Skip to main content
Glama
README.md4.16 kB
# MCP Planning Server MCP server for Claude Code that helps organize and track software development tasks. ## What is this? This is a plugin for Claude Code (Anthropic's AI coding assistant). It adds planning tools that help: - Break down complex tasks into steps - Track what's done and what's next - Record technical decisions - Compare different solution approaches ## Prerequisites Before installing, make sure you have: - **Node.js** (version 18 or higher) - download from https://nodejs.org - **Claude Code** - the CLI tool from Anthropic To check if Node.js is installed, run in terminal: ```bash node --version ``` ## Installation 1. Clone or download this repository 2. Open terminal in the project folder 3. Install dependencies: ```bash npm install ``` 4. Build the project: ```bash npm run build ``` After build, you should see a `dist` folder with compiled files. ## Configuration To use this server with Claude Code, you need to add it to MCP settings. ### Step 1: Find your settings file Settings can be placed in one of these locations: | Location | File path | Use case | |----------|-----------|----------| | Project (private) | `.claude/settings.local.json` | Only for you, not shared | | Project (shared) | `.claude/settings.json` | Shared with team via git | | Global | `~/.claude/settings.json` | All your projects | ### Step 2: Add server configuration Open or create the settings file and add: ```json { "mcpServers": { "planning": { "command": "node", "args": ["FULL_PATH_TO/planner/dist/index.js"], "env": { "MCP_PLANNING_STORAGE_PATH": "./.mcp-plans" } } } } ``` **Important:** Replace `FULL_PATH_TO` with the actual path to this folder. Example paths: - Windows: `D:/Projects/planner/dist/index.js` - Mac/Linux: `/home/user/projects/planner/dist/index.js` ### Step 3 (Optional): Allow tools without confirmation By default, Claude Code asks permission for each tool use. To skip confirmation for planning tools, add this section to your settings file: ```json { "permissions": { "allow": [ "mcp__planning" ] } } ``` Your complete settings file will look like this: ```json { "mcpServers": { "planning": { "command": "node", "args": ["FULL_PATH_TO/planner/dist/index.js"], "env": { "MCP_PLANNING_STORAGE_PATH": "./.mcp-plans" } } }, "permissions": { "allow": [ "mcp__planning" ] } } ``` **Note:** Use exactly `mcp__planning` (not `mcp__planning__*`). Wildcards are not supported. ## How to use Once configured, just talk to Claude Code naturally. ### Starting a new plan Tell Claude Code what you want to build: > "I need to build a user login system" Claude Code will automatically: 1. Create a plan 2. Add requirements 3. Suggest solutions 4. Break work into phases ### Checking progress Ask Claude Code about your plan: > "What's the status of my plan?" > "What should I work on next?" ### Updating progress Tell Claude Code when you finish something: > "I finished the login form" Claude Code will update the plan and suggest what to do next. ## Available tools The server provides tools in 8 categories: | Category | Description | |----------|-------------| | Plan | Create and manage plans | | Requirement | Define what needs to be built | | Solution | Compare different approaches | | Decision | Record technical choices | | Phase | Break work into steps | | Link | Connect related items | | Query | Search and analyze plans | | Artifact | Store code and configs | ## Where plans are stored Plans are saved as JSON files: ``` .mcp-plans/ plans/ {plan-id}/ manifest.json requirements.json solutions.json decisions.json phases.json links.json artifacts.json active-plans.json ``` ## Troubleshooting **Claude Code doesn't see the planning tools** - Check that the path in settings is correct - Make sure you ran `npm run build` - Restart Claude Code after changing settings **Permission errors** - Check that the storage path is writable - On Windows, avoid paths with spaces if possible ## License MIT

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/cppmyjob/cpp-mcp-planner'

If you have feedback or need assistance with the MCP directory API, please join our Discord server