MCP Work History

🤖 MCP server for logging AI tool activities with comprehensive metrics tracking

javascript

2

GitHub Stars

javascript

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "nocoo-mcp-work-history": {
      "command": "node",
      "args": [
        "/Users/your-username/path/to/mcp-work-history/src/index.js"
      ]
    }
  }
}

The MCP Work History Server lets AI tools log their activities into daily Markdown worklog files with precise timestamps, tool and model tracking, and useful metrics. It makes keeping a detailed, searchable activity history easy for your team and your tooling, while staying MCP-compatible for seamless integration with any MCP-enabled client.

How to use

You use the server by running it locally or remotely and then have your MCP-enabled AI tools send log_activity events. Each entry records the tool name, model used, time, and a concise description, plus optional metrics like tokens, duration, cost, and tags. Review today’s worklog file to get a fast, readable summary of what happened and when.

To start logging from your tools, configure the MCP client to target the Work History server. When an AI task completes, call log_activity with the required details and any optional metrics you track. The server creates a daily markdown file under logs/ named worklog-YYYY-MM-DD.md and appends a bullet entry for the event, making it simple to scan for outcomes, performance, and costs.

How to install

Prerequisites you need installed on your machine are Node.js and npm. Node provides a runtime for the MCP server, and npm handles package management and scripts.

Step by step install and run flow:

Install dependencies locally
Start the MCP server

Configuration and usage notes

The server exposes a single logging tool named log_activity. You configure your MCP clients to call this tool with the required and optional parameters shown below. The server writes to a per-day log in logs/worklog-YYYY-MM-DD.md using a concise, bullet-point format.

MCP server configurations

The following configuration blocks show how to connect Warp and Claude Desktop to the Work History server. These entries are standard MCP stdio configurations and assume you replace placeholder paths with your actual installation paths.

Examples of logging a new entry

Basic usage from a tool that logs simple activity: log_activity({ tool_name: "Warp", log_message: "Created React component for user dashboard" })

With full metrics: log_activity({ tool_name: "Warp", ai_model: "gemini-2.5-pro", log_message: "Refactored authentication system with OAuth integration", tokens_used: 1250, context_length: 8, duration_ms: 2300, cost_usd: 0.0043, success: true, tags: ["refactoring", "auth", "oauth"] })

Error example: log_activity({ tool_name: "GitHub Copilot", ai_model: "gpt-4", log_message: "Attempted to implement user profile endpoint", input_tokens: 800, output_tokens: 0, success: false, error_message: "Timeout error", tags: ["coding", "api"] })

Log file structure

Logs are stored in the logs/ directory with files named worklog-YYYY-MM-DD.md. Each file begins with a date header and contains timestamped bullet entries that include tool name, model, and a brief description.

Notes on security and maintenance

Keep the server up to date with security patches. Use least-privilege access for any filesystem paths and ensure logs contain no sensitive information beyond what you intend to capture. Regularly rotate or archive old logs if needed to manage disk usage.

Troubleshooting

If log entries fail to appear, verify that the MCP client can reach the Work History server and that the log_activity call includes the required parameters. Check the daily log file for write permissions and confirm that the logs/ directory exists. Review any error_message fields in failed entries to identify the underlying issue.

Available tools

log_activity

Main MCP tool that logs an AI tool's activity to the current day's worklog file. It accepts required fields for tool_name and log_message and optional metrics such as ai_model, tokens_used, context_length, duration_ms, cost_usd, success, error_message, and tags.