@dev-workflow/skill 0.1.0

package/skills/execution/figmaSync/scripts/figma-sync-report.mjs

@@ -0,0 +1,448 @@
+/**
+ * Figma Sync Report
+ *
+ * 把本次 figmaSync 任务的「AI 做了什么」与「产物长什么样」交叉量化输出。
+ *
+ * 用法:
+ *   pnpm figma:report                                # 默认 plan 命令、写到 .agent/skills/figmaSync/
+ *   pnpm figma:report --command=plan|apply           # 报告文件名前缀：session-report-<command>.md
+ *   pnpm figma:report --target-dir=src/foo           # 同时把报告复制一份到指定目录
+ *   pnpm figma:report --plan=src/foo/PLAN.md         # 解析该 PLAN.md §4 表格，统计 variables 来源
+ *   pnpm figma:report --reset                        # 清空 session log（任务开始时调）
+ *
+ * 主指标（新版排序）：
+ *   1. 硬编码计数（产物侧）—— 需要解释
+ *   2. CSS variables 命中数（解析 PLAN §4 表格）
+ *   3. 脚本调用率 lookupVar / matchToken —— 辅助参考
+ *
+ * 告警条件：
+ *   - 硬编码 > 0 → ⚠️
+ *   - Fast Path 覆盖率 < 50% 且 PLAN.md §4 无 variables 来源 → ⚠️
+ */
+import { execSync } from 'node:child_process'
+import fs from 'node:fs'
+import path from 'node:path'
+
+import { readSessionLog, resetSessionLog } from './_shared/session-log.mjs'
+
+const DEFAULT_REPORT_DIR = '.agent/skills/figmaSync'
+const FAST_PATH_LOW_THRESHOLD = 0.5
+
+function parseArgs(argv) {
+  const out = { reset: false, command: 'plan', targetDir: null, plan: null }
+  for (const arg of argv.slice(2)) {
+    if (arg === '--reset') out.reset = true
+    else if (arg.startsWith('--command=')) out.command = arg.slice('--command='.length)
+    else if (arg.startsWith('--target-dir=')) out.targetDir = arg.slice('--target-dir='.length)
+    else if (arg.startsWith('--plan=')) out.plan = arg.slice('--plan='.length)
+  }
+  if (!['plan', 'apply'].includes(out.command)) out.command = 'plan'
+  return out
+}
+
+function getChangedFiles() {
+  try {
+    const out = execSync('git status --porcelain', { encoding: 'utf-8' })
+    return out
+      .split('\n')
+      .filter(Boolean)
+      .map((line) => ({
+        status: line.slice(0, 2).trim(),
+        file: line.slice(3).replace(/^"|"$/g, '')
+      }))
+      .filter((c) => /\.(ts|tsx|css|md)$/.test(c.file))
+      .filter((c) => fs.existsSync(c.file))
+  } catch {
+    return []
+  }
+}
+
+function auditFile(filePath) {
+  const source = fs.readFileSync(filePath, 'utf-8')
+  const apexImports = []
+  const apexImportRe = /import\s+\{([^}]+)\}\s+from\s+['"]@frontend\/apex-ui--react['"]/g
+  let m
+  while ((m = apexImportRe.exec(source)) !== null) {
+    const names = m[1]
+      .split(',')
+      .map((n) => n.replace(/\bas\s+\w+/, '').trim())
+      .filter(Boolean)
+    apexImports.push(...names)
+  }
+
+  const cssVariables = {}
+  const cssVarRe = /var\(\s*(--[a-zA-Z0-9_-]+)/g
+  while ((m = cssVarRe.exec(source)) !== null) {
+    cssVariables[m[1]] = (cssVariables[m[1]] ?? 0) + 1
+  }
+
+  const hardcodedRe = /(['"`])(#[0-9a-fA-F]{3,8})\1|\b(\d{2,4})px\b/g
+  const hardcoded = []
+  while ((m = hardcodedRe.exec(source)) !== null) {
+    const raw = m[2] ?? m[3] + 'px'
+    const before = source.lastIndexOf('\n', m.index)
+    const lineIdx = source.slice(0, m.index).split('\n').length
+    const snippet = source.slice(before + 1, source.indexOf('\n', m.index)).trim()
+    hardcoded.push({ value: raw, line: lineIdx, snippet })
+  }
+
+  const commonRefs = []
+  const commonImportRe = /from\s+['"]([^'"]*src\/common[^'"]*|[^'"]*\/common\/[^'"]*)['"]/g
