hslee16/Archy ? reverse-engineered prompt
Reverse engineered prompt
Build me a Python tool that reads a codebase’s import relationships and turns that into something I can actually use to keep the structure honest. I want a command line app and an MCP server that both use the same cached graph, so I can ask things like what a change would affect, which tests are likely impacted, where the cycles are, and whether the real imports still match the layers I declared in a config file. It should work well in CI and fail when the architecture rules are broken, not just warn about them.
Please make it feel practical for agent assisted development, so the answers are easy to pull into a coding session and quick enough to run often. If you need current docs for any libraries or Python features, look them up online. Also include a small set of tests and a clean setup so the project can be run locally without much fuss.
Are you gonna build this?
make sure you review the code using coderabbit