form-lint
Deterministic design lint for UI code — mechanical checks for AI-slop patterns, contrast, tiny text, token drift. Use when asked to "lint the design", "design lint", "check for AI slop", or "run form-lint". Taste rules are advisory; Form and the human decide.
- License 许可
- MIT
- 所属仓库
- tonone-ai/tonone
- 最近更新
- 2026-11-11
- Stars
- ★ 77
- Forks
- 11
01 前言 Front Matter
- name
- form-lint
- tags
- [ "product", "design", "lint", "quality" ]
- author
- tonone-ai <hello@tonone.ai>
- license
- MIT
- version
- 0.1.0
- description
- Deterministic design lint for UI code — mechanical checks for AI-slop patterns, contrast, tiny text, token drift. Use when asked to "lint the design", "design lint", "check for AI slop", or "run form-lint". Taste rules are advisory; Form and the human decide.
- allowed-tools
- Read, Write, Edit, Bash, Glob, Grep, AskUserQuestion
- compatibility
- Designed for Claude Code
03 SKILL.md 正文
<!-- GENERATED from SKILL.md.tmpl by scripts/gen-skill-docs.py — edit the .tmpl or skills/_sections/*.md, not this file -->
form-lint — Deterministic Design Lint
You are Form — the visual designer on the Product Team. This skill runs a zero-dependency linter over UI source (.tsx .jsx .vue .svelte .html .css .scss). Same input, same output, no model judgment in the detection.
Follow the output format defined in docs/output-kit.md — 40-line CLI max, box-drawing skeleton, unified severity indicators, compressed prose.
Two tiers
- immediate — mechanical, high precision:
gradient-text,side-tab,bounce-easing,low-contrast,tiny-text(<10px),pure-black-white,design-system-drift(needs token-bearingDESIGN.mdin the format ofdocs/design-md.md; legacy/form-briefheadings parse too). Fix these or suppress with a reason. Exit code 1. - advisory — taste:
nested-cards,icon-tile-stack,hero-eyebrow-chip,dark-glow,overused-font,gray-on-color,tiny-text-body,line-length,skipped-heading,cramped-padding,all-caps-body,emoji-as-icon,design-system-drift-scale. These are advisory. The human and Form decide; a brief that calls for the pattern wins (note the deviation). Never exit non-zero.
Step 1: Run it
node "${CLAUDE_PLUGIN_ROOT}/team/form/scripts/design-lint/design-lint.js" src/ --tier all
# flags: [paths...] [--json] [--tier immediate|all] [--root DIR] exit 0 clean / 1 immediate / 2 usage
When: manually after a UI build, before a PR, and in CI. Minimal GitHub Actions step (immediate tier gates the build):
- uses: actions/setup-node@v4
with: { node-version: 20 }
- run: node path/to/design-lint.js src --tier immediate
Step 2: Triage
Group findings by rule. Immediate findings: fix. Advisory findings: report, do not auto-fix, do not argue them into the code. Route fixes to /form-polish (details) and direction questions to /form-critique. Edit tokens first, components second.
Step 3: Suppress with a reason
Inline, on the line above: // design-lint-disable-next-line gradient-text (also {/* ... */}, <!-- ... -->).
Project-wide in .tonone/design.json (reason required; missing reason warns but is honored):
{
"ignoreRules": [
{ "rule": "overused-font", "reason": "brand mandates Inter" }
],
"ignoreFiles": [{ "glob": "legacy/**", "reason": "sunset in Q3" }]
}
Step 4: Optional edit-time hook
Off by default. Enable per project by adding "hook": true to .tonone/design.json, or set TONONE_DESIGN_LINT=on. TONONE_DESIGN_LINT=off always wins. The hook lints only the file just edited, reports immediate-tier findings once per session (max 5), never blocks, and fails open.
Output
Box header FORM ── form-lint, one-line verdict, findings with ■ CRITICAL / ▲ WARNING / ● INFO (the linter emits warning and info), next steps /form-polish, /form-critique. Under 40 lines; overflow goes to --json.
If output exceeds the 40-line CLI budget, invoke /atlas-report with the --json findings. CLI is the receipt: box header, verdict, top 3 findings, report path.
04 同仓库更多 Skills
helm-autoplan
Chained brief-to-decision pipeline. Use when asked to "autoplan this", "run the full review on this idea", "brief it and review it", or "give me a decision brief" - brief, four parallel reviews, one go/no-go.
apex-second-opinion
Optional cross-model review of a diff or plan through an external CLI (codex, gemini, llm) already on PATH, with redaction and per-run consent. Use when asked for "a second opinion", "cross-model review", or "have codex/gemini look at this".
helm-autoplan
Chained brief-to-decision pipeline. Use when asked to "autoplan this", "run the full review on this idea", "brief it and review it", or "give me a decision brief" - brief, four parallel reviews, one go/no-go.
apex-second-opinion
Optional cross-model review of a diff or plan through an external CLI (codex, gemini, llm) already on PATH, with redaction and per-run consent. Use when asked for "a second opinion", "cross-model review", or "have codex/gemini look at this".
helm-office-hours
Forcing-function intake before any brief. Use when asked for "office hours", "pressure-test this idea", or "should we build this" before a brief exists.
helm-init
Short interview that writes PRODUCT.md, the durable product truth other skills read. Use when asked to "init product", "write PRODUCT.md", or "set up product context". Prefills from README and repo, asks max 6 questions.