frontend-aestheticslisted

# frontend-aesthetics Generated UI is bad in a specific, boring way: the model does not make a mistake, it makes **the average**. Left to its defaults it emits the statistical center of every landing page it ever saw, which is why AI sites all look like each other rather than like anything. Fixing that is bias-correction, not taste. This skill is a method plus an enforcement mechanism: 1. **Read the room** before designing (below). 2. **Set the dials** explicitly, with reasons. 3. **Design against the defaults**, using `references/tells.md`. 4. **Gate the output** on `scripts/slop_check.py` - mechanical, deterministic, no model in the loop. FAIL blocks. 5. **Answer the judgment questions** the linter cannot. Steps 1-3 are craft. Step 4 is the part most design skills lack: their "checklist" is run by the same model that wrote the code, which is a model grading its own homework. A rule a machine can check should be checked by a machine. ## Step 1 - The design read (one line, before any code) Declare what you think you are building. Get it wrong out loud and the user can correct you in one word, before you have written 400 lines in the wrong key. > Reading this as: **\<page kind>** for **\<audience>**, in a **\<vibe>** language, > leaning toward **\<system / aesthetic family>**. The audience picks the aesthetic, not your taste and not mine. A public-sector service and a fashion label want opposite things and neither is "better design". Constraints that **override** a