MCP Server

Active

MIT

Kagan MCP Server

io.github.kagan-sh/kagan

Supervised kanban board for AI coding agents with isolated git worktrees and human review gates.

What is the Kagan MCP server?

Kagan is an OpenCode plugin that organizes AI agent coding work into supervised kanban tasks. Each task runs in an isolated git worktree, progressing through Backlog → In Progress → Review → Done with mandatory human review gates at every transition. The agent never touches your main checkout, and merges only happen with your explicit approval.

Kagan turns autonomous AI coding agent work into a managed workflow with human oversight. It creates isolated git worktrees for each task, routes agent findings through a reviewer agent, and requires you to triage and approve all changes before merge. Use it to safely delegate coding tasks to agents while maintaining full control and visibility.

How to install Kagan

Copy-paste configuration for popular MCP clients.

transport: stdio

Config generated by PluginBench — verify against the source before use.

Environment / auth

KAGAN_OPENAI_API_KEY
secret
OpenAI API key for agent execution (if using OpenAI backend)
KAGAN_ANTHROPIC_API_KEY
secret
Anthropic API key for agent execution (if using Claude backend)
KAGAN_DB_PATH
Custom path for Kagan database file

~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "kagan": {
      "command": "uvx",
      "args": [
        "kagan"
      ],
      "env": {
        "KAGAN_OPENAI_API_KEY": "<YOUR_KAGAN_OPENAI_API_KEY>",
        "KAGAN_ANTHROPIC_API_KEY": "<YOUR_KAGAN_ANTHROPIC_API_KEY>",
        "KAGAN_DB_PATH": "<YOUR_KAGAN_DB_PATH>"
      }
    }
  }
}

Tools & capabilities

Tools this server exposes to the agent.

Kanban board interface — TUI-based kanban board for managing AI agent tasks with Backlog, In Progress, Review, and Done columns
Task creation — Create and manage tasks conversationally via /kagan-task command
Isolated worktree management — Automatically creates and manages isolated git worktrees for each task on kagan/<slug> branches
Review agent — Automated reviewer agent that files ranked findings against completed tasks
Settings management — Configure Kagan behavior via /kagan-settings command

Use cases

Delegate coding tasks to AI agents while maintaining human review and approval control
Organize multiple concurrent agent-driven development tasks on a shared kanban board
Review and triage AI-generated code changes before merging to main branches
Track agent work progress from intake through completion with mandatory gates at each stage
Manage isolated feature branches for agent-generated code without touching your main checkout

Kagan MCP server FAQ

What is Kagan?

Kagan is an OpenCode plugin that creates a supervised kanban workflow for AI coding agents. Each task runs in an isolated git worktree and requires human review and approval before merging.

Is Kagan free?

Yes, Kagan is open-source under the MIT license.

How do I install Kagan?

Install globally with `opencode plugin -g @kagan-sh/kagan` (requires OpenCode 1.17.13+), or add a local clone to your OpenCode config. Open the board with `/kagan` from the command palette or press `<leader>k`.

Do I need authentication to use Kagan?

No special authentication is required. Kagan works with your existing OpenCode and git setup.

Can the AI agent modify my main code?

No. Kagan isolates all agent work to separate git worktrees on `kagan/<slug>` branches. Your main checkout is never touched until you explicitly approve and merge.

What happens when an agent completes a task?

A reviewer agent files ranked findings against the task. You then triage every finding and decide whether and where to merge the changes.

README (reference)

Source of truth, from the repository.

<h3 align="center"> <img src="https://raw.githubusercontent.com/kagan-sh/kagan/main/docs/public/favicon.svg" width="80" alt="Kagan" /><br /> kagan </h3> <p align="center">Supervised kanban for AI coding agents inside OpenCode.</p> <p align="center"> <a href="https://docs.kagan.sh/">Docs</a> · <a href="https://docs.kagan.sh/quickstart/">Quickstart</a> </p>

Kagan is an OpenCode plugin that turns agent work into supervised tasks on a kanban board. Each task is an OpenCode session running in an isolated git worktree, moving through Backlog → In Progress → Review → Done with a gate at every transition — intake before the agent runs, review before you approve, merge only on your say-so.

The agent never touches your checkout. It works on a kagan/<slug> branch in its own worktree, a reviewer agent files ranked findings against the original task, and nothing reaches Done until you've triaged every finding and chosen where — or whether — to merge.

Install

You need OpenCode 1.17.13 or newer (see engines.opencode in package.json).

Install globally so Kagan is available in every project:

opencode plugin -g @kagan-sh/kagan

Or add a local clone to both OpenCode config files:

{
  "plugin": ["/path/to/kagan"]
}

Open the board with /kagan from the command palette, the kagan palette command, or <leader>k (the leader key defaults to ctrl+x).

From any regular OpenCode session, run /kagan-task to create board tasks conversationally — useful when planning several tickets at once without opening the board create dialog for each one.

For global npm installs, Kagan checks npm for newer stable releases and shows an available version in the board footer. Press u, run /kagan-update, or use the command palette to review and confirm the update; Kagan then installs the exact release and asks you to restart OpenCode. Local, file, and development installs are never updated automatically. See Updating.

Pass options by using the array-of-array form, or open /kagan-settings from the project — see the configuration reference.

Docs

Documentation is available in docs/.

License

MIT

Related MCP servers

Kagura Memory Cloud

Active

Persistent memory for AI assistants: store, search, and connect knowledge across conversations with adaptive neural learning.

Python

Apache-2.0

View repository →

RTL awareness for coding agents

Give coding agents right-to-left awareness: lint, normalise Arabic, detect direction.

io.github.kahlilashanti/mantis-mcp

Active

Build proven high converting product showrooms and websites from your existing ecommerce catalog

TypeScript

MIT

View repository →

io.github.kailayerhq/kai

Semantic code intelligence — call graphs, dependencies, impact analysis, and test coverage.

View repository →

io.github.kairosastro-sketch/mcp-ephemeris

Active

NASA-validated computed planetary positions for your AI: 19 astrology tools, never hallucinated.

TypeScript

View repository →

TeTsuki Gateway

Checks whether a website has a usable llms.txt. Score 0–100. Free 3/day/IP, then ¥10.

View repository →