ci: build registry-index.json in-repo for dashboard real-time updates

Pair with the existing plugins-index.json path (which feeds the
daemon's signed install lane). registry-index.json mirrors the
dict-shaped payload the registry-worker's cron currently builds via
40+ GitHub Contents API calls — but built locally from the checked-out
tree by scripts/build-registry-index.mjs, so the worker only fetches
ONE file per category type (2 total: plugins + registry) on refresh.

Workflow now picks up content changes across all 8 category dirs
(was: plugins/ only) so dashboard updates land within seconds of a
push instead of waiting for the 02:00 UTC cron tick.

The dashboard's /api/registry endpoint reads kv_store('registry_data');
the worker's forced-refresh now writes that key with these bytes and
purges the Cache-API entry, so the next dashboard hit sees fresh
data instead of the 1h-cached previous payload.

Generated counts on first build: 11p 17h 32a 60s 44c 57pr 22w 33mcp.

.github/workflows/refresh-cache.yml

@@ -1,26 +1,34 @@
 name: Refresh registry-worker cache
 
-# On every push that changes plugins/, regenerate plugins-index.json
-# (the daemon-shaped flat array the registry-worker signs and the
-# librefang daemon installs from), commit it back, then poke the
-# worker's forced-refresh endpoint so it re-signs the new bytes
-# immediately — without this, daemon installs lag up to ~24h behind
-# the next 02:00 UTC cron tick.
+# On every push that changes any registry content, regenerate the two
+# in-repo indexes the registry-worker ingests:
+#   plugins-index.json  — daemon-shaped flat plugins array (signed)
+#   registry-index.json — dict-shaped dashboard payload (unsigned)
+# then poke the worker's forced-refresh endpoint so it pulls both
+# (2 subrequests total, regardless of registry size — important under
+# Workers Free's 50-subrequest budget) and re-signs / stores them.
 #
-# Walking the on-disk repo (instead of the worker hitting GitHub
-# Contents API per-file) keeps the worker's refresh path under the
-# Workers Free 50-subrequest budget regardless of registry size.
+# Without this, dashboard + daemon would have to wait for the next
+# 02:00 UTC cron tick to see content changes (up to ~24h delay).
 
 on:
   push:
     branches: [main]
     paths:
       - 'plugins/**'
+      - 'agents/**'
+      - 'skills/**'
+      - 'hands/**'
+      - 'channels/**'
+      - 'providers/**'
+      - 'workflows/**'
+      - 'mcp/**'
       - 'scripts/build-plugins-index.mjs'
-  workflow_dispatch: # manual trigger for ops / first-deploy
+      - 'scripts/build-registry-index.mjs'
+  workflow_dispatch:
 
 permissions:
-  contents: write   # needed to commit the regenerated index back
+  contents: write   # commit regenerated index files back
 
 jobs:
   refresh:
@@ -32,18 +40,20 @@ jobs:
         with:
           node-version: '20'
 
-      - name: Rebuild plugins-index.json
-        run: node scripts/build-plugins-index.mjs
+      - name: Rebuild indexes
+        run: |
+          node scripts/build-plugins-index.mjs
+          node scripts/build-registry-index.mjs
 
-      - name: Commit regenerated index if changed
+      - name: Commit regenerated indexes if changed
         run: |
           git config user.name "github-actions[bot]"
           git config user.email "github-actions[bot]@users.noreply.github.com"
-          git add plugins-index.json
+          git add plugins-index.json registry-index.json
           if git diff --cached --quiet; then
-            echo "plugins-index.json already up-to-date"
+            echo "indexes already up-to-date"
           else
-            git commit -m "chore: regenerate plugins-index.json"
+            git commit -m "chore: regenerate registry indexes"
             git push
           fi
 