+  while ((m = commonImportRe.exec(source)) !== null) commonRefs.push(m[1])
+
+  const pandaForbidden = []
+  const forbiddenRe = /@pandacss|styled-system|css\(|cva\(|sva\(/g
+  while ((m = forbiddenRe.exec(source)) !== null) {
+    const lineIdx = source.slice(0, m.index).split('\n').length
+    pandaForbidden.push({ value: m[0], line: lineIdx })
+  }
+
+  return { apexImports, cssVariables, hardcoded, commonRefs, pandaForbidden }
+}
+
+function aggregate(files) {
+  const acc = {
+    apexImports: {},
+    cssVariables: {},
+    hardcoded: [],
+    commonRefs: {},
+    pandaForbidden: [],
+    files: []
+  }
+  for (const file of files) {
+    const r = auditFile(file)
+    if (
+      r.apexImports.length === 0 &&
+      Object.keys(r.cssVariables).length === 0 &&
+      r.hardcoded.length === 0 &&
+      r.commonRefs.length === 0 &&
+      r.pandaForbidden.length === 0
+    )
+      continue
+    acc.files.push(file)
+    for (const name of r.apexImports) acc.apexImports[name] = (acc.apexImports[name] ?? 0) + 1
+    for (const [k, v] of Object.entries(r.cssVariables))
+      acc.cssVariables[k] = (acc.cssVariables[k] ?? 0) + v
+    for (const h of r.hardcoded) acc.hardcoded.push({ ...h, file })
+    for (const ref of r.commonRefs) acc.commonRefs[ref] = (acc.commonRefs[ref] ?? 0) + 1
+    for (const hit of r.pandaForbidden) acc.pandaForbidden.push({ ...hit, file })
+  }
+  return acc
+}
+
+function auditForbiddenFiles(changedFiles) {
+  const forbidden = [
+    'src/routeTree.gen.ts',
+    'src/services/request.ts',
+    'src/shared/wujie-bridge.ts',
+    'src/stores/shell-store.ts'
+  ]
+  return changedFiles
+    .map((c) => c.file)
+    .filter((file) => forbidden.includes(file) || file.startsWith('src/common/'))
+}
+
+function tallyLog(log) {
+  const summary = {
+    lookupVar: { hit: 0, miss: 0, byLayer: {} },
+    matchToken: { hit: 0, miss: 0, byLayer: {} },
+    misses: []
+  }
+  for (const entry of log) {
+    const target =
+      entry.script === 'lookup-var'
+        ? summary.lookupVar
+        : entry.script === 'match-token'
+          ? summary.matchToken
+          : null
+    if (!target) continue
+    if (entry.hit) {
+      target.hit++
+      const layer = entry.layer ?? 'unknown'
+      target.byLayer[layer] = (target.byLayer[layer] ?? 0) + 1
+    } else {
+      target.miss++
+      summary.misses.push({ script: entry.script, input: entry.input })
+    }
+  }
+  return summary
+}
+
+/**
+ * 解析 PLAN.md §4 「CSS 变量映射表」，统计「命中方式」列里 variables 来源出现次数。
+ * 表格契约（PLAN.md.tpl 强制）：列顺序为 来源 | 类型 | CSS variable | /theme 路径 | 原始值 | 命中方式 | nodeId | designContextStatus。
+ * 命中方式取值约定：
+ *   - "Fast Path 0 (variables)" → MCP get_variable_defs 直接命中
+ *   - "Fast Path 1"              → lookup-var 命中
+ *   - "Slow Path"                → match-token 命中
+ *   - "未命中"                   → 降级
+ */
+function parsePlanTokenTable(planPath) {
+  if (!planPath || !fs.existsSync(planPath)) {
+    return { source: 'none', stats: null }
+  }
+  const text = fs.readFileSync(planPath, 'utf-8')
+  // 切 §4 整段：从 "## 4. CSS 变量映射表" 到下一个 "## " 之间（不含下一个标题）
+  const headRe = /^##\s+4\.\s+CSS\s+变量映射表\s*$/m
+  const headMatch = headRe.exec(text)
+  if (!headMatch) return { source: planPath, stats: null }
+  const start = headMatch.index + headMatch[0].length
