vibe-kanban
Manage AI coding agents on a visual Kanban board. Run parallel agents through a To Do→In Progress→Review→Done flow with automatic git worktree isolation and GitHub PR creation.
What this skill does
## Platform Support Status (Current)
| Platform | Current Support | Requirements |
|---|---|---|
| Claude | Native MCP integration | Register in `mcpServers` |
| Codex | MCP script integration | `scripts/mcp-setup.sh --codex` or equivalent config |
| Gemini | MCP registration | `mcpServers`/bridge configuration |
| OpenCode | MCP/bridge integration | `omx`/`ohmg` or equivalent setup |
Whether this skill alone is sufficient:
- Claude/Gemini: **Yes**
- Codex: **Yes (requires script-based setup)**
- OpenCode: **Yes (via orchestration)**
# Vibe Kanban — AI Agent Kanban Board
> Manage multiple AI agents (Claude/Codex/Gemini) from a single Kanban board.
> Moving a card (task) to In Progress automatically creates a git worktree and starts the agent.
## When to use this skill
- When breaking an epic into independent tasks for parallel agent assignment
- When you want to visually track the status of ongoing AI work
- When you want to review agent results as diffs/logs in the UI and retry them
- When combining GitHub PR-based team collaboration with AI agent work
---
## Prerequisites
```bash
# Node.js 18+ required
node --version
# Complete agent authentication beforehand
claude --version # Set ANTHROPIC_API_KEY
codex --version # Set OPENAI_API_KEY (optional)
gemini --version # Set GOOGLE_API_KEY (optional)
opencode --version # No separate setup needed (GUI-based)
```
> **Verified versions (as of 2026-02-22)**
> - vibe-kanban: v0.1.17
> - claude (Claude Code): 2.1.50
> - codex: 0.104.0
> - gemini: 0.29.5
> - opencode: 1.2.10
---
## Installation & Running
### npx (fastest)
```bash
# Run immediately (no installation needed)
npx vibe-kanban
# Specify port (default port 3000)
npx vibe-kanban --port 3001
# Specify port and environment variable together
PORT=3001 npx vibe-kanban --port 3001
# Use wrapper script
bash scripts/vibe-kanban-start.sh
```
Browser opens `http://localhost:3000` automatically.
> ⚠️ **Port conflict warning**: If another dev server like Next.js is using port 3000,
> run `PORT=3001 npx vibe-kanban --port 3001`.
> Confirm `Main server on :3001` in the startup log, then visit `http://localhost:3001`.
Normal startup log:
```
Starting vibe-kanban v0.1.17...
No user profiles.json found, using defaults only
Starting PR monitoring service with interval 60s
Remote client initialized with URL: https://api.vibekanban.com
Main server on :3001, Preview proxy on :XXXXX
Opening browser...
```
### Clone + dev mode
```bash
git clone https://github.com/BloopAI/vibe-kanban.git
cd vibe-kanban
pnpm i
pnpm run dev
```
---
## Environment Variables
| Variable | Description | Default |
|------|------|--------|
| `PORT` | Server port | `3000` |
| `HOST` | Server host | `127.0.0.1` |
| `VIBE_KANBAN_REMOTE` | Allow remote connections | `false` |
| `VK_ALLOWED_ORIGINS` | CORS allowed origins | Not set |
| `DISABLE_WORKTREE_CLEANUP` | Disable worktree cleanup | Not set |
| `ANTHROPIC_API_KEY` | For Claude Code agent | — |
| `OPENAI_API_KEY` | For Codex/GPT agent | — |
| `GOOGLE_API_KEY` | For Gemini agent | — |
Set in `.env` file before starting the server.
> **API key location per agent (Settings → Agents → Environment variables)**
> - Claude Code: `ANTHROPIC_API_KEY`
> - Codex: `OPENAI_API_KEY`
> - Gemini: `GOOGLE_API_KEY`
> - Opencode: No separate setup needed (built-in auth)
---
## MCP Configuration
Vibe Kanban runs as an MCP (Model Context Protocol) server, letting agents control the board directly.
### Claude Code MCP Setup
`~/.claude/settings.json` or project `.mcp.json`:
```json
{
"mcpServers": {
"vibe-kanban": {
"command": "npx",
"args": ["vibe-kanban", "--mcp"],
"env": {
"MCP_HOST": "127.0.0.1",
"MCP_PORT": "3001"
}
}
}
}
```
### OpenCode MCP Setup
Add to `~/.config/opencode/opencode.json`:
```json
{
"mcp": {
"vibe-kanban": {
"command": "npx",
"args": ["vibe-kanban", "--mcp"],
"env": {
"MCP_HOST": "127.0.0.1",
"MCP_PORT": "3001"
}
}
}
}
```
After restarting, `vk_*` tools are available directly in your OpenCode session.
### MCP Tool List
| Tool | Description |
|------|------|
| `vk_list_cards` | List all cards (workspaces) |
| `vk_create_card` | Create a new card |
| `vk_move_card` | Change card status |
| `vk_get_diff` | Get card diff |
| `vk_retry_card` | Re-run a card |
> ⚠️ **Tool name changes from older versions**: `vk_list_tasks` → `vk_list_cards`, `vk_create_task` → `vk_create_card`
> These are the confirmed tool names from the actual MCP API as of v0.1.17.
### Codex MCP Integration
To connect Vibe Kanban with Codex, run the following from your project root:
```bash
bash scripts/mcp-setup.sh --codex
```
This command adds the `vibe-kanban` MCP server config to `~/.codex/config.toml`.
Hook-based auto-looping is not default Codex behavior, so retry/loop management is handled via board card progress states or a higher-level orchestrator.
---
## Workspace → Parallel Agents → PR Workflow
> **v0.1.17 actual UI structure**: Vibe Kanban is a Kanban board, but
> the actual unit of work is a **Workspace**.