registry-index.json

@@ -0,0 +1 @@
+{"hands":[{"id":"analytics","name":"Analytics Hand","description":"Autonomous data analytics agent — data collection, analysis, visualization, dashboards, and automated reporting","category":"data","icon":"lucide:trending-up","version":"1.1.0","i18n":{"zh":{"description":"自主数据分析——数据采集、分析、可视化、仪表盘与自动报告"},"zh-TW":{"description":"自主資料分析——資料採集、分析、視覺化、儀表板與自動報告"},"ja":{"description":"自律型データ分析エージェント——データ収集、分析、可視化、ダッシュボードと自動レポート"},"es":{"description":"Agente autónomo de análisis de datos — recopilación, análisis, visualización, dashboards e informes automatizados"},"fr":{"description":"Agent autonome d'analyse de données — collecte, analyse, visualisation, tableaux de bord et rapports automatisés"},"de":{"description":"Autonomer Datenanalyse-Agent — Datenerfassung, Analyse, Visualisierung, Dashboards und automatisierte Berichte"},"ko":{"description":"자율 데이터 분석 에이전트 — 데이터 수집, 분석, 시각화, 대시보드 및 자동 보고"}}},{"id":"apitester","name":"API Tester Hand","description":"Autonomous API testing agent — endpoint discovery, request validation, load testing, and regression detection","category":"development","icon":"lucide:plug","version":"1.1.0","i18n":{"zh":{"description":"自主 API 测试——端点发现、请求验证、压力测试与回归检测"},"zh-TW":{"description":"自主 API 測試——端點發現、請求驗證、壓力測試與回歸偵測"},"ja":{"description":"自律型APIテストエージェント——エンドポイント発見、リクエスト検証、負荷テスト、回帰検出"},"es":{"description":"Agente autónomo de pruebas API — descubrimiento de endpoints, validación de solicitudes, pruebas de carga y detección de regresiones"},"fr":{"description":"Agent autonome de test d'API — découverte de points de terminaison, validation de requêtes, tests de charge et détection de régression"},"de":{"description":"Autonomer API-Test-Agent — Endpoint-Erkennung, Request-Validierung, Lasttests und Regressionserkennung"},"ko":{"description":"자율 API 테스트 에이전트 — 엔드포인트 발견, 요청 검증, 부하 테스트, 회귀 감지"}}},{"id":"browser","name":"Browser Hand","description":"Autonomous web browser — navigates sites, fills forms, clicks buttons, and completes multi-step web tasks with user approval for purchases","category":"productivity","icon":"lucide:globe","version":"1.1.0","tags":["popular"],"i18n":{"zh-TW":{"description":"自主瀏覽器——導航網站、填寫表單、點擊按鈕，購買需用戶審批"}}},{"id":"clip","name":"Clip Hand","description":"Turns long-form video into viral short clips with captions and thumbnails","category":"content","icon":"lucide:film","version":"1.1.0","tags":["popular"],"i18n":{"zh-TW":{"description":"將長影片自動轉為帶字幕和封面的直式短影音"}}},{"id":"collector","name":"Collector Hand","description":"Autonomous intelligence collector — monitors any target continuously with change detection and knowledge graphs","category":"data","icon":"lucide:search","version":"1.1.0","tags":["popular"],"i18n":{"zh-TW":{"description":"自主情報收集器——持續監控目標，變化偵測與知識圖譜"}}},{"id":"creator","name":"Creator Hand","description":"AI media studio — generates images, videos, music, and speech from text prompts","category":"content","icon":"lucide:palette","version":"1.0.0","i18n":{"zh":{"description":"AI 媒体工作室——根据文本提示生成图片、视频、音乐和语音"},"zh-TW":{"description":"AI 媒體工作室——根據文字提示生成圖片、影片、音樂和語音"},"ja":{"description":"AIメディアスタジオ——テキストから画像、動画、音楽、音声を生成"},"ko":{"description":"AI 미디어 스튜디오 — 텍스트 프롬프트로 이미지, 비디오, 음악, 음성 생성"},"es":{"description":"Estudio de medios IA — genera imágenes, videos, música y voz a partir de texto"},"fr":{"description":"Studio média IA — génère images, vidéos, musique et voix à partir de texte"},"de":{"description":"AI-Medienstudio — erzeugt Bilder, Videos, Musik und Sprache aus Textprompts"}}},{"id":"devops","name":"DevOps Hand","description":"Autonomous DevOps engineer — CI/CD management, infrastructure monitoring, deployment automation, and incident response","category":"development","icon":"lucide:hard-hat","version":"1.1.0","i18n":{"zh":{"description":"自主 DevOps 工程师——CI/CD 管理、基础设施监控、部署自动化与事件响应"},"zh-TW":{"description":"自主 DevOps 工程師——CI/CD 管理、基礎設施監控、部署自動化與事件回應"},"ja":{"description":"自律型DevOpsエンジニア——CI/CD管理、インフラ監視、デプLine truncated