+  const tail = text.slice(start)
+  const nextRe = /^##\s/m
+  const nextMatch = nextRe.exec(tail)
+  const section = nextMatch ? tail.slice(0, nextMatch.index) : tail
+
+  const lineRe = /^\|[^\n]+\|$/gm
+  const rows = section.match(lineRe) ?? []
+  const stats = {
+    'Fast Path 0 (variables)': 0,
+    'Fast Path 1 (lookup-var)': 0,
+    'Slow Path (value-match)': 0,
+    未命中: 0
+  }
+  for (const row of rows) {
+    if (/^\|\s*-+/.test(row)) continue
+    if (/来源(Figma|（Figma）)|命中方式|CSS variable/.test(row)) continue
+    if (/^\|\s*<|^\|\s*\(占位/.test(row)) continue
+    const cells = row
+      .split('|')
+      .slice(1, -1)
+      .map((c) => c.trim())
+    const hitWay = cells[5] ?? ''
+    for (const key of Object.keys(stats)) {
+      if (hitWay.includes(key)) stats[key]++
+    }
+  }
+  return { source: planPath, stats }
+}
+
+function renderReport({
+  log,
+  audit,
+  planTokens,
+  generatedAt,
+  command,
+  changedFiles,
+  forbiddenFiles
+}) {
+  const lines = []
+  lines.push(`# figmaSync Session Report (${command}) — ${generatedAt}`)
+  lines.push('')
+
+  lines.push('## Review 摘要')
+  lines.push('')
+  lines.push(`- 改动文件：${changedFiles.length} 个`)
+  lines.push(
+    `- 禁止 / 高风险文件命中：${forbiddenFiles.length} 个${forbiddenFiles.length === 0 ? ' ✅' : ' ⚠️'}`
+  )
+  lines.push(
+    `- Panda / styled-system 禁用内容：${audit.pandaForbidden.length} 处${audit.pandaForbidden.length === 0 ? ' ✅' : ' ⚠️'}`
+  )
+  lines.push(
+    `- 硬编码嫌疑：${audit.hardcoded.length} 处${audit.hardcoded.length === 0 ? ' ✅' : ' ⚠️'}`
+  )
+  lines.push('')
+
+  lines.push('## 改动文件')
+  lines.push('')
+  if (changedFiles.length === 0) {
+    lines.push('- 未检测到工作树改动')
+  } else {
+    for (const file of changedFiles) lines.push(`- ${file.status || '??'} ${file.file}`)
+  }
+  lines.push('')
+
+  lines.push('## 成功项 / 跳过项 / 风险项')
+  lines.push('')
+  const success = []
+  const skipped = []
+  const risks = []
+  if (Object.keys(audit.cssVariables).length > 0) success.push('使用 CSS variables')
+  else skipped.push('未检测到 CSS variable 引用')
+  if (Object.keys(audit.apexImports).length > 0) success.push('复用 Apex UI 组件')
+  else skipped.push('未检测到 Apex UI import')
+  if (Object.keys(audit.commonRefs).length > 0) success.push('复用 common module')
+  if (audit.hardcoded.length > 0) risks.push('存在硬编码嫌疑，需要解释或确认')
+  if (audit.pandaForbidden.length > 0) risks.push('存在 Panda / styled-system 禁用内容')
+  if (forbiddenFiles.length > 0) risks.push('命中禁止或高风险文件，需要 Human 再确认')
+  lines.push(`- 成功项：${success.length > 0 ? success.join('；') : '无自动识别项'}`)
+  lines.push(`- 跳过项：${skipped.length > 0 ? skipped.join('；') : '无'}`)
+  lines.push(`- 风险项：${risks.length > 0 ? risks.join('；') : '无'}`)
+  lines.push('')
+
+  if (forbiddenFiles.length > 0) {
+    lines.push('## 禁止 / 高风险文件')
+    lines.push('')
+    for (const file of forbiddenFiles) lines.push(`- ${file}`)
+    lines.push('')
+  }
+
+  // ---------- 主指标 ----------
+  lines.push('## 🎯 主指标')
+  lines.push('')
+  const hardcodedCount = audit.hardcoded.length
+  lines.push(`- 硬编码嫌疑：${hardcodedCount} ${hardcodedCount === 0 ? '✅' : '⚠️ 需解释'}`)
+
+  let variablesHits = 0
+  if (planTokens.stats) {
+    variablesHits = planTokens.stats['Fast Path 0 (variables)'] ?? 0
+    const total = Object.values(planTokens.stats).reduce((a, b) => a + b, 0)
