What it does
The tool writes a ranked queue of refactoring hypotheses without modifying your code. All current evaluation labels were generated by an LLM rather than human reviewers, and the tool is currently in an alpha state.
Agent tooling
jev-scanr
A read-only command-line tool that uses the Jev model to rank TypeScript and JavaScript functions for potential refactoring.
Bookmark: jev-scanr
Keep this in your collection.
Leave a noteWhat would you try with this?
: jev-scanr
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
How you can use it
To use this tool, you run a command in your terminal that scans your project files. It asks the model specific questions about your code, like whether two functions do the exact same thing.
After the scan, it gives you a list of places in your code to look at, ranked by how likely they are to need fixing. You or your coding assistant must check these suggestions manually before making any changes.