Skill

Redesign Existing Projects

leonxlnx

Reads the code already there and names what is weak in it, then lifts type, rhythm, colour and the states nobody built, on whatever styling the project happens to use. The binding constraint is that it improves in place rather than replacing, so structure and behaviour survive the pass.

A situation it fits

A four-year-old logistics app, Vue and styled-components, is consistent and tidy and looks like 2021. Nothing in it is inconsistent enough to call a defect. The whole product sits behind a login: dashboards, a five-step booking form, dense tables. There is no budget for a rewrite, and the team will not migrate off styled-components to get a nicer look.

Redesign Existing Projects. Read the code that is there, name what is weak, then lift type, rhythm, colour and the states nobody built, on whatever styling the project happens to use: improving in place is the binding constraint, which is what a no-rewrite, no-migration brief actually requires. The most tempting wrong answer is design-taste-frontend, which does take on visual overhauls of sites that already exist - its trap is scope: it is aimed at surfaces meant to be scrolled and read, and hands dashboards and multi-step wizards off to a product-UI system, which is this entire application. gpt-taste is the same push aimed at a different model, and nothing here is about repetition. image-to-code would draw a new look and build it, which is the rewrite that was ruled out.

Skills it gets confused with

These share a family with Redesign Existing Projects, which is another way of saying they are the ones you might reach for by mistake.

Knowing what Redesign Existing Projects does is the easy half. Telling it apart from the others under time pressure is the game.

Today's session

Redesign Existing Projects is part of leonxlnx/taste-skill. Licence: MIT. The description above was written for this game, not taken from the skill.