Slack-native
One agent per project. Each agent listens to its own Slack channel. Switch between projects by switching channels.
You → Slack #project-a → y0mcp → Claude Code in ~/projects/project-a ↕ permission relay (✅/❌ in Slack) → Dozzle API (optional)| OpenClaw | Claude Code Channels | y0mcp | |
|---|---|---|---|
| Slack | ✓ | ✗ | ✓ |
| Multi-project routing | ✓ | ✗ | ✓ |
| Permission relay in Slack | ✗ | ✗ | ✓ |
| claude.ai subscription | ✗ | ✓ | ✓ |
| API key required | ✓ | ✗ | ✗ |
| Auto token refresh | ✗ | ✗ | ✓ |
| Always-on (launchd/systemd) | ✗ | ✗ | ✓ |
git clone https://github.com/es-ua/y0mcpcd y0mcp && bash scripts/install.shbash scripts/new-agent.shSlack-native
One agent per project. Each agent listens to its own Slack channel. Switch between projects by switching channels.
Permission relay
Claude asks before risky actions. Approve or deny with a reaction directly in Slack — no terminal needed.
Always-on
Agents run as launchd/systemd services. Auto-restart on crash. OAuth tokens refresh automatically every 2 hours.
Container logs
Optional Dozzle integration for fetching container logs through Claude Code.