MCPFast / Tools / Per-repo memory and telemetry for Claude Code with MCP

GitHubMCP★★★★☆

Per-repo memory and telemetry for Claude Code with MCP

Per-repo memory, outcome telemetry, and confidence gate for Claude Code, MCP-compatible and readable by other AI coding tools.

View on GitHub

Per-Repo Memory and Telemetry for Claude Code with MCP

This tool provides essential per-repository memory and outcome telemetry for Claude Code, integrating seamlessly with the MCP (Memory, Code, and Prompt) framework. Designed for developers, it enhances the capabilities of AI coding assistants by offering persistent, context-aware memory and actionable insights into model performance. This allows for more efficient development cycles and improved AI agent behavior.

What it Does

The core functionality of this MCP tool is to establish a dedicated memory space for each individual code repository. This means Claude Code, when interacting with a specific project, can access and retain context relevant only to that project, avoiding cross-contamination of information. Furthermore, it captures outcome telemetry, logging the results of code generation and modification tasks. This data is crucial for understanding the AI's effectiveness and identifying areas for improvement. A confidence gate is also implemented, allowing for programmatic control over when the AI's output is considered reliable enough for integration.

Key Features

Who it's For

This tool is targeted at AI developers, machine learning engineers, and software architects who are building or integrating AI coding assistants. It is particularly beneficial for those working with Claude Code and leveraging the MCP framework. Developers seeking to improve the context awareness, performance tracking, and reliability of their AI coding agents will find this tool invaluable. It empowers users to create more sophisticated and dependable AI-powered development workflows.