reconcile-paper-sync
millionco/skills · Agent Skill
Controlla e ripara l'implementazione Paper-to-code rispetto al suo rendering canonico paper-sync, inclusi caratteri, token precisi e confronto degli screenshot.
Vai all'installazione ↓Panoramica
| Autore | |
|---|---|
| Repository | millionco/skills ↗ |
| Cartella dello skill | reconcile-paper-sync |
| Stelle | 34 |
| Fork | 2 |
| File inclusi | 1 |
| Licenza | MIT |
| Aggiornato | 30 giugno 2026 |
| Link | Sorgente dello skill ↗ · Repository ↗ |
Verifica Mozgi
Cosa abbiamo trovato nei file dello skill. I dettagli con l’abbonamento.
Installazione
Uno skill è una cartella con un file SKILL.md. Scegli il tuo agente ed esegui i comandi in un terminale (macOS, Linux o Git Bash su Windows) dalla radice del progetto.
Incolla il prompt in Claude, Codex o un altro assistente: controllerà lo skill e lo installerà.
Install the “reconcile-paper-sync” skill from mozgi.io (https://mozgi.io/skills/millionco-reconcile-paper-sync/) for me. 1. Download the archive: https://mozgi.io/skills/millionco-reconcile-paper-sync/download If it fails, take the skill folder from the source: https://github.com/millionco/skills/tree/main/skills/reconcile-paper-sync 2. Read SKILL.md and the other files. Check the skill is safe: no hidden commands, no sending data elsewhere, no scripts doing something other than described. If anything looks suspicious, stop and tell me. 3. Put the “reconcile-paper-sync” folder into your agent's skills folder: .claude/skills/ for Claude Code, .agents/skills/ for Codex, .cursor/skills/ for Cursor, .gemini/skills/ for Gemini CLI. Inside the project if we are in one, otherwise in the home folder (~/.claude/skills/ etc.). 4. Make sure <skills folder>/reconcile-paper-sync/SKILL.md exists. 5. Briefly tell me what this skill does and offer to try it on my task. If you cannot do a step yourself, tell me what to click or type.
git clone --depth 1 --branch main https://github.com/millionco/skills.git /tmp/skills mkdir -p .claude/skills cp -r /tmp/skills/skills/reconcile-paper-sync .claude/skills/reconcile-paper-sync
Riavvia Claude Code: lo skill compare nell'elenco; per richiamarlo a mano usa /reconcile-paper-sync. Per tutti i progetti: ~/.claude/skills.
git clone --depth 1 --branch main https://github.com/millionco/skills.git /tmp/skills mkdir -p .agents/skills cp -r /tmp/skills/skills/reconcile-paper-sync .agents/skills/reconcile-paper-sync
Codex cerca gli skill in .agents/skills dalla cartella corrente fino alla radice del repo; richiamalo con /skills o $reconcile-paper-sync. Per tutti i progetti: ~/.agents/skills.
git clone --depth 1 --branch main https://github.com/millionco/skills.git /tmp/skills mkdir -p .cursor/skills cp -r /tmp/skills/skills/reconcile-paper-sync .cursor/skills/reconcile-paper-sync
Cursor carica lo skill all'avvio; richiamalo con “/” nella chat Agent. Cursor legge anche .claude/skills e .agents/skills. Per tutti i progetti: ~/.cursor/skills.
git clone --depth 1 --branch main https://github.com/millionco/skills.git /tmp/skills mkdir -p .gemini/skills cp -r /tmp/skills/skills/reconcile-paper-sync .gemini/skills/reconcile-paper-sync
Gemini CLI propone di attivare lo skill quando il compito corrisponde: tu confermi. Per tutti i progetti: ~/.gemini/skills.
Leggi SKILL.md e i file correlati prima di installare. Clicca un file per l'anteprima.
Contenuto di SKILL.md
Testo originale dell'autore · sola letturaReconcile a Built App with paper-sync
You make a running app match the Paper design it was built from, using the project's paper-sync output as the canonical reference render.
