@rokkit/cli 1.1.19 → 1.3.0

package/dist/src/doctor.d.ts

@@ -20,25 +20,6 @@ export function runChecks(fs: {
     fix: string;
     autoFix?: string;
 }>;
-/**
- * Scan files for legacy z-scale utilities and suggest named-token equivalents.
- * Advisory only — z-scale still resolves via the preset's back-compat layer.
- * @param {Array<{ path: string, content: string }>} files
- * @returns {{ count: number, byFile: Array<{ path: string, hits: Array<{ token: string, suggestion: string }> }> }}
- */
-export function findLegacyZUtilities(files: Array<{
-    path: string;
-    content: string;
-}>): {
-    count: number;
-    byFile: Array<{
-        path: string;
-        hits: Array<{
-            token: string;
-            suggestion: string;
-        }>;
-    }>;
-};
 /**
  * Validate a parsed rokkit config for the named-token system. All advisory (warn).
  * @param {Record<string, unknown> | null} config

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@rokkit/cli",
-  "version": "1.1.19",
+  "version": "1.3.0",
   "description": "Command line utilities for icon bundling and conversion.",
   "repository": {
     "type": "git",
@@ -27,7 +27,7 @@
   },
   "dependencies": {
     "@mdi/svg": "^7.4.47",
-    "@rokkit/unocss": "1.1.19",
+    "@rokkit/unocss": "1.3.0",
     "sade": "^1.8.1",
     "@iconify/tools": "5.0.3",
     "@iconify/types": "^2.0.0",

package/skills/semantic-styles-rokkit/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: semantic-styles-rokkit
-description: Use when building, styling, or auditing a Rokkit-powered app (Svelte 5 + UnoCSS presetRokkit) — covers the named-token vocabulary (paper / ink / primary / on-primary / *-soft), EXTENDING it with custom tokens via `overrides:` and the full palette ladder via `tokens: 'extended'`, the rokkit.config.js palette→skin→tokens pipeline, dual-palette dark mode, migrating legacy z-scale utilities (bg-surface-z0, text-primary-z5) to named tokens, and the rokkit init / rokkit doctor CLI.
+description: Use when building, styling, or auditing a Rokkit-powered app (Svelte 5 + UnoCSS presetRokkit) — covers the named-token vocabulary (paper / ink / primary / on-primary / *-soft), EXTENDING it with custom tokens via `overrides:` and the full palette ladder via `tokens: 'extended'`, the rokkit.config.js palette→skin→tokens pipeline, dual-palette dark mode, and the rokkit init / rokkit doctor CLI.
 ---
 
 # Semantic Styles — Rokkit
@@ -17,7 +17,6 @@ rokkit.config.js
   UnoCSS preset (presetRokkit)  →  CSS custom properties + utility classes
        ↓
   Components use named-token utilities:  bg-paper · text-ink · bg-primary · text-on-primary
-  (legacy z-scale utilities — bg-surface-z1 — still resolve via back-compat aliases)
 ```
 
 The pipeline has three layers. Most authoring happens in **Layer 3 (named tokens)**.
@@ -56,7 +55,7 @@ The preset wraps values into complete colors when it emits CSS vars (see Layer 3
 **What:** A skin maps semantic role names to palette names. The core roles that back the
 named-token vocabulary are: `surface`, `ink`, `primary`, `accent`, `success`, `warning`,
 `danger`, `error`, `info`. (`secondary`/`tertiary` are accepted but no named token reads
-them — they only matter if you also use `extended` z-scale utilities for those roles.)
+them — they only matter if you opt into `tokens: 'extended'` for those roles.)
 
 **Why:** Decoupling role from palette lets you swap the whole scheme in config, not code.
 `bg-paper` doesn't know whether `surface` is `kami`, `slate`, or `zinc`.
@@ -84,7 +83,7 @@ skins: {
 
 ### `tokens:` — which CSS the preset emits
 ```js
-tokens: 'core'      // default — emit the 24 named tokens (palette values inlined) + z-aliases
+tokens: 'core'      // default — emit the 24 named tokens (palette values inlined)
 tokens: 'extended'  // also emit the full 11-shade ladder (--color-{role}-{shade}) per role
 ```
 `'core'` is the default and what apps should use. Reach for `'extended'` only when you need
@@ -226,16 +225,7 @@ directly in CSS-only contexts; for alpha, use `color-mix` (the form the preset i
 .thing      { background: var(--paper); border-color: var(--paper-edge); }
 .thing-50pc { background: color-mix(in oklch, var(--primary) 50%, transparent); }
 ```
-Prefer utility classes over raw vars where you can. (In `core` mode the legacy
-`--color-{role}-z{n}` vars are aliases pointing at named tokens, so they too resolve to
-complete colors — the old "wrap bare components in `rgb()`/`oklch()`" rule no longer
-applies in core mode.)
-
-### Legacy z-scale utilities (back-compat)
-
-The older z-scale utilities (`bg-surface-z0`, `text-surface-z9`, `text-primary-z5`, …)
-**still resolve** as a back-compat layer that collapses onto the named tokens above —
-existing code keeps working. New code should use named tokens. See the migration section.
+Prefer utility classes over raw vars where you can.
 
