use link, expires in 5 minutes)
!done / !release end the session and release the machine

Fleet-wide, outside a session thread: !fleet-status, !free-oldest.

Switching agents with !provider moves the conversation to a fresh machine: history carries over as a transcript bridge, working files do not.

Give the agent a bounded task, the repository or system in scope, and the verification required before it reports completion:

@nori in the api repository, reproduce the failing login test, fix the cause,
run the focused suite, and prepare a pull request. Do not merge it.

The thread becomes both the control surface and the result log. Reply with new constraints, use !interrupt when the current turn should change immediately, and use !status when you only need lifecycle information.

!done is the only action that makes a thread human-only. After it, plain replies are ignored until a new mention. Every other way a session winds down (idle timeout, restart, transport drop) keeps reply-to-resume working: your next thread reply revives the session, with a catch-up digest if it advanced elsewhere.

Three modes govern who can drive the bot (set in Fleet Setup):

Mode Behavior
open anyone in the workspace can start and drive sessions
guarded org members trusted; other users’ thread messages are queued as context, not executed
strict only allowlisted emails and admins; others are blocked

A separate knob controls the agent’s own reach into Slack, from channel-only (default: proxied, scoped to the thread’s channel) up to direct-token (raw token in the session; internal environments only).

The intermediate reach modes are public-member, for public channels the bot can access, and all-channels, for public or private channels the bot has joined. Neither mode grants access to direct messages.

The Integrations checklist installs the app with least-privilege scopes. DM support, file upload, and file download are optional scope groups you can toggle off before installing. Both socket mode and HTTP mode are supported.

  • If a mention gets no response, confirm the app is in the channel and the user is allowed by the organization’s Slack access mode.
  • If a thread reply is ignored, check whether someone previously used !done or !release. Start a new session with another mention.
  • If the agent cannot read an attachment, verify that file download was enabled when the app was installed and that the file is within the documented limit.
  • If a switch loses uncommitted files, return to the prior agent or recover from source control. Agent switches preserve transcript context, not the working directory.

this page as markdown: /sessions/guides/slack.md· index for agents: /llms.txt