codecks-mcp
A TypeScript MCP server for Codecks project management that provides over 32 tools to manage cards, decks, milestones, and PM workflows. It enables users to perform actions like creating cards, tracking activity, and managing project conversations through the Model Context Protocol.
README
codecks-mcp
TypeScript MCP server for Codecks project management. Exposes 32+ tools for managing cards, decks, milestones, tags, and PM workflows via the Model Context Protocol.
Built with @modelcontextprotocol/server, zod, and Express.
Quick Start
# Run directly (no install needed)
npx codecks-mcp
# Or install globally
npm install -g codecks-mcp
codecks-mcp
Configuration
Create a .env file in your working directory:
CODECKS_TOKEN=your_session_cookie
CODECKS_ACCOUNT=your_account_slug
Tokens
| Variable | Source | Expires |
|---|---|---|
CODECKS_TOKEN |
Browser DevTools > Cookie at |
Yes |
CODECKS_REPORT_TOKEN |
CLI generate-token command |
No |
CODECKS_ACCOUNT |
Your Codecks account slug | N/A |
CODECKS_USER_ID |
Auto-discovered if unset | N/A |
Optional Settings
| Variable | Default | Description |
|---|---|---|
CODECKS_MCP_RESPONSE_MODE |
legacy |
legacy or envelope |
CODECKS_HTTP_TIMEOUT_SECONDS |
30 |
Request timeout |
CODECKS_HTTP_MAX_RETRIES |
2 |
Retry count |
Transports
stdio (default)
codecks-mcp
HTTP
codecks-mcp --transport http --port 3000
Endpoints:
POST /mcp— MCP protocol (Streamable HTTP)GET /ping— Health check
IDE Setup
Claude Code
{
"mcpServers": {
"codecks": {
"command": "npx",
"args": ["-y", "codecks-mcp"]
}
}
}
Cursor
{
"mcpServers": {
"codecks": {
"command": "npx",
"args": ["-y", "codecks-mcp"]
}
}
}
Remote (HTTP)
{
"mcpServers": {
"codecks": {
"url": "http://localhost:3000/mcp"
}
}
}
Tools
Read (10)
| Tool | Description |
|---|---|
get_account |
Current account info |
list_cards |
List/filter cards with pagination |
get_card |
Full card details |
list_decks |
All decks |
list_projects |
Projects with decks |
list_milestones |
Milestones |
list_tags |
Project-level tags |
list_activity |
Recent activity feed |
pm_focus |
PM dashboard (blocked, stale, suggested) |
standup |
Daily standup summary |
Hand (3)
| Tool | Description |
|---|---|
list_hand |
Personal work queue |
add_to_hand |
Add cards to hand |
remove_from_hand |
Remove cards from hand |
Mutation (9)
| Tool | Description |
|---|---|
create_card |
Create card (supports sub-cards) |
update_cards |
Batch update properties |
mark_done |
Mark cards done |
mark_started |
Mark cards started |
archive_card |
Archive (reversible) |
unarchive_card |
Restore archived |
delete_card |
Permanent delete |
scaffold_feature |
Hero + lane sub-cards |
split_features |
Batch-split features |
Comments (5)
| Tool | Description |
|---|---|
create_comment |
Start thread |
reply_comment |
Reply to thread |
close_comment |
Resolve thread |
reopen_comment |
Reopen thread |
list_conversations |
List threads |
PM Session (3)
| Tool | Description |
|---|---|
get_pm_playbook |
Methodology guide |
get_workflow_preferences |
Load preferences |
save_workflow_preferences |
Save preferences |
Feedback (2)
| Tool | Description |
|---|---|
save_cli_feedback |
Save feedback |
get_cli_feedback |
Read feedback |
Security
- Prompt injection detection (6 regex patterns)
[USER_DATA]boundary tagging on user-authored content- Input validation with zod schemas + length limits
- UUID validation on all card IDs
- Control character stripping
Development
# Install
pnpm install
# Dev mode (tsx, auto-reload)
pnpm dev
# Build
pnpm build
# Test
pnpm test
# Lint + format
pnpm lint
pnpm format:check
# Type check
pnpm typecheck
Also Available
- Python version: codecks-mcp-python — same tools, Python runtime
- Full CLI: codecks-cli — CLI + library with formatters, GDD sync, and more
License
MIT
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。