MCP Hub
Back to servers

backlog-mcp

A lightweight task tracking system for AI agents that allows for managing, filtering, and archiving project tasks through a single unified interface.

Tools
1
Updated
Dec 27, 2025
Validated
Jan 11, 2026

backlog-mcp

Minimal task backlog MCP server for Claude and AI agents.

Quick start: Tell your LLM: Add backlog-mcp to .mcp.json and use it to track tasks

Task Schema

{
  id: string;           // TASK-0001
  title: string;
  description?: string;
  status: 'open' | 'in_progress' | 'blocked' | 'done' | 'cancelled';
  created_at: string;   // ISO8601
  updated_at: string;   // ISO8601
  blocked_reason?: string;
  evidence?: string[];
}

MCP Tool

Single unified tool with action parameter:

backlog action="list"                         # List all tasks
backlog action="list" summary=true            # Get counts by status
backlog action="list" status=["open"]         # Filter by status
backlog action="get" id="TASK-0001"           # Get task details
backlog action="create" title="Fix bug"       # Create task
backlog action="update" id="TASK-0001" set_status="done"  # Update task

Installation

Add to your MCP config (.mcp.json or Claude Desktop config):

{
  "mcpServers": {
    "backlog": {
      "command": "npx",
      "args": ["-y", "backlog-mcp"]
    }
  }
}

Or build from source:

git clone https://github.com/gkoreli/backlog-mcp.git
cd backlog-mcp
npm install && npm run build
npm start

Storage

  • Default: data/backlog.json (local to project)
  • Global: Set BACKLOG_DATA_DIR=~/.backlog for cross-project persistence
  • Completed/cancelled tasks auto-archive to archive.json
  • Atomic writes via temp + rename

License

MIT

Reviews

No reviews yet

Sign in to write a review