Agent Progress Tracker MCP Server

An MCP (Model Context Protocol) server that enables AI agents to track, search, and retrieve their progress across projects. Provides persistent memory and context sharing for multi-step or multi-agent workflows.

Features

• Log Progress - Record completed work with structured metadata
• Search Logs - Discover relevant entries by title, date, keywords, or tags
• Get Context - Retrieve summarized information about prior work by ID
• Project-scoped - All entries organized by project to keep context relevant
• LLM-powered Summarisation - Automatic summarisation using OpenAI-compatible APIs
• SQLite Storage - Fast, local database with full-text search capabilities
• Comprehensive Error Handling - Graceful fallbacks and detailed error reporting

Quick Start

1. Install

npm install -g @thesammykins/agent-progress-mcp

2. Configure API Key

export OPENAI_API_KEY=your-api-key-here

3. Add to Claude Desktop

Add to your Claude Desktop claude_desktop_config.json:

{
  "mcpServers": {
    "agent-progress": {
      "command": "agent-progress-mcp",
      "env": {
        "OPENAI_API_KEY": "your-api-key-here"
      }
    }
  }
}

4. Start Tracking Progress

• Log completed work: log_progress
• Search previous work: search_logs
• Get context: get_context

Documentation

• Configuration Guide - Environment variables, MCP host setup, alternative providers
• Development Guide - Contributing, testing, building from source
• Tools Overview - Technical details about each tool
• How-To Guide - Practical usage patterns and workflows
• Troubleshooting - Common issues and solutions
• Technical Specification - Complete technical specification

Usage Example

# Log completed work
log_progress(
  projectId="web-app",
  title="Implemented JWT authentication",
  content="Created JWT token generation/validation utilities, added auth middleware, implemented login/register endpoints with bcrypt password hashing. All tests passing.",
  tags=["authentication", "security", "api"]
)

# Search for related work
search_logs(projectId="web-app", query="authentication")

# Get detailed context
get_context(projectId="web-app", id="found-entry-id", includeFull=true)

Installation Options

Development Install

git clone https://github.com/thesammykins/context-mcp.git
cd context-mcp
npm install
npm run build
npm link  # Optional: for local development

MCP Host Integration

The server works with any MCP-compatible host:

• Claude Desktop - Native integration
• Cursor - Add to MCP servers configuration
• MCP CLI - Command-line usage
• Custom hosts - STDIO transport protocol
• OpenCode - Add to opencode.jsonc configuration

OpenCode Configuration

Add to your opencode.jsonc:

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "agent-progress": {
      "type": "local",
      "command": ["npx", "-y", "@thesammykins/agent-progress-mcp"],
      "environment": {
        "OPENAI_API_KEY": "{env:OPENAI_API_KEY}"
      }
    }
  }
}

Use use the agent-progress tool in your prompts to enable progress tracking.

Database & Privacy

• Local Storage: All data stored locally on your machine
• SQLite Database: Fast, reliable, file-based storage
• Optional Encryption: Password-protect your progress data
• No Telemetry: Your work context never leaves your environment (except for optional LLM summarisation)

License

MIT License - See LICENSE file for details.

Contributing

We welcome contributions! Please see the Development Guide for details on contributing, testing, and building from source.

Support

• GitHub Issues - Report bugs or request features
• GitHub Discussions - Ask questions and share ideas
• Documentation - Complete documentation and guides