 ---
 
@@ -305,38 +295,6 @@ Default stays `'core'`.
 
 ---
 
-## Migrating from z-scale to named tokens
-
-`rokkit doctor` scans `src/**` and prints the suggested replacement per occurrence
-(advisory — it never rewrites your code). The mapping it uses:
-
-**Surface role** (`{prefix}-surface-z{n}` → `{prefix}-{token}`):
-
-| z | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
-|---|---|---|---|---|---|---|---|---|---|---|----|
-| token | `paper` | `paper-soft` | `paper-mute` | `paper-mute` | `paper-edge` | `ink-soft` | `ink-soft` | `ink-mute` | `ink-mute` | `ink` | `ink` |
-
-**Ink role** (inverted scale):
-
-| z | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
-|---|---|---|---|---|---|---|---|---|---|---|----|
-| token | `ink` | `ink-mute` | `ink-mute` | `ink-soft` | `ink-soft` | `paper-edge` | `paper-edge` | `paper-mute` | `paper-mute` | `paper-soft` | `paper` |
-
-**Accent / status roles** (`primary`, `accent`, `success`, `warning`, `danger`, `error`,
-`info`): `z0–z2 → {role}-soft` (tinted background), `z3+ → {role}` (solid). Exception:
-`primary` has **no** `-soft` companion — `primary-z*` all map to `primary`; for a tinted
-primary background use `accent-soft` or a custom token.
-
-Common rewrites:
-```
-bg-surface-z0   → bg-paper            text-surface-z9 → text-ink
-bg-surface-z1   → bg-paper-soft       text-surface-z7 → text-ink-mute
-border-surface-z3 → border-paper-edge text-surface-z5 → text-ink-soft
-text-primary-z5 → text-primary        bg-success-z1   → bg-success-soft
-```
-
----
-
 ## Rokkit CLI — `init` and `doctor`
 
 **`rokkit init`** scaffolds `rokkit.config.js` in the named-token shape (a header comment
@@ -354,7 +312,6 @@ fail the command or block `--fix`:
 | `skin-ink-role` | the active colormap has no `ink` role (text tokens fall back) |
 | `oklch-needs-palettes` | `colorSpace: 'oklch'` but no `palettes` block |
 | `colors-alias` | config uses the legacy `colors:` instead of `skin:` |
-| legacy z-utilities | `src/**` contains `bg-surface-z*` / `text-primary-z*` etc. (prints named-token suggestions) |
 
 The "active colormap" is resolved the same way the preset does: `skins.default ?? skin ??
 colors`. Run `rokkit doctor --fix` to auto-write a real starter config when one is missing.
@@ -447,7 +404,6 @@ export default {
 | "core only emits `accent`/`accent-soft`, so I can't have `accent-2`/`on-accent`" | the vocabulary is extensible, not fixed | add them under `overrides:` (or use `tokens: 'extended'`) — see Layer 3b |
 | Override palette ref `'sky-600'` (hyphen) | refs use **dot** notation | `'sky.600'` (and `sky` must be in `palettes:`) |
 | `color: #3D3730` in a component | breaks on theme change | `text-ink` (or the right named token) |
-| Using `bg-surface-z5` / `text-primary-z5` in new code | legacy back-compat layer | named tokens — see migration table (or run `rokkit doctor`) |
 | Single-palette `surface` but expecting dark mode | no `[data-mode]` block is emitted | make `surface`/`ink` dual-palette `{ light, dark }` |
 | Skin with no `ink` role | `ink-*` text tokens fall back to surface | add `ink: '<palette>'` (reuse surface is fine) |
 | Using `skin:` and `skins:` together | `skins:` silently wins | pick one |

package/src/doctor.js

@@ -1,6 +1,6 @@
 /* eslint-disable no-console */
-import { existsSync, readFileSync, writeFileSync, readdirSync } from 'fs'
-import { resolve, extname } from 'path'
+import { existsSync, readFileSync, writeFileSync } from 'fs'
+import { resolve } from 'path'
 import { loadConfig as loadParsedConfig } from './config.js'
 import {
 	generateUnoConfig,
@@ -363,85 +363,6 @@ function handleResults(checks, cwd, failures, fix) {
 	}
 }
 
