SkillAgentSearch skills...

kanzo

Self-hosted kanban for humans and agents. One 4 MB binary, starts in 2 ms, zero setup.

Install / Use

claude mcp add emanuele-em -- npx -y github:emanuele-em/kanzo

If the server publishes to npm under a different name, use that package instead — check the repo README.

About this skill
🔌

MCP Server

Model Context Protocol server

Quality Score

78/100

Supported Platforms

Claude Code
Claude Desktop
OpenAI Codex

Our assessment of kanzo

kanzo scores 78/100 on our quality scale, 533rd of 599 Data & Analytics skills we index.

Its MCP Server is 5.0 KB long, split into 7 sections with 4 code examples: a solid amount of guidance for an agent.

It has 46 GitHub stars, so there is little community track record yet; judge it on its content.

Substance
26/30
Structure
18/20
Description
12/15
Adoption
7/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated today, so kanzo is actively maintained.
  • It is released under AGPL-3.0, a copyleft license: you can use it, but modified versions you distribute must carry the same license.
  • Its trust signals score 92/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands (1 minor note below). An AI review of the same text found nothing harmful.

  • noteInstalls by piping a downloaded script into a shellline 24
    curl -LsSf https://github.com/emanuele-em/kanzo/releases/latest/download/kanzo-installer.sh | sh

AI review by kimi-k2.7-code on 2026-10-27. Automated pattern scan on 2026-10-27. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

kanzo compared with similar skills

All 4 of these similar skills score higher than kanzo; compare them before choosing.

SkillScoreStarsUpdatedFormat
kanzo (this skill)by emanuele-em7846todayMCP Server
headroomby headroomlabs-ai10078.0ktodayCLAUDE.md
Agent-Reachby Panniantong100104.1k2d agoMCP Server
Scraplingby D4Vinci10089.9k1d agoMCP Server
crawl4aiby unclecode10086.4ktodayMCP Server

Frequently asked questions

How do I install kanzo?
Run claude mcp add emanuele-em -- npx -y github:emanuele-em/kanzo. The install tabs above show the steps for each supported agent.
Which AI agents does kanzo work with?
It is written for Claude Code, Claude Desktop and OpenAI Codex, as a MCP Server file. Other agents that read the same format can often use it too.
Is kanzo safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands (1 minor note below). An AI review of the same text found nothing harmful. It is AGPL-3.0-licensed and scores 92/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is kanzo still maintained?
The repository was last updated today, so kanzo is actively maintained.
<p align="center"> <picture> <source media="(prefers-color-scheme: dark)" srcset="docs/img/kanzo-dark.jpg"> <img alt="kanzo: a tiny kanban board for you and your coding agents" src="docs/img/kanzo-light.jpg"> </picture> </p> <p align="center"> <a href="https://github.com/emanuele-em/kanzo/actions/workflows/ci.yml"><img src="https://github.com/emanuele-em/kanzo/actions/workflows/ci.yml/badge.svg" alt="CI"></a> <a href="https://crates.io/crates/kanzo"><img src="https://img.shields.io/crates/v/kanzo" alt="crates.io"></a> <a href="LICENSE"><img src="https://img.shields.io/badge/license-AGPL--3.0-blue" alt="AGPL-3.0"></a> </p>

Kanzo is a kanban board in your browser and an MCP server in one small program, so you and your coding agents work on the same tasks. The whole board is one folder, and a private git repo keeps it in step on every machine you use: no account, no server.

<picture> <source media="(prefers-color-scheme: dark)" srcset="docs/img/board-dark.gif"> <img alt="The Kanzo board: agents move cards while you watch, and ⌘K finds a task in any project" src="docs/img/board-light.gif"> </picture>

Install

curl -LsSf https://github.com/emanuele-em/kanzo/releases/latest/download/kanzo-installer.sh | sh

For macOS and Linux, including every Raspberry Pi. Or with Cargo:

cargo install kanzo

Use

| Command | Does | | --- | --- | | kanzo | Starts the board at http://localhost:4242 and opens it; --port for another port, --no-open to not open it | | kanzo service install | Keeps the board running and starts it at login; uninstall removes it | | kanzo mcp | The MCP server for agents: they start it, not you | | kanzo sync | Commits the board and pulls or pushes it; --repo <url> sets the repo, --keep-local or --keep-remote when both sides changed | | kanzo export | Prints the board as Markdown; -f json for JSON, -p <project> for one project, -o <file> to save it | | kanzo update | Installs the latest release and restarts the board's service; the board offers new versions too | | kanzo init | Creates the data folder; the first run does this too | | kanzo --version | Prints the version |

Every command takes --home <dir> to use another data folder.

n adds a task, / or ⌘K searches every project, Tab moves to the description, ⌘↵ saves and closes, Esc cancels.

Paste or drop an image into a task or a comment to attach it, as in ChatGPT. Agents get its file path and can open it.

When a new version is out, the sidebar offers it: one click installs it and restarts the board.

Connect your agents

claude mcp add kanzo --scope user -- kanzo mcp   # Claude Code
codex mcp add kanzo -- kanzo mcp                 # Codex

Other MCP clients: run kanzo mcp (apps like Raycast need the full path, from which kanzo), or use the running board at http://localhost:4242/mcp.

Then ask: "take the next task from kanzo and do it." The agent moves the card to doing and comments as it works. It finds the right project from the git remote of the repo it works in; link more repos to a project from its ⋯ menu.

<picture> <source media="(prefers-color-scheme: dark)" srcset="docs/img/task-dark.jpg"> <img alt="A task open on the board, with Claude Code's comments and a reply" src="docs/img/task-light.jpg"> </picture>

Your board on every machine

A laptop, a desktop, a Raspberry Pi: they all share one board through a private git repo you own.

  1. Create an empty private repo on GitHub, or anywhere git can push.
  2. Press the sync button in the sidebar and paste its URL. Kanzo makes ~/kanzo a git repo and pushes the board.
  3. On each other machine, do the same with the same URL: Kanzo offers to use the board that's already there.

From then on, press sync, or run kanzo sync, when you switch machines. It commits the board and pulls or pushes it, and it's safe while the board and your agents are running. If both machines changed the board, Kanzo asks which one to keep. To move to another repo, click its name at the bottom of the sidebar.

Settings

Everything lives in ~/kanzo: the board in kanzo.db, settings in config.toml. Use another folder with --home or KANZO_HOME.

port = 4242
listen = "127.0.0.1"   # "0.0.0.0" opens the board to your network: no logins, so only on one you trust
statuses = ["todo", "doing", "review", "done"]   # new tasks go first, the last counts as done
mcp_http = true        # MCP at /mcp

Small and fast

| | | | --- | --- | | Program | 4.1 MB, SQLite included | | Memory | 4.5 MB idle, 8 MB with 3,000 tasks | | Start | 2 ms | | Board of 3,000 tasks | 17 ms to load in the browser | | Agents' tool list | about 900 tokens, once per session | | Agents' replies | 10 to 60 tokens of plain text |

Measured on an Apple M4.

License

AGPL-3.0: use it, change it and share it; if you offer a changed version over a network, share its source too. To help, see CONTRIBUTING.

<sub>mcp-name: io.github.emanuele-em/kanzo</sub>

Related Skills

View on GitHub
GitHub Stars46
CategoryData
Updated10h ago
Forks2

Languages

Rust

Trust signals

92/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 low1 info