Telling Technology · Learning in public
All episodes
EP.113 — POLISH LOOP
A Claude Code workflow, for pages you use every day

My page passed its check. Then a loop found seventeen more things.

The Rolodex looked done at one width. A Claude Code loop then polished it for two hours at night, one commit roughly every eight minutes, and every round fixed something my first check could not see: another width, another state, the biggest dataset.

One dark card wheel shown seven times side by side, from a wide desktop down to a phone and a phone held sideways, each frame with its width written under it in teal
⌁ v02 → v18 · 23:10 → 01:17 · 7 viewports · stopped itself after 3 empty rounds
First, the day it looked finished.
01 The false finish

It passed. It was not done.

The Rolodex is my Ctrl+Alt+R card wheel of every project, ticket, todo and session. I built it on 2026-09-19 in one day, twelve tasks, and it passed its own gate. That gate shot one width: 1680 by 980. So it never saw that on a 390 px phone the top bar was 1,180 px wide and pushed the wheel off to the right. Not careless. Invisible by construction. “Looks right at one width” is not done.

claude code — desktop
> the rolodex passes verify.mjs. is it done?
It passes the gate it has. That gate shoots 1680 x 980 only, against today’s data. It has never seen a phone, a 900 px half screen, an empty search, or the biggest deck. Passing means nothing broke where we looked. It does not mean finished.
One large desktop screenshot of the card wheel with a green PASS stamp, and six small greyed-out frames beside it for the widths the check never looked at, the phone frame marked in coral with 1,180 px
the one width it checked
Mockup. The 1,180 px top bar is from the loop’s baseline notes (v00).
02 The loop

Seventeen commits between 23:10 and 01:17.

On 2026-09-21 I typed one line: “make a loop the improves and polishes the rolodex just like you did on the web page earlier”. Claude Code wrote the rules and ran a self-paced /loop on the ticket’s own branch (tix itm_1c50f271). Every round: shoot the page in every view, read every picture, compare with the last version, fix three to five things, pass a hard speed budget (first card under 600 ms, a flip under 4 ms, zero network requests while flipping), then commit polish(rolodex): vNN with a changelog line. v02 landed at 23:10, v18 at 01:17. Then it ran three more rounds, found nothing worth changing, and stopped itself at 01:34.

claude code — desktop
> polish round on the rolodex, ticket itm_1c50f271. read the loop prompt and follow it exactly.
Gate shots read: 1680, 1280, 900, phone, phone on its side. Found: facts and every action button clipped away at 844 x 390. Fixed: any screen under 500 px tall acts like a phone, with a caption and a tap-up sheet. New gate scenarios added. Gate PASS, worst flip 1.0 ms. Committed v11.
A vertical timeline of seventeen commits from v02 at 23:10 to v18 at 01:17, each with its time and a short subject, colour-coded by kind: states, widths, scale, data and feel, ending in three grey empty rounds and a stop marker at 01:34
v02 → v18, one ticket
Real commit times and subjects. Terminal card paraphrased from the v11 changelog entry.
03 The honest epilogue

The loop ran for two hours. The merge waited eight days.

The loop was not allowed to merge, on purpose: it only commits on its branch, and I ship. So the 17 commits sat on tix/itm_1c50f271-rolodex-polish from 2026-09-22 until 2026-09-30. By then main had moved: another ticket had added a local time stamp to the same “Last touched” cell. The rebase replayed all 17 rounds, and the shots caught what a clean rebase hides: v03’s recency bars sat on top of the new stamp at every width. One more fix, then the full gate and the new seven-viewport check both passed, 654 of 654 tests green, and it landed on main as 91367ab. The Rolodex I open with Ctrl+Alt+R is v18 now. I checked.

claude code — desktop
> rebase the polish rounds on main and check the shots before I merge
17 rounds replayed, none dropped. Conflicts in v02, v03, v08, all on the Last touched line. The shots show v03’s bars covering main’s stamp (“just now · 09” at 900 px). Moved the stamp to its own line. Gate PASS, page-gate 7 viewports PASS, 654/654 tests.
A before and after crop of one info panel row: before, small coloured recency bars drawn over a time stamp so it reads just now 09; after, the stamp on its own line under the age with the bars beside it
found in the rebase shots
Mockup of the collision. The wording is from the rebase notes.
⚠️

Honest part: the loop was not perfect either.

In v02 it dropped keyboard focus rings as “low value here”. Five rounds later, v07 put them back. v15 went out as perf(rolodex) instead of polish(rolodex), and it left the typo in place because its rules forbid rewriting history. And one problem it could only hide: session cards titled /clear or a single dot need a fix in the indexer, not the page, so it wrote that down as a proposal instead of building it.

04 What the rounds found

Widths, states, the biggest deck, and the rule it left behind

Sorted by kind, the seventeen rounds found states (an empty search, cards with no picture, a calm empty state in v10), widths (a phone info sheet in v04, a 900 px Sessions bar in v10, a phone on its side in v11), scale (1,668 session cards broke the sheet, the scrubber and the A-Z rail in v15 to v17), data (Claude Code markup as session titles in v18) and feel (pull to close, a sliding rail, a one-line toast, Esc that keeps your card). Afterwards it all went into a written finish bar and one script, page-gate.mjs: seven viewports, overflow measured in pixels, page errors, failed requests. Tap any image to enlarge it and read the exact prompt that drew it.

LOOP STOPPED ITSELF 2026-09-22 01:34 · ON MAIN SINCE 2026-09-30

The gate tells you it works. Only the loop tells you it is finished.

A plan lists features. Polish only comes from looking at the built thing, once per round, with one job: find the worst thing on screen. So that is now a rule in my workspace: every page runs the seven-viewport gate before anyone calls it done, and anything I use daily gets two or three polish rounds after the gate. Two hours of machine time at night is cheaper than one visitor holding a phone on its side.

A bold card with two lines: the gate says it works, in grey, and only the loop says it is finished, in teal, above a row of seventeen small commit dots ending in three hollow ones
AIOS-redesign-3merged · v02 1aebc03 → v18 227368f · rebase 91367ab · finish bar 2026-09-22 · 2026-09-21/30
06 Steal this

Steal the finish bar

The gate script and the loop live in my private workspace. What travels is the pattern: seven viewports with overflow measured in pixels, a list of states to shoot beyond the plain load, a hard speed budget, three to five fixes a round with every screenshot read, and a loop that stops after three empty rounds. Everything in this episode is free and open — clone it, run it, make it yours.

skill page-finish-bar.md: widths, states, contracts cmd page-gate.mjs: 7 viewports, overflow, errors, requests template polish loop prompt: 3 to 5 fixes, read every shot, stop after 3 empty assets v02 → v18 changelog as a worked example
run it # repo soon. the pattern: build → gate at 7 widths → read every shot → fix 3-5 → commit vNN → repeat → stop after 3 empty rounds

No GitHub? Comment POLISH on the post and the bot DMs you the link.

Next episode

The loop ran off one ticket. Next: the board it came from.

Every round here was tied to ticket itm_1c50f271, and the merge still waited for me. Next: the ticket board my agents work from, a flow page that counts tickets per stage per day and splits worker from human, and a Retry merge button for the moment a branch goes stale.

A darkened teaser card from the ticket board series
drops soon · follow so you don’t miss it