-// z-slot → named token (matches @rokkit/core Z_COLLAPSE maps).
-const SURFACE_Z = {
-	z0: 'paper', z1: 'paper-soft', z2: 'paper-mute', z3: 'paper-mute', z4: 'paper-edge',
-	z5: 'ink-soft', z6: 'ink-soft', z7: 'ink-mute', z8: 'ink-mute', z9: 'ink', z10: 'ink'
-}
-const INK_Z = {
-	z0: 'ink', z1: 'ink-mute', z2: 'ink-mute', z3: 'ink-soft', z4: 'ink-soft',
-	z5: 'paper-edge', z6: 'paper-edge', z7: 'paper-mute', z8: 'paper-mute', z9: 'paper-soft', z10: 'paper'
-}
-
-/**
- * Map a z-scale role+slot to its nearest named token.
- * @param {string} role
- * @param {string} z — e.g. 'z5'
- * @returns {string}
- */
-function suggestNamedToken(role, z) {
-	if (role === 'surface') return SURFACE_Z[z]
-	if (role === 'ink') return INK_Z[z]
-	// accent + status roles collapse z0–z2 to their `-soft` tint (matches
-	// @rokkit/core theme.ts getZAliasesOther: zNum <= 2). primary has no `-soft`.
-	if (role !== 'primary' && Number(z.slice(1)) <= 2) return `${role}-soft`
-	return role
-}
-
-const Z_UTILITY_RE =
-	/\b(bg|text|border(?:-[tblr])?|fill|stroke|ring|outline|divide)-(surface|ink|primary|accent|success|warning|danger|error|info)-(z(?:10|[0-9]))\b/g
-
-/**
- * Scan files for legacy z-scale utilities and suggest named-token equivalents.
- * Advisory only — z-scale still resolves via the preset's back-compat layer.
- * @param {Array<{ path: string, content: string }>} files
- * @returns {{ count: number, byFile: Array<{ path: string, hits: Array<{ token: string, suggestion: string }> }> }}
- */
-export function findLegacyZUtilities(files) {
-	const byFile = []
-	let count = 0
-	for (const { path: filePath, content } of files) {
-		const hits = []
-		for (const match of content.matchAll(Z_UTILITY_RE)) {
-			const [token, prefix, role, z] = match
-			hits.push({ token, suggestion: `${prefix}-${suggestNamedToken(role, z)}` })
-		}
-		if (hits.length > 0) {
-			byFile.push({ path: filePath, hits })
-			count += hits.length
-		}
-	}
-	return { count, byFile }
-}
-
-const SCAN_EXTS = new Set(['.svelte', '.css', '.ts', '.js'])
-
-/**
- * Recursively gather scannable source files under cwd/src.
- * Skips node_modules, symlinks (avoids recursion loops), and uno.config.js
- * (its safelist legitimately lists z-utilities).
- * @param {string} cwd
- * @returns {Array<{ path: string, content: string }>}
- */
-function gatherSrcFiles(cwd) {
-	const root = resolve(cwd, 'src')
-	if (!existsSync(root)) return []
-	const out = []
-	const walk = (dir) => {
-		for (const dirent of readdirSync(dir, { withFileTypes: true })) {
-			if (dirent.isSymbolicLink() || dirent.name === 'node_modules') continue
-			const full = resolve(dir, dirent.name)
-			if (dirent.isDirectory()) {
-				walk(full)
-			} else if (dirent.isFile() && SCAN_EXTS.has(extname(dirent.name)) && !full.endsWith('uno.config.js')) {
-				out.push({ path: full.slice(cwd.length + 1), content: readFileSync(full, 'utf-8') })
-			}
-		}
-	}
-	walk(root)
-	return out
-}
-
 /**
  * Validate a parsed rokkit config for the named-token system. All advisory (warn).
  * @param {Record<string, unknown> | null} config
@@ -505,17 +426,6 @@ export async function doctor(opts = {}) {
 		}
 	}
 
-	const { count, byFile } = findLegacyZUtilities(gatherSrcFiles(cwd))
-	if (count > 0) {
-		console.info(`\nLegacy z-scale utilities (advisory — ${count} in ${byFile.length} file(s)):`)
-		for (const { path: filePath, hits } of byFile.slice(0, 10)) {
-			const examples = [...new Map(hits.map((h) => [h.token, h.suggestion]))]
-			const shown = examples.slice(0, 3).map(([t, s]) => `${t} → ${s}`).join(', ')
-			console.info(`  ${filePath}: ${shown}${examples.length > 3 ? ', …' : ''}`)
-		}
-		console.info('        Named tokens are preferred; z-scale still works via back-compat.')
-	}
-
 	handleResults(checks, cwd, failures, opts.fix ?? false)
 	console.info('')
 }

package/src/init.js

@@ -20,9 +20,7 @@ const NAMED_TOKEN_HEADER = `/**
  *   Accent   bg-primary · text-on-primary · bg-accent · bg-accent-soft
  *   Status   bg-success-soft · text-success   (+ warning / danger / error / info)
  *
- * Tokens flip automatically under [data-mode="dark"]. The older z-scale
- * utilities (bg-surface-z0, text-primary-z5) still resolve for back-compat,
- * but new code should prefer the named tokens above.
+ * Tokens flip automatically under [data-mode="dark"].
  *
  * tokens: 'core' emits the named vocabulary; 'extended' also emits the full
  * 11-shade palette ladder per role (for charts / data-viz).