Sttts Slagent
- 21 repo stars
- Updated March 27, 2026
- sttts/slagent
sttts-slagent is a Claude Code marketplace distributing slagent, a Go library and slaude CLI that mirror Claude Code agent sessions to Slack threads so developers can watch, comment, and steer from Slack.
About
sttts-slagent packages slagent, a Go library for streaming agent sessions to Slack, and slaude, its CLI for mirroring Claude Code into Slack channels or threads. slaude start launches a new mirrored session; slaude join attaches another agent instance to an existing thread; slaude resume continues from a cursor URL. Access modes include locked owner-only, open, selective, and observe variants toggled via /open, /lock, or /observe slash commands in Slack. Permission requests can route to Slack for manual approval, with optional --dangerous-auto-approve green/yellow levels and known-host network policies classified by claude --model haiku. The standalone claude-command-classifier-hook shares classifier.yaml with slaude. slagent also exposes a Go API for custom Slack-integrated agent UIs. The project is experimental and marked likely insecure. Install slaude via Homebrew tap sttts/slagent or go install github.com/sttts/slagent/cmd/slaude@latest. macOS session-token extraction is actively tested. Use sttts-slagent when Claude Code work must be visible and steerable from Slack instead of the terminal alone.
- slaude CLI mirrors Claude Code sessions to Slack channels or threads
- Supports start, join, resume, auth, channels, and share plan-file commands
- Thread access modes: locked, open, selective, and observe via Slack /commands
- Routes tool permissions to Slack with haiku-based risk classification
- slagent Go library builds custom Slack-streaming agent UIs on same core
Sttts Slagent by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
/plugin marketplace add sttts/slagentAdd your badge
Show developers this marketplace is listed on Skillselion. Paste this into your README.
| repo stars | ★ 21 |
|---|---|
| Last updated | March 27, 2026 |
| Repository | sttts/slagent ↗ |
How do you mirror Claude Code sessions to Slack?
Run slaude start #channel "task" to mirror a Claude Code session into a Slack thread with steerable access modes and Slack-side permission approvals.
Who is it for?
Teams already on Slack who want experimental shared visibility into Claude Code sessions with thread-level steering and permission routing.
Skip if: Security-sensitive repos, air-gapped environments, or workflows requiring guaranteed-safe production-grade Slack integrations without experimental caveats.
What you get
Slack thread transcript of Claude turns, steer commands, permission approval prompts, and resumable session URLs.
- Slack thread mirror of agent session
- Resumable session URL
- Slack permission approval prompts
By the numbers
- slaude CLI provides start, join, resume, auth, channels, and share commands
- Permission classifier uses claude --model haiku for speed and cost
- macOS is the actively tested platform for Slack session-token extraction
Plugins in this marketplace
1 plugin - install individually after you add the marketplace.
Recommended Marketplaces
How it compares
Pick sttts-slagent over screen sharing when Slack-native steering, threaded resumes, and permission approvals must wrap Claude Code sessions.
FAQ
What is the difference between slagent and slaude?
Sttts Slagent distributes slagent, the Go library that streams agent turns to Slack threads, and slaude, the CLI built on slagent that mirrors Claude Code sessions with start, join, resume, and auth commands.
How do Slack thread access modes work in slaude?
Sttts Slagent slaude sessions default to locked owner-only mode. Teams can switch threads to open, selective user lists, or observe mode using Slack /open, /lock, /observe, and :fox_face: shortcut commands without restarting Claude.
Is Sttts Slagent production ready?
Sttts Slagent README marks slaude as experimental and likely insecure, warning that mirroring Claude Code to Slack exposes session content. Use at your own risk and prefer locked mode for sensitive repositories.