> Each workspace handles one task independently.
### 1. Start the server
```bash
# Default run
npx vibe-kanban
# → http://localhost:3000
# Port conflict (e.g., Next.js)
PORT=3001 npx vibe-kanban --port 3001
# → http://localhost:3001
```
### 2. (Optional) Review epic plan with planno
```text
Review the implementation plan for this feature with planno
```
planno (plannotator) is an independent skill — usable without Vibe Kanban.
### 3. Create a Workspace
1. Open the UI → click **"+ Create Workspace"** or the `+` button in the left sidebar
2. **Which repositories?** screen:
- **Browse** → select a git repo from the filesystem (manual path entry supported)
- **Recent** → previously used repos
- Select a repo, then choose a branch (default: `main`)
- Click **Continue**
3. **What would you like to work on?** screen:
- Select an agent (Opencode, Claude Code, Codex, Gemini, Amp, Qwen Code, Copilot, Droid, Cursor Agent)
- Enter a task description (Markdown supported)
- Select a mode (Default, Build, etc.)
- Click **Create**
### 4. Automatic agent execution
When a workspace is created:
- A `vk/<hash>-<slug>` branch is created automatically (e.g., `vk/3816-add-a-comment-to`)
- A git worktree is created automatically (fully isolated per agent)
- The selected agent CLI runs with log streaming
Workspace states:
- **Running**: Agent is executing (left sidebar)
- **Idle**: Waiting
- **Needs Attention**: Agent finished or needs input
### 5. Review results
- **Changes panel**: View file diffs
- **Logs panel**: Agent execution logs
- **Preview panel**: Web app preview
- **Terminal**: Run commands directly
- **Notes**: Write notes
### 6. Create PR & finish
- Workspace detail → **"Open pull request"** button
- PR merge → workspace moves to Archive
- Worktree cleaned up automatically
---
## Git Worktree Isolation Structure
Workspace directory (configurable in Settings → General → Workspace Directory):
```
~/.vibe-kanban-workspaces/ ← default location (under home directory)
├── <workspace-uuid-1>/ ← workspace 1 isolated environment
├── <workspace-uuid-2>/ ← workspace 2 isolated environment
└── <workspace-uuid-3>/ ← workspace 3 isolated environment
```
Branch naming (configurable in Settings → General → Git → Branch Prefix):
```
vk/<4-char ID>-<task-slug>
e.g.: vk/3816-add-a-comment-to-readme
```
Internal behavior:
```bash
git worktree add <workspace-dir> -b vk/<hash>-<task-slug> main
<agent-cli> -p "<task-description>" --cwd <workspace-dir>
```
> **Recommended .gitignore entries:**
> ```
> .vibe-kanban-workspaces/
> .vibe-kanban/
> ```
---
## Remote Deployment
### Docker
```bash
# Official image
docker run -p 3000:3000 vibekanban/vibe-kanban
# Pass environment variables
docker run -p 3000:3000 \
-e ANTHROPIC_API_KEY=$ANTHROPIC_ARelated in Code Review
gstack
IncludedFast headless browser for QA testing and site dogfooding. Navigate pages, interact with elements, verify state, diff before/after, take annotated screenshots, test responsive layouts, forms, uploads, dialogs, and capture bug evidence. Use when asked to open or test a site, verify a deployment, dogfood a user flow, or file a bug with screenshots. (gstack)
startup-due-diligence
IncludedLegal due diligence review for seed-stage and Series A startups (US, Delaware C-Corp focus). Supports both investor and founder perspectives. Capabilities include: (1) Interactive document review and issue spotting; (2) Document request list generation; (3) Cap table and SAFE/convertible note analysis; (4) Red flag identification with severity ratings; (5) Diligence report generation. TRIGGERS: due diligence, DD, startup investment, cap table review, Series A, seed round, investor diligence, legal review startup, SAFE analysis, convertible note, 409A, founder vesting.
interview-master
IncludedThis skill should be used when the user asks to "generate interview questions", "prepare for interview", "optimize resume", "conduct mock interview", "analyze git commits for resume", "generate resume from code", "review my resume", or mentions interview preparation, career assistance, or extracting project experience from git history. Provides comprehensive interview and career development guidance for both job seekers and interviewers.
fix-issue
IncludedFixes GitHub issues using parallel analysis agents for root cause investigation, code exploration, and regression detection. Reads issue context from gh CLI, searches codebase and memory for related patterns, generates a fix with tests, and links the resolution back to the issue via PR. Includes prevention analysis to avoid recurrence. Use when debugging errors, resolving regressions, fixing bugs, or triaging issues.
sf-apex
IncludedGenerates and reviews Salesforce Apex code with 150-point scoring. TRIGGER when: user writes, reviews, or fixes Apex classes, triggers, test classes, batch/queueable/schedulable jobs, or touches .cls/.trigger files. DO NOT TRIGGER when: LWC JavaScript (use sf-lwc), Flow XML (use sf-flow), SOQL-only queries (use sf-soql), or non-Salesforce code.
swift-development
IncludedComprehensive Swift development for building, testing, and deploying iOS/macOS applications. Use when Claude needs to: (1) Build Swift packages or Xcode projects from command line, (2) Run tests with XCTest or Swift Testing framework, (3) Manage iOS simulators with simctl, (4) Handle code signing, provisioning profiles, and app distribution, (5) Format or lint Swift code with SwiftFormat/SwiftLint, (6) Work with Swift Package Manager (SPM), (7) Implement Swift 6 concurrency patterns (async/await, actors, Sendable), (8) Create SwiftUI views with MVVM architecture, (9) Set up Core Data or SwiftData persistence, or any other Swift/iOS/macOS development tasks.