Kanban

README

# OpenClaw Kanban

Task board and agent status UI for OpenClaw. Runs as a standalone HTTP service and integrates with OpenClaw via the task-dashboard skill and REST API.

## Features

- Kanban board: Todo / In progress / Done / Failed columns with drag-and-drop
- P.A.R.A. areas and task assignment
- Agent sidebar: list agents from your OpenClaw agents directory with working/idle status
- Task logs and LLM activity per task
- REST API for tasks, areas, agents (for cron and agent tools)

## Requirements

- Python 3.8+
- OpenClaw installed and configured (for full integration)

No extra pip dependencies; the server uses the Python standard library only.

## Quick start

1. Clone or copy this repo to a directory of your choice.

2. Set environment variables (optional):
   - `KANBAN_PORT` — HTTP port (default: 9999)
   - `OPENCLAW_KANBAN_DIR` — Directory containing server.py and assets (default: script directory)
   - `OPENCLAW_AGENTS` — OpenClaw agents root (default: ~/.openclaw/agents)
   - `OPENCLAW_WORKSPACE` — OpenClaw workspace root (default: ~/.openclaw/workspace)

3. Run: `python3 server.py` then open http://127.0.0.1:9999/

## Integration with OpenClaw

- Use a reverse proxy (e.g. Nginx): / for Gateway, /kanban for this server.
- Enable the task-dashboard skill in OpenClaw and point task API to this server.
- Cron jobs can create/update tasks via the Kanban API.

See OpenClaw docs for Gateway, skills, and cron setup.

## Project layout

- server.py — HTTP server and API
- db.py — SQLite and PARA access
- dashboard.html — Single-page Kanban UI
- requirements.txt — Dependencies (currently none)
- tests/ — Unit tests: `python -m unittest discover -s tests`

## License

MIT. See LICENSE.