Code Modernization

Point Claude at a legacy codebase and get four things: an understanding of what it is and what it does, a plan you approve, the modernized code, and proof that the new code behaves like the old. It works with any language and any kind of move: a newer version of the same technology, a rewrite in another one, or a rebuild on a new architecture.

Install

You need Claude Code. Then:

/plugin install code-modernization@claude-plugins-official

Some steps start many agents at once, so expect real usage on a large system (see What to expect). Start with one module or one unit as a pilot, not the whole estate.

Start here

Open a folder for the work and type /code-modernization:modernize. It asks what you want to do with your code (a couple of short questions), finds the code, writes your answers down once, shows the road ahead, and gives you the exact first command. Every later command reads what you said, so nothing is asked twice. Not sure what you want? Choose Understand it first: you get the assessment, the map, the business rules and a plan, and no code is rebuilt.

To go step by step instead, point the first command at your code and follow the "next step" line each command ends with:

/code-modernization:modernize-preflight <name> --source <path to your code>
/code-modernization:modernize-status <name>       # where am I, and what is next: run it any time

is a short label of your choice (letters, digits, -, _). --source makes a link at legacy/ and copies nothing, so a huge repository stays where it is. Nothing edits your legacy source. Commands write only to analysis/ / and modernized/ (a build check writes build output where your build normally does), and each refreshes analysis/ /REPORT.html: one page, with everything found so far, that you can open or share. A few steps need someone who knows how the system is built and run, so bring an engineer for the first questions.

The path

Each step stands alone, so you can stop and review after any of them.

assess --portfolio surveys many systems and ranks them on one page. modernize-status says where you are and gives the exact command to paste.

A person decides at six points, and the plugin never decides them for you: the five preflight answers, the rules that look wrong (review), approving the brief, accepting any difference the proof finds, signing the proof, and applying the security patch.

How it proves the result

Modernization fails quietly: the new code passes the tests someone wrote, and differs on the inputs nobody thought of. So the proof is built from evidence a script can check, not from the model's opinion. modernize-verify runs after a build, ideally in a fresh session, and re-does the work instead of trusting the notes the build left behind:

• The tests, run again from clean. Build output is deleted, the full suite runs, and the counts come from the runner's own result files or a saved log. A run that executed no tests, or skipped them, is a failure, and a count the command merely typed in does not count.