This is the verification-and-repair phase that follows paper-to-code-components. The conversion produces maintainable code; this skill confirms the result still looks like the design and closes the gaps where it drifted.
When this applies
- A Paper-derived app (e.g. a Next.js site) renders close to the design but "slightly off."
- A
paper-syncproject exists (a.paper-sync/<fileId>/directory and usually a dev server such aslocalhost:6006). - You want to verify a fresh Paper-to-code implementation before calling it done.
If there is no paper-sync project, fall back to the Paper screenshot as the oracle and apply the same method.
RED/GREEN Intent
This skill prevents these reconciliation failures:
- Declaring parity from reading code instead of comparing rendered pixels
- Leaving the app on
system-ui/Geist/Georgia fallbacks while the design uses real licensed fonts - Collapsing the design's several typefaces into one font token
- "Fixing" a difference that is actually newer design work in paper-sync by inventing data to match it
- Dropping small decorations (code-chip backgrounds, hairlines, dashed outlines, badges) during conversion and never noticing
GREEN outcome: the running app, screenshotted at the frame's width, matches the paper-sync render on typeface, type scale, color, spacing, and per-element decoration — and any remaining differences are explicitly attributed to newer design iterations, not conversion bugs.
1. Find the paper-sync canonical
paper-sync mirrors the Paper file into the repo. For each Paper file it keeps:
1. .paper-sync/<fileId>/.canonical/*.tsx — the verbatim Paper export per frame, the source of truth for exact values. Frame files are named page-1--frame-<id>-0.tsx and carry a header comment with the Paper URL (.../<FILE>/1-0/<FRAME>).
2. .paper-sync/<fileId>/paper.css — @font-face declarations and the design's CSS.
3. .paper-sync/<fileId>/public/paper-assets/fonts/*.otf — the real font files.
4. A dev server with a preview route: ?preview#<FRAME> renders a single frame alone (the frame id is the uppercase hash, e.g. N66-0).
Map the frame you are reconciling to its canonical file by matching visible text (grep -l "We introduce React Bench" .canonical/*.tsx). The on-disk .canonical can lag the running server; trust the live ?preview#<frame> render for what the user sees, and the canonical .tsx for exact numeric values.
2. Fonts first — the biggest lever
Substituted fonts are the most common cause of "slightly off." Do this before chasing spacing.
1. List the families the canonical frame uses: grep -ohE "font-\['[^']+'" <frame>.tsx | sort | uniq -c | sort -rn. Expect several distinct families by role (body, serif display, UI/label, numeric/mono, code).
2. For each family, confirm the app self-hosts the real font via @font-face and maps it to its own token. If it instead falls back to Geist/Georgia/system, that is the bug.
3. Port any missing fonts from public/paper-assets/fonts/*.otf into the app (e.g. public/fonts/), add @font-face per weight/style, and point each token at the real family with the system stack only as fallback. Keep distinct typefaces as distinct tokens — do not collapse them.
See paper-to-code-components §8 for the porting rules.
3. Extract exact values from the canonical
The .canonical/*.tsx holds the real numbers. For any element that looks off, read its canonical class list and match:
1. Type: text-[15px]/6 is 15px / line-height 24px; replicate size AND line-height, not just size.
2. Color: color(display-p3 ...) and hex — keep the exact gamut/value.
3. Decoration: backgrounds, borders, dashed outlines, underlines, radii, small markers.
Do not read sizes or colors off a screenshot when the canonical source is available.
4. Screenshot-diff the running app against the frame
1. Render both at the same width. A headless browser screenshot works without installing Playwright: drive a cached Chromium, e.g. chrome-headless-shell --hide-scrollbars --force-device-scale-factor=2 --virtual-time-budget=6000 --window-size=<w>,<h> --screenshot=out.png <url>. Allow time for @font-face to load before the shot.
