TaskTracker MCP Server
A DAG-based task tracking MCP server for structured bug analysis and investigation workflows, with dependency management, priority-based execution, and automatic circular dependency detection.
README
TaskTracker MCP Server
A DAG-based task tracking server for bug analysis and investigation workflows, built as a Model Context Protocol (MCP) server.
Features
- DAG-based task dependencies — declare dependencies for structured investigation workflows
- Circular dependency detection — automatic validation prevents invalid chains
- Priority-based execution — tasks sorted by high/medium/low priority
- Atomic bulk operations — add entire investigation plans in one call
- Rich status tracking —
pending→in_progress→completed/skipped/blocked - Evidence-based resolution — every resolved task requires a finding
- Analysis gate —
conclude_analysis()blocks until all tasks are resolved
Installation
Option 1 — npx (no install needed)
npx tasktracker-mcp
Option 2 — Global install
npm install -g tasktracker-mcp
tasktracker-mcp
Option 3 — From source
git clone https://github.com/letscodekrkumar/tasktracker-mcp.git
cd tasktracker-mcp
npm install
npm run build
npm start
MCP Server Configuration
Claude Desktop
Edit claude_desktop_config.json:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Using npx (recommended):
{
"mcpServers": {
"tasktracker": {
"command": "npx",
"args": ["tasktracker-mcp"]
}
}
}
Using global install:
{
"mcpServers": {
"tasktracker": {
"command": "tasktracker-mcp"
}
}
}
Using local source build:
{
"mcpServers": {
"tasktracker": {
"command": "node",
"args": ["/absolute/path/to/tasktracker-mcp/dist/index.js"]
}
}
}
Claude Code (CLI)
claude mcp add tasktracker npx tasktracker-mcp
Or with a local build:
claude mcp add tasktracker node /absolute/path/to/tasktracker-mcp/dist/index.js
Verify it connected:
claude mcp list
# tasktracker: npx tasktracker-mcp - ✓ Connected
Other MCP Clients
Any client that supports the MCP stdio transport can use:
{
"command": "npx",
"args": ["tasktracker-mcp"]
}
Tools
| Tool | Purpose |
|---|---|
add_task |
Register a single task with optional dependencies |
add_tasks_bulk |
Register an entire investigation DAG atomically |
update_task |
Resolve task with evidence, update status, or rewire deps |
get_ready_tasks |
Get all executable tasks (deps resolved), priority-sorted |
get_all_tasks |
Full DAG snapshot with statuses, findings, and dependency state |
conclude_analysis |
Gate that blocks until all tasks resolved; returns summary |
reopen_task |
Reopen a blocked task when its blocker is resolved |
reset |
Clear all tasks and start fresh |
Quick Start
# 1. Define your investigation DAG
add_tasks_bulk([
{"title": "Fetch bug fields", "category": "log_check", "priority": "high"},
{"title": "Extract machine ID", "category": "log_check", "priority": "high"},
{"title": "Fetch run.log — search for TIMEOUT errors", "category": "log_check", "priority": "high", "depends_on": ["T2"]},
{"title": "Identify root cause", "category": "root_cause", "priority": "high", "depends_on": ["T1", "T2", "T3"]}
])
# 2. Execute ready tasks
tasks = get_ready_tasks() # returns T1, T2
# 3. Resolve tasks as you go
update_task("T1", status="completed", finding="Bug filed 2026-03-28. Build: 4.2.1-rc3.")
update_task("T2", status="completed", finding="Machine ID: X200-lot-7.")
# 4. T3 is now unblocked
update_task("T3", status="completed", finding="3 TIMEOUT errors at 14:22:01")
# 5. Conclude when all done
conclude_analysis()
Task Status Transitions
| Transition | Allowed | Notes |
|---|---|---|
| pending → in_progress | Yes | No dependency check |
| pending/in_progress → completed | Yes | All deps must be resolved; finding required |
| pending/in_progress → skipped | Yes | Finding required |
| pending/in_progress → blocked | Yes | Finding required |
| completed / skipped → any | No | Permanently resolved |
| blocked → any | No | Use reopen_task() to reopen |
Development
npm test # run tests
npm run test:coverage # with coverage report
npm run build # compile TypeScript
npm run dev # build + run
npm run benchmark # performance benchmarks
Project Structure
src/
index.ts # MCP server entry point
tracker.ts # DAG engine (task state, dependency resolution)
types.ts # TypeScript interfaces and validators
monitor.ts # Progress monitoring
examples.ts # Usage examples
benchmark.ts # Performance benchmarks
__tests__/ # Test suite
docs/ # Design docs, deployment guide, specs
License
MIT — see LICENSE
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。