Skip to main content

create-styleguide

Per-surface visual styleguide lifecycle — capture, annotate with token callouts, audit via /review-design, track token debt, and detect visual drift between rounds. Composes /create-design-board and /review-design into a repeatable pipeline.

Quellinformationen

Repository
grahama1970/agent-stack-public
Letzte Quellaktivität
24. September 2026 um 15:51
Erkannte Sprache von SKILL.md
Englisch
Sterne
0
Forks
0

Installationsoptionen

Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.

Quelldateien prüfen

Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.

Datei-Explorer
9 Dateien

SKILL.md wird angezeigt

SKILL.md
Quellanweisungen · Schreibgeschützte Vorschau
name
create-styleguide
version
1.0.0
description
Per-surface visual styleguide lifecycle — capture, annotate with token callouts, audit via /review-design, track token debt, and detect visual drift between rounds. Composes /create-design-board and /review-design into a repeatable pipeline.
provides
["styleguide","visual-audit","token-debt-tracking"]
composes
["create-design-board","review-design","memory","agentic-evals"]
triggers
["create styleguide","build styleguide","visual styleguide"]
allowed-tools
["Bash","Read","Write"]
taxonomy
["design","compliance","visual"]
disciplines
["ui-design-engineering"]
> STOP. READ THIS ENTIRE SKILL.MD BEFORE CALLING ANY ENDPOINT. # /create-styleguide Per-surface visual styleguide lifecycle for Embry OS's 10 UX surfaces. ## Commands - `build` — Full pipeline: annotate -> audit -> debt -> assemble STYLEGUIDE.md - `annotate` — Just the Pillow annotation step (token callouts on screenshots) - `audit` — Run /review-design on a surface - `diff` — Visual regression between rounds (pixel diff + block-SSIM) - `status` — Show which surfaces have styleguides, debt counts, last audit dates - `dry-run` — Generate sample output without LLM calls (used by sanity.sh) ## Usage ```bash ./run.sh build --surface S4 --skip-audit ./run.sh annotate --screenshots ./shots/ --tokens ./tokens.json --output ./annotated/ ./run.sh diff --surface S4 --before ./round1/ --after ./round2/ ./run.sh status ```
Auf GitHub ansehen