@bagaking/projmcp
v0.3.11
Published
MCP server for AI agent project management using best practices methodology
Maintainers
Readme
MCP Project Plan Server
A Model Context Protocol (MCP) server designed to help AI agents better utilize project organization best practices. This server provides standardized APIs for project management, documentation organization, and workflow automation.
Features
- File Management: List and organize project files by type (sprint, doc, code, opinion)
- Document Operations: Record and retrieve documentation with proper categorization
- Project Status: Quick access to current project state and planning documents
- Template Generation: Initialize project structures following best practices
- Token Efficient: Optimized responses to minimize AI agent token consumption
Installation
npm install -g @bagaking/projmcpQuick Start
One-command setup:
claude mcp add projmcp -- npx -y @bagaking/projmcpThat's it! All tools are now available in Claude Code conversations.
MCP Tools
The server provides these tools for AI agents:
Available Tools
list_files- List and analyze project files by type (all/sprint/doc/code/opinion)show_current- Display current project status from CURRENT.mdshow_plan- Display project plan from PLAN.mdinit_project_plan- Initialize complete project structurerecord- Record structured documentation (doc/code/opinion types)query_sprint- Query specific sprint information by IDright_now- Get current time in multiple formats
Usage: Just call tools directly in conversations or use /mcp tool_name syntax.
Response Format
All tools return structured JSON with:
- Content: Main response data
- Metadata: Statistics and context
- Timestamps: Right-now time information
Project Directory Structure
After running init_project_plan, your project will have this structure:
project_plan/
├── PLAN.md # Master project planning document
├── CURRENT.md # Current execution status
├── M01_S01.initial_setup.md # Sprint execution plans
├── DOCREF_001.topic.md # External documentation references
├── CODEREF_001.topic.md # Code implementation references
└── OPINIONS_001.topic.md # Key decisions and observationsBest Practices Methodology
This server implements the "项目组织的最佳实践" (Project Organization Best Practices):
- Structured Planning: Separate planning (PLAN.md) from execution tracking (CURRENT.md)
- Sprint Organization: Each milestone/sprint gets dedicated planning documents
- Documentation Types:
DOCREF: External documentation and researchCODEREF: Implementation patterns and code examplesOPINIONS: Decision records and key observations
- Metadata Tracking: All documents include creation dates, purposes, and evidence levels
- Flat File Structure: All files in single directory with semantic naming
Source Code Structure
src/
├── index.ts # Server entry point
├── server.ts # MCP server core logic
├── interfaces/ # TypeScript interfaces
│ └── core-interfaces.ts # Core abstractions (ITool, IFileManager, etc.)
├── services/ # Core services
│ ├── logger.ts # Structured logging service
│ ├── mcp-server.ts # Enhanced MCP server with DI
│ └── tool-registry.ts # Tool registration and management
├── tools/ # MCP tool implementations
│ ├── list-files.ts # File listing with metadata
│ ├── init-project.ts # Project initialization
│ ├── show-status.ts # Status display tools
│ ├── record.ts # Documentation recording
│ └── query-sprint.ts # Sprint querying
├── utils/ # Utilities
│ ├── file-manager.ts # Secure file operations
│ ├── security-validator.ts # File security validation
│ └── template-generator.ts # Document template generation
└── types.ts # TypeScript type definitionsDevelopment
# Install dependencies
npm install
# Build project
npm run build
# Run linting
npm run lint
# Start in development mode
npm run devRequirements
- Node.js >= 18.0.0
- TypeScript >= 5.0.0
License
MIT
