Show HN: Context.md – A proposed standard for AI project context
Context.md is a proposed standard for defining AI project context, aiming to help AI tools better understand and work with codebases by providing structured metadata and documentation within a project.
Background
- This is a GitHub project by user kerbelp proposing a standard file called `context.md` that would sit inside a code repository and describe the project's purpose, tech stack, architecture, conventions, and other background info in a structured format.
- The idea is that AI coding assistants (like GitHub Copilot, Cursor, or Claude-powered tools) could automatically read this file to understand the project before generating code, rather than guessing or missing important context.
- Currently, developers using AI tools often have to manually explain their project setup or rely on AI models inferring incomplete context from the codebase. A standard like `context.md` aims to solve this fragmentation.
- The proposal is loosely inspired by existing conventions like `README.md`, but is specifically targeted at machine consumption, not human reading — though it would be human-readable too.
- The discussion on Hacker News (Show HN) will likely debate whether this becomes a real convention, whether it overlaps with existing files, and what should go in it.