geek-prateek/prompt-grapher ? reverse-engineered prompt
Reverse engineered prompt
Build me a Python CLI tool that scans a codebase and automatically writes helpful Cursor rules for that repo.
I want it to look at a Graphify output file like graphify out graph.json or manifest.json, figure out the project’s main language, framework hints, architecture style, naming patterns, important folders, entry points, and whether tests exist, then send that summary to an OpenAI compatible model. It should generate a .cursor/rules/project-rules.mdc file and also write an AGENTS.md file at the repo root as a backup.
Please make it usable from the command line, with an init command that creates a .env file in the current project and an analyze command that runs on the current folder, refreshes the graph when needed, and supports reusing an existing graph file if I want. Keep it flexible enough to work on arbitrary projects, not just one framework. If you need to check current docs for the model or API compatibility, go ahead and look them up online.
Are you gonna build this?
make sure you review the code using coderabbit