+    lines.push(`- PLAN §4 CSS variable 命中分布（共 ${total} 行）：`)
+    for (const [key, n] of Object.entries(planTokens.stats)) {
+      lines.push(`    - ${key}: ${n}`)
+    }
+  } else if (planTokens.source !== 'none') {
+    lines.push(`- PLAN §4 CSS variable 表格解析失败（${planTokens.source}）`)
+  } else {
+    lines.push('- PLAN §4 CSS variable 表格：未指定 --plan 路径，跳过解析')
+  }
+
+  const fastTotal = log.lookupVar.hit + log.lookupVar.miss
+  const allLookupTotal = fastTotal + log.matchToken.hit + log.matchToken.miss
+  const fastCoverage = allLookupTotal > 0 ? log.lookupVar.hit / allLookupTotal : null
+  if (fastCoverage !== null) {
+    const pct = Math.round(fastCoverage * 100)
+    lines.push(`- 脚本调用率（辅助参考）：lookup-var ${pct}% / 总查询 ${allLookupTotal}`)
+    if (fastCoverage < FAST_PATH_LOW_THRESHOLD && variablesHits === 0) {
+      lines.push(
+        '    - ⚠️ 调用率 < 50% 且 PLAN §4 无 variables 来源；AI 既没查脚本也没标 variables 命中'
+      )
+    }
+  } else {
+    lines.push('- 脚本调用率（辅助参考）：本 session 未触发匹配脚本')
+  }
+  lines.push('')
+
+  // ---------- AI 行为 ----------
+  lines.push('## 🔎 AI 调用轨迹')
+  lines.push('')
+  const total = log.lookupVar.hit + log.lookupVar.miss + log.matchToken.hit + log.matchToken.miss
+  if (total === 0) {
+    lines.push('- 本 session 未调用 lookup-var / match-token')
+  } else {
+    lines.push(`- lookup-var 命中：${log.lookupVar.hit} / ${fastTotal}`)
+    for (const [layer, n] of Object.entries(log.lookupVar.byLayer)) {
+      lines.push(`    - ${layer}.* × ${n}`)
+    }
+    const slowTotal = log.matchToken.hit + log.matchToken.miss
+    if (slowTotal > 0) {
+      lines.push(`- match-token 命中：${log.matchToken.hit} / ${slowTotal}`)
+      for (const [layer, n] of Object.entries(log.matchToken.byLayer)) {
+        lines.push(`    - 推荐 ${layer} 层 × ${n}`)
+      }
+    } else {
+      lines.push('- match-token 本次未触发')
+    }
+    if (log.misses.length > 0) {
+      lines.push('- 未命中输入：')
+      for (const miss of log.misses) {
+        lines.push(`    - \`${miss.script}\`  ← \`${miss.input}\``)
+      }
+    }
+  }
+  lines.push('')
+
+  // ---------- 产物审计 ----------
+  lines.push('## 📦 产物审计')
+  lines.push('')
+  if (audit.files.length === 0) {
+    lines.push('- 未在工作树中检测到相关 ts/tsx 改动')
+  } else {
+    lines.push(`改动文件：${audit.files.length} 个`)
+    for (const f of audit.files) lines.push(`  - ${f}`)
+    lines.push('')
+
+    const apexEntries = Object.entries(audit.apexImports)
+    if (apexEntries.length > 0) {
+      lines.push(`apex-ui 组件复用：${apexEntries.length} 种`)
+      for (const [name, n] of apexEntries) lines.push(`  - ${name} × ${n}`)
+      lines.push('')
+    } else {
+      lines.push('apex-ui 组件复用：0  ⚠️ 检查是否漏用组件库')
+      lines.push('')
+    }
+
+    const variableEntries = Object.entries(audit.cssVariables)
+    if (variableEntries.length > 0) {
+      lines.push('CSS variable 引用：')
+      for (const [name, n] of variableEntries) lines.push(`  - ${name} × ${n}`)
+      lines.push('')
+    }
+
+    const commonEntries = Object.entries(audit.commonRefs)
+    if (commonEntries.length > 0) {
+      lines.push('common component / common module 复用：')
+      for (const [ref, n] of commonEntries) lines.push(`  - ${ref} × ${n}`)
+      lines.push('')
+    }
+
+    if (audit.hardcoded.length > 0) {
+      lines.push(`硬编码嫌疑：${audit.hardcoded.length} 处`)
