Slackgentic Team — Communication skill for Claude Code

slackgentic-team

Local Codex and Claude agents, coordinated from Slack.

Slackgentic turns one Slack channel into a control room for local coding agents: start work with a normal message, watch agent output stream back into a thread, track sessions started outside Slack, and keep all credentials and transcripts strictly local.

Use agents exactly the way you already do. They still run on your laptop, with the same tools, deployment access, repo access, and approval levels. Slack just becomes the place where you can step away, touch some grass, and reply when an agent needs you.

Why It Exists

• One Slack channel becomes the surface for local agent work.
• Everything stays on your machine.
• The agent channel is private by default, so only you see what your agents are up to.
• Codex and Claude work under one roster.
• No API-key billing detour: it drives the Codex and Claude Code you already use.
• Cross-model reviews are easy: ask for one directly, or let an agent pull in another agent when it wants a second opinion.
• Thread subtasks are natural: reply with somebody ..., another agent picks up that slice, then the original agent resumes with the added context.
• Sessions started outside Slack still get tracked in Slack threads.
• Hire/fire keeps parallelism intentional.

Quick Install

Recommended — installs `slackgentic` on `PATH` globally via [uv](https://docs.astral.sh/uv/) so the command works in any shell:

uv tool install --with pip git+https://github.com/unbounded-ai/slackgentic-team.git
slackgentic slack setup
slackgentic service install && slackgentic service status

`--with pip` lets the in-app updater apply new releases in place.

Alternatively, install from a source checkout — needed if you want to track `main` or hack on the code:

git clone https://github.com/unbounded-ai/slackgentic-team.git && cd slackgentic-team
python3.13 -m venv .venv && source .venv/bin/activate && pip install -e . && slackgentic slack setup
sla