SaravananJaichandar/world-model-mcp

What it does

World Model MCP is an MCP server that creates a temporal knowledge graph for codebases. It captures decision traces throughout code development, links test failures directly to code changes, analyzes co-edit patterns, and predicts regression risks. This tool helps development teams understand the relationship between code changes and test outcomes while learning patterns from their codebase history.

Key features

• Temporal knowledge graph construction from codebase history
• Decision trace capture and visualization
• Test failure to code change correlation
• Co-edit pattern learning and analysis
• Regression risk prediction
• Learned constraint enforcement at edit boundaries
• Support for multiple platforms (Windows, Linux, macOS)

How to set up

As an MCP server, installation involves adding it to your MCP configuration. Clone the repository from GitHub, ensure Python 3 is installed, and configure it with your Claude client or compatible AI tool that supports MCP protocol. Refer to the project repository for detailed setup instructions specific to your platform.