+      for (const h of audit.hardcoded) {
+        lines.push(`  - ${h.file}:${h.line}  \`${h.value}\`  — ${h.snippet}`)
+      }
+      lines.push('')
+    } else {
+      lines.push('硬编码嫌疑：0 ✅')
+      lines.push('')
+    }
+
+    if (audit.pandaForbidden.length > 0) {
+      lines.push(`Panda 禁用内容：${audit.pandaForbidden.length} 处 ⚠️`)
+      for (const hit of audit.pandaForbidden) {
+        lines.push(`  - ${hit.file}:${hit.line}  \`${hit.value}\``)
+      }
+      lines.push('')
+    }
+  }
+
+  return lines.join('\n') + '\n'
+}
+
+function writeReport({ report, command, targetDir }) {
+  const fileName = `session-report-${command}.md`
+  const primary = path.join(DEFAULT_REPORT_DIR, fileName)
+  fs.mkdirSync(DEFAULT_REPORT_DIR, { recursive: true })
+  fs.writeFileSync(primary, report)
+  const written = [primary]
+
+  if (targetDir) {
+    fs.mkdirSync(targetDir, { recursive: true })
+    const copy = path.join(targetDir, fileName)
+    fs.writeFileSync(copy, report)
+    written.push(copy)
+  }
+  return written
+}
+
+function main() {
+  const args = parseArgs(process.argv)
+  if (args.reset) {
+    resetSessionLog()
+    console.log('[figma:report] session log reset.')
+    return
+  }
+
+  const log = tallyLog(readSessionLog())
+  const changedFiles = getChangedFiles()
+  const audit = aggregate(
+    changedFiles.filter((c) => /\.(ts|tsx|css)$/.test(c.file)).map((c) => c.file)
+  )
+  const planTokens = parsePlanTokenTable(args.plan)
+  const forbiddenFiles = auditForbiddenFiles(changedFiles)
+  const report = renderReport({
+    log,
+    audit,
+    planTokens,
+    command: args.command,
+    changedFiles,
+    forbiddenFiles,
+    generatedAt: new Date().toISOString().replace(/\.\d{3}Z$/, 'Z')
+  })
+
+  const written = writeReport({ report, command: args.command, targetDir: args.targetDir })
+  console.log(report)
+  for (const w of written) console.log(`[figma:report] saved -> ${w}`)
+}
+
+main()

package/skills/execution/figmaSync/scripts/icon-inventory.mjs

@@ -0,0 +1,165 @@
+/**
+ * Icon Inventory
+ *
+ * 扫描仓库中真实存在的 icon 资产源，供 figmaSync plan 做语义匹配。
+ *
+ * 用法:
+ *   node .agent/skills/figmaSync/scripts/icon-inventory.mjs
+ *   node .agent/skills/figmaSync/scripts/icon-inventory.mjs --pretty
+ */
+import fs from 'node:fs'
+import path from 'node:path'
+
+const ROOT = process.cwd()
+const ICON_EXTENSIONS = new Set(['.svg', '.png', '.jpg', '.jpeg', '.webp', '.gif', '.tsx', '.ts'])
+const STATIC_EXTENSIONS = new Set(['.svg', '.png', '.jpg', '.jpeg', '.webp', '.gif'])
+const SEARCH_ROOTS = ['assets', 'icons', 'public', 'src']
+const IGNORE_DIRS = new Set([
+  '.git',
+  '.agent',
+  'dist',
+  'node_modules',
+  'styled-system',
+  'coverage',
+  '.turbo',
+  '.vite'
+])
+
+function main() {
+  const pretty = process.argv.includes('--pretty')
+  const roots = SEARCH_ROOTS.filter((dir) => fs.existsSync(path.join(ROOT, dir)))
+  const files = []
+
+  for (const root of roots) {
+    walk(path.join(ROOT, root), files)
+  }
+
+  const icons = files
+    .map((file) => toIconRecord(file))
+    .flat()
+    .sort((a, b) => a.name.localeCompare(b.name) || a.path.localeCompare(b.path))
+
+  const out = {
+    roots,
+    count: icons.length,
+    icons,
+    note:
+      icons.length === 0
+        ? '仓库未发现可用 icon 源；请用户提供其他 icon 导入方案，或确认不关心 icon。'