2. Capture the app (localhost:3000) and the frame (localhost:6006/?preview#<FRAME>), then compare side by side.
3. Reconcile element stacking order (e.g. nav above vs below a header band), content width and text measure (same paragraph → same line count?), typeface, type scale, color, spacing, and every decoration. A flow/responsive layout will not be pixel-identical to absolute Paper coordinates — judge structural, typographic, and decorative fidelity, not coordinate equality.
5. Separate conversion drift from version drift
A difference has two possible causes; treat them differently:
- Conversion drift — the app dropped or substituted something the source frame has (a font, a code-chip background, a hairline). Fix it to match.
- Version drift — the paper-sync frame is a newer design than the export the app was built from (the frame header date / frame id differs; new elements like error bars or bullets appear that need data the app does not have). Do NOT invent data to fake it. Flag it to the user and let them decide whether to re-sync the app to the newer frame.
State which differences you fixed and which you attributed to a newer frame.
Escape Hatches Closed
| You Think | Do This Instead |
|---|---|
| "The code matches the design, so it's fine." | Screenshot the build against the paper-sync frame and compare pixels. |
| "The font has a system-ui fallback, that's close enough." | The fallback is not the design. Self-host the real .otf from public/paper-assets/fonts and @font-face it. |
| "I'll match the new error bars / bullets in the frame." | Check the frame date/id first. If it's newer than the built export, that's version drift — flag it, don't invent data. |
| "I'll eyeball the sizes from the screenshot." | Read exact text-[..]/.. and color(display-p3 ..) values from the .canonical/*.tsx. |
| "No Playwright installed, so I can't screenshot." | Drive a cached chrome-headless-shell with --screenshot; no install needed. |
Before Marking Complete, You MUST:
1. Identify the canonical frame (file id + frame hash) the app section corresponds to
2. Confirm every typeface the frame uses is self-hosted via @font-face and mapped to its own token — no design text left on a fallback font
3. Match exact type size + line-height, colors, and decorations from the .canonical/*.tsx, not from a screenshot
4. Screenshot the running app against ?preview#<frame> at matched width and reconcile the differences
5. Classify each remaining difference as conversion drift (fixed) or version drift (flagged to the user)
6. Re-run the app's lint/build after changes and report the screenshot comparison
Domande frequenti
Che cosa fa lo skill reconcile-paper-sync?
Controlla e ripara l'implementazione Paper-to-code rispetto al suo rendering canonico paper-sync, inclusi caratteri, token precisi e confronto degli screenshot.
Come si installa reconcile-paper-sync?
Apri il blocco “Installazione” in questa pagina e scegli il tuo agente — Claude Code, Codex, Cursor o Gemini CLI — per i comandi pronti che copiano la cartella dello skill nel posto giusto. Oppure copia il prompt dalla scheda “Con l'assistente” e lascia che sia l'assistente a controllare e installare lo skill.
Con quali modelli AI funziona reconcile-paper-sync?
Gli skill sono semplici istruzioni in Markdown (formato Agent Skills), quindi reconcile-paper-sync funziona con qualsiasi agente che legge SKILL.md: Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode e altri. Il modello carica le istruzioni solo quando il compito corrisponde alla descrizione dello skill.
reconcile-paper-sync è gratuito?
Lo skill è pubblicato apertamente da millionco con licenza MIT. Lo skill in sé non costa nulla: paghi solo il tuo fornitore AI per i token usati.
È sicuro installare gli skill?
Uno skill è un insieme di istruzioni che il tuo agente seguirà: leggi SKILL.md e i file correlati prima di installarlo. Gli script nella cartella dello skill vengono eseguiti solo se li esegue l'agente: controlla cosa fa e non concedere permessi superflui.
In cosa differisce uno skill da un server MCP o da un plugin?
Un server MCP o un plugin danno al modello nuovi strumenti: codice che viene eseguito e restituisce un risultato. Uno skill fornisce conoscenza e metodo: come affrontare il compito, quali passi seguire e com'è un buon risultato. Non servono server né chiavi API.