A
aloki-alok/kage
Run your own Claude Code session from Slack or WhatsApp
View on GitHub ↗Link checked 2026-11-14 · HTTP 200 · opens the vendor's own site
Official descriptionRun your own Claude Code session from Slack or WhatsApp. Tag @kage in a thread and your local agent does the work and replies there.
- Star count
- 1 ★ · aloki-alok/kage · read from the GitHub search API · (checked 2026-11-13 20:00:50)
- Source
- read directly from its GitHub page · type: open-source project
- Harvest history
- first seen 2026-11-14 · last seen 2026-11-14 · seen 1× · listed by 1 sources
- Score
- score 103 · 1 ★ · 1 sources · repo
FAQ (answers come from our records)
Who is it for?
We filed it under these tags: AI Agent. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
aloki-alok/kage;first catalogued 2026-11-14;last seen 2026-11-14。There is no official certification or quality rating here — we record sources and dates only.Kage lets you tag @kage in any Slack thread, DM or channel, and your own Claude
以下摘录自该仓库的 README(片段) · 在 GitHub 读完整 README →
#incidents ▸ thread
Priya checkout is throwing 500s since the 14:05 deploy
You @kage what changed in that deploy?
Kage 🥷 Looking at the deploy diff and the error logs...
Kage The 14:05 deploy renamed order.total to order.amount.
3 callers still read the old name. Fix is on branch fix/order-amount.
## Set it up with your agent
Paste this into Claude Code (or any coding agent with a shell):
text
Set up Kage for me: https://github.com/aloki-alok/kage
Install it with: curl -fsSL https://raw.githubusercontent.com/aloki-alok/kage/main/install.sh | bash
Then follow AGENTS.md in ~/.local/share/kage from start to finish.
Do every step yourself, and stop only for the steps it marks for me.
Finish when kage doctor ends with "All good".
Your agent does the work. You only do what a person has to do:
In Slack, click Create on the prefilled app, generate one token, and click Allow.
Paste the three tokens into the hidden prompts in the terminal window your agent opens. The tokens never pass through the agent.
The rest of this page is for doing it by hand, and for knowing how it works.
## Before you start
| You need | Why |
| macOS or Linux, Node 22.18 or newer | Kage runs TypeScript directly on Node. macOS (launchd) or Linux (systemd) keeps it running. tmux only for tmux panes. |
| A recent Claude Code, logged in with a claude.ai account | Each thread is a normal Claude Code session with your settings, skills, tools and logins. |
| Permission to create an app in your Slack workspace | If your workspace restricts apps, an admin approves it once. |
## Set it up by hand (about 5 minutes)
1. Install it. One line, no git needed: it downloads Kage, puts the kage command on your PATH, and starts the setup.
sh
curl -fsSL https://raw.githubusercontent.com/aloki-alok/kage/main/install.sh | bash
Run the same line again later to update. Your settings stay.
2. Pick a name when asked. kage is the default. Whatever you pick becomes the @name you tag.
3. A browser tab opens with your Slack app already filled in. Pick your workspace, then click Next and Create.
4. Under Basic Information > App-Level Tokens, click Generate, add the scope connections:write, and copy the token (xapp-...).
5. Under Install App, click Install to Workspace, allow it, and copy both tokens: User OAuth (xoxp-...) and Bot (xoxb-...).
6. Paste the three tokens into the setup script. It checks each one with Slack and saves them to ~/.config/kage/env, readable only by you.
## Start it
sh
kage start
The first start installs the Agent SDK and turns on the service: macOS or Linux keeps Kage running, brings it back within about 10 s after a crash, and starts it at login (on Linux, at boot once linger is on; kage doctor prints the command). Three starts in 10 minutes send you a DM. Each Slack thread gets its own Claude Code session inside it.
kage alone says whether Kage is running and shows the six commands you use most. kage help lists all of them, kage help shows one with an example, and a mistyped command gets the closest real one.
| To | Run |
| Check every part | kage doctor prints the command that fixes each problem |
| See every session live | kage top, by mouse or keys. kage status prints one look. |
Priya checkout is throwing 500s since the 14:05 deploy
You @kage what changed in that deploy?
Kage 🥷 Looking at the deploy diff and the error logs...
Kage The 14:05 deploy renamed order.total to order.amount.
3 callers still read the old name. Fix is on branch fix/order-amount.
## Set it up with your agent
Paste this into Claude Code (or any coding agent with a shell):
text
Set up Kage for me: https://github.com/aloki-alok/kage
Install it with: curl -fsSL https://raw.githubusercontent.com/aloki-alok/kage/main/install.sh | bash
Then follow AGENTS.md in ~/.local/share/kage from start to finish.
Do every step yourself, and stop only for the steps it marks for me.
Finish when kage doctor ends with "All good".
Your agent does the work. You only do what a person has to do:
In Slack, click Create on the prefilled app, generate one token, and click Allow.
Paste the three tokens into the hidden prompts in the terminal window your agent opens. The tokens never pass through the agent.
The rest of this page is for doing it by hand, and for knowing how it works.
## Before you start
| You need | Why |
| macOS or Linux, Node 22.18 or newer | Kage runs TypeScript directly on Node. macOS (launchd) or Linux (systemd) keeps it running. tmux only for tmux panes. |
| A recent Claude Code, logged in with a claude.ai account | Each thread is a normal Claude Code session with your settings, skills, tools and logins. |
| Permission to create an app in your Slack workspace | If your workspace restricts apps, an admin approves it once. |
## Set it up by hand (about 5 minutes)
1. Install it. One line, no git needed: it downloads Kage, puts the kage command on your PATH, and starts the setup.
sh
curl -fsSL https://raw.githubusercontent.com/aloki-alok/kage/main/install.sh | bash
Run the same line again later to update. Your settings stay.
2. Pick a name when asked. kage is the default. Whatever you pick becomes the @name you tag.
3. A browser tab opens with your Slack app already filled in. Pick your workspace, then click Next and Create.
4. Under Basic Information > App-Level Tokens, click Generate, add the scope connections:write, and copy the token (xapp-...).
5. Under Install App, click Install to Workspace, allow it, and copy both tokens: User OAuth (xoxp-...) and Bot (xoxb-...).
6. Paste the three tokens into the setup script. It checks each one with Slack and saves them to ~/.config/kage/env, readable only by you.
## Start it
sh
kage start
The first start installs the Agent SDK and turns on the service: macOS or Linux keeps Kage running, brings it back within about 10 s after a crash, and starts it at login (on Linux, at boot once linger is on; kage doctor prints the command). Three starts in 10 minutes send you a DM. Each Slack thread gets its own Claude Code session inside it.
kage alone says whether Kage is running and shows the six commands you use most. kage help lists all of them, kage help shows one with an example, and a mistyped command gets the closest real one.
| To | Run |
| Check every part | kage doctor prints the command that fixes each problem |
| See every session live | kage top, by mouse or keys. kage status prints one look. |
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/aloki-alok-kage/)
- Source
- github · https://github.com/aloki-alok/kage
- Browse by field
- AI agents
- Traction
- 1 ★
- added
- today
More like this
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.