+        : null
+  }
+
+  console.log(JSON.stringify(out, null, pretty ? 2 : 0))
+}
+
+function walk(dir, files) {
+  const entries = fs.readdirSync(dir, { withFileTypes: true })
+  for (const entry of entries) {
+    if (entry.name.startsWith('.')) continue
+    if (entry.isDirectory()) {
+      if (IGNORE_DIRS.has(entry.name)) continue
+      const next = path.join(dir, entry.name)
+      if (isIconSearchDir(next)) walk(next, files)
+      continue
+    }
+    if (!entry.isFile()) continue
+    const file = path.join(dir, entry.name)
+    const ext = path.extname(entry.name).toLowerCase()
+    if (!ICON_EXTENSIONS.has(ext)) continue
+    if (isInAllowedIconPath(file) || looksLikeIconFile(file)) {
+      files.push(file)
+    }
+  }
+}
+
+function isIconSearchDir(absDir) {
+  const rel = toPosix(path.relative(ROOT, absDir))
+  if (!rel) return true
+  const parts = rel.split('/')
+  if (parts.some((p) => IGNORE_DIRS.has(p))) return false
+  if (parts[0] === 'assets' || parts[0] === 'icons' || parts[0] === 'public') return true
+  if (parts[0] === 'src') return true
+  return false
+}
+
+function isInAllowedIconPath(absFile) {
+  const rel = toPosix(path.relative(ROOT, absFile))
+  const parts = rel.split('/')
+  if (parts[0] === 'assets' || parts[0] === 'icons' || parts[0] === 'public') return true
+  return parts[0] === 'src' && (parts.includes('assets') || parts.includes('icons'))
+}
+
+function looksLikeIconFile(absFile) {
+  const basename = path.basename(absFile, path.extname(absFile))
+  return /\bicon\b/i.test(basename) || /(^|[-_])ic([-_]|$)/i.test(basename)
+}
+
+function toIconRecord(absFile) {
+  const ext = path.extname(absFile).toLowerCase()
+  const rel = toPosix(path.relative(ROOT, absFile))
+  if (STATIC_EXTENSIONS.has(ext)) {
+    const baseName = path.basename(absFile, ext)
+    return [
+      {
+        name: baseName,
+        normalizedName: normalizeName(baseName),
+        type: ext.slice(1),
+        path: rel,
+        exportName: null
+      }
+    ]
+  }
+
+  const source = fs.readFileSync(absFile, 'utf-8')
+  const exportNames = extractIconExports(source)
+  if (exportNames.length === 0) return []
+  return exportNames.map((exportName) => ({
+    name: exportName,
+    normalizedName: normalizeName(exportName),
+    type: 'component',
+    path: rel,
+    exportName
+  }))
+}
+
+function extractIconExports(source) {
+  const names = new Set()
+  const patterns = [
+    /export\s+(?:const|function)\s+([A-Z][A-Za-z0-9]*Icon|Icon[A-Z][A-Za-z0-9]*)\b/g,
+    /export\s*\{\s*([^}]+)\s*\}/g
+  ]
+
+  let m
+  while ((m = patterns[0].exec(source)) !== null) {
+    names.add(m[1])
+  }
+
+  while ((m = patterns[1].exec(source)) !== null) {
+    for (const raw of m[1].split(',')) {
+      const name = raw
+        .trim()
+        .replace(/\s+as\s+.+$/, '')
+        .trim()
+      if (/([A-Z][A-Za-z0-9]*Icon|Icon[A-Z][A-Za-z0-9]*)$/.test(name)) names.add(name)
+    }
+  }
+
+  return [...names]
+}
+
+function normalizeName(name) {
+  return name
+    .replace(/Icon$/i, '')
+    .replace(/^Icon/i, '')
+    .replace(/([a-z0-9])([A-Z])/g, '$1-$2')
+    .replace(/[_\s]+/g, '-')
+    .replace(/[^a-zA-Z0-9-]/g, '')
+    .toLowerCase()
+    .replace(/^-+|-+$/g, '')
+}
+
+function toPosix(p) {
+  return p.split(path.sep).join('/')
+}
+
+main()