WildCrane plugin
Plugins and CLI for a cloud task board: an idea becomes a card with a preview, and your agent delivers it once you approve.
- 51 ★ on the repository
- Last commit
Extended access
Why this level
- The Runner executes agent tasks in your repository, and `akb` commands read and change tasks in the cloud.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/wildcraneDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The Runner executes agent tasks in your repository, and `akb` commands read and change tasks in the cloud.
This is third-party code. Review the repository files before installing.
What it does
WildCrane turns a wish into a task with a preview you review before anything is built. After approval your coding agent delivers it. The repository holds the parts that run on your computer: the `akb` CLI (npm package `wildcrane`), a Runner that executes tasks in your repository and plugins for Claude Code, ZCode, Codex, Cursor, Kimi Code and Grok Build. Through `akb` an agent reads and changes cloud tasks straight from a conversation: `akb task list`, `create`, `approve`, `chat`, `resume`, `stop`. Results show up in the web app at once. The cloud and web app are closed source and not in the repository, and per the authors code, branches and model keys are not stored in the cloud.
Who it is for. Indie hackers and small teams that track tasks on a board and want to preview the outcome before an agent starts working.
Good fit when
- Before coding you want to approve a mock-up of a screen or subtitles instead of reviewing a finished diff
- Tasks from the web board should be picked up by an agent on your laptop rather than a cloud machine
- A task has to be created, approved or stopped from the conversation without opening a browser
- One project is run from Claude Code while some tasks launch from Codex or Cursor
- You need studios and preview components for your own result types
Not a fit when
- You need a standalone tool without an account: the board and web app live in the closed WildCrane cloud
- The repository must not be linked to an external task service
- You want a board without previews and approvals, a plain task list is enough
Example request
List project tasks with status waiting_review and approve the one whose subtitle preview looks right.Limitations
The repository contains only the local parts. WildCrane Cloud and the web app are closed, so you need an account there and access to its public API. Some documentation and command output is in Chinese. The old `ai4kanban` package must be removed before installing `wildcrane`, since both install `akb`. `akb install` and the `akb skill` tree were removed: the Runner writes board skills when a folder is bound. The Grok Build plugin installs with `--trust`, so verify the source first. The license is Apache-2.0.
How to disable. Disable the `wildcrane` plugin through `/plugin` and remove the package with `npm uninstall -g wildcrane`. Delete the project binding (`.akb/`) and the Runner on the machine separately.
FAQ
How do I install WildCrane in Claude Code?
Run `claude plugin marketplace add wildcrane-ai/wildcrane`, then `claude plugin install wildcrane@wildcrane`.
Which agents are supported?
The repository holds plugins for Claude Code, ZCode, Codex, Cursor, dsh, Kimi Code and Grok Build. Kimi Code uses the `kimi-plugin` branch and Grok Build the `grok-plugin` branch.
Where is my code stored?
Per the authors, on your computer: the Runner works in your repository, and the cloud stores no code, branches or model keys.
What does `akb task approve` do?
It approves a task awaiting review and hands it to the agent for delivery. A free comment is added with the `--comment` flag.
Do I need an account?
Yes, the board and web app run in WildCrane Cloud, and the repository holds client parts that talk to the cloud only through its public API.
Discussion
No questions or reviews yet.
Related
PluginsA self-hosted issue tracker and wiki on Cloudflare with MCP: 126 tools, issue leases and file claims for several agents
A local task board with the taskctl CLI and a skill for Codex, Claude Code, and other agents: creates issues, moves statuses, and keeps comments under a fixed w
A plugin and MCP for a self-hosted task board where an agent claims a card, does the work and hands off with token telemetry.
CodeKanban gathers terminals and Claude Code and Codex sessions on one page, and codekanban-cli drives them from an agent
A skill sets up one shared task board for people and agents: kanban, a dependency graph and a code map on one page.
A scaffold for Claude Code and Codex: tasks in files, developer and QA roles, manual approval and shared state.
README in short
The README briefly describes the product, lists repository folders (`cli`, `runner`, `plugins`, `artifact-sdk`, `official`) and gives installation for five hosts. Details live on the documentation site.