scripts/build-registry-index.mjs

@@ -0,0 +1,265 @@
+#!/usr/bin/env node
+// Build registry-index.json from the on-disk content tree.
+//
+// Mirrors the dict-shaped payload the registry-worker's
+// `refreshRegistryCache` would produce by walking the GitHub Contents API
+// — but does it from the checked-out repo, so the worker's forced-refresh
+// path stays at a single subrequest regardless of registry size.
+//
+// Output shape (consumed by dashboard via /api/registry):
+//   {
+//     hands:[{id,name,description,category,icon,tags?,i18n?}, ...],
+//     channels:[...], providers:[...], workflows:[...], agents:[...],
+//     plugins:[...], skills:[...], mcp:[...],
+//     handsCount: ..., ..., pluginsCount: ..., skillsCount: ...,
+//     fetchedAt: <ISO>,
+//     signature: <directory-signature>,    // for cron-skip parity
+//   }
+
+import fs from 'node:fs'
+import path from 'node:path'
+import crypto from 'node:crypto'
+
+const repoRoot = path.resolve(new URL('.', import.meta.url).pathname, '..')
+const outPath = path.join(repoRoot, 'registry-index.json')
+
+// Layouts: directory-with-manifest categories use `<name>/<manifest>`;
+// flat-file categories list `*.toml` directly. mcp tolerates both
+// (matches the existing worker `m.name.endsWith('.toml') ? mcp/<name> :
+// mcp/<name>/MCP.toml`).
+const DIR_MANIFEST = {
+  hands: 'HAND.toml',
+  agents: 'agent.toml',
+  plugins: 'plugin.toml',
+}
+const SKILL_DIR = 'SKILL.md'
+const FLAT_TOML = ['channels', 'providers', 'workflows']
+
+function pickString(text, key) {
+  const m = text.match(new RegExp(`^${key}\\s*=\\s*"([^"]*)"`, 'm'))
+  return m ? m[1] : ''
+}
+
+function pickStringArray(text, key) {
+  const m = text.match(new RegExp(`^${key}\\s*=\\s*\\[([^\\]]*)\\]`, 'm'))
+  if (!m) return undefined
+  const items = m[1].match(/"([^"]*)"/g)?.map(s => s.replace(/"/g, ''))
+  return items?.length ? items : undefined
+}
+
+function parseI18n(text) {
+  const i18n = {}
+  const re = /\[i18n\.([a-zA-Z-]+)\]\s*\n([^[]*?)(?=\n\[|\n*$)/g
+  let m
+  while ((m = re.exec(text)) !== null) {
+    const descM = (m[2] || '').match(/description\s*=\s*"([^"]*)"/)
+    if (descM) i18n[m[1]] = { description: descM[1] }
+  }
+  return i18n
+}
+
+function parseToml(text, fallbackId) {
+  const out = {
+    id: pickString(text, 'id') || fallbackId,
+    name: pickString(text, 'name') || fallbackId,
+    description: pickString(text, 'description'),
+    category: pickString(text, 'category'),
+    icon: pickString(text, 'icon'),
+  }
+  const version = pickString(text, 'version')
+  if (version) out.version = version
+  const tags = pickStringArray(text, 'tags')
+  if (tags?.length) out.tags = tags
+  const needs = pickStringArray(text, 'needs')
+  if (needs?.length) out.needs = needs
+  const i18n = parseI18n(text)
+  if (Object.keys(i18n).length) out.i18n = i18n
+  return out
+}
+
+function parseSkillMd(text, fallbackId) {
+  const fm = text.match(/^---\s*\n([\s\S]*?)\n---/)
+  if (!fm) return null
+  const get = key => {
+    const m = fm[1].match(new RegExp(`^${key}\\s*:\\s*"?([^"\\n]*?)"?\\s*$`, 'm'))
+    return m ? m[1].trim() : ''
+  }
+  return {
+    id: get('id') || fallbackId,
+    name: get('name') || fallbackId,
+    description: get('description'),
+    category: 'skills',
+    icon: '',
+  }
+}
+
+function loadDirCategory(category, manifest) {
+  const dir = path.join(repoRoot, category)
+  if (!fs.existsSync(dir)) return { items: [], rawNames: [] }
+  const entries = fs
+    .readdirSync(dir, { withFileTypes: true })
+    .filter(d => d.isDirectory() && d.name !== 'README.md')
+    .sort((a, b) => a.name.localeCompare(b.name))
+
+  const items = []
+  for (const e of entries) {
+    const f = path.join(dir, e.name, manifest)
+    if (!fs.existsSync(f)) continue
+    items.push(parseToml(fs.readFileSync(f, 'utf8'), e.name))
+  }
+  return { items, rawNames: entries.map(e => e.name) }
+}
+
+function loadFlatTomlCategory(category) {
+  const dir = path.join(repoRoot, category)
+  if (!fs.existsSync(dir)) return { items: [], rawNames: [] }
+  const files = fs
+    .readdirSync(dir)
+    .filter(n => n.endsWith('.toml') && n !== 'README.md')
+    .sort((a, b) => a.localeCompare(b))
+
+  const items = []
+  for (const f of files) {
+    const id = f.replace(/\.toml$/, '')
+    items.push(parseToml(fs.readFileSync(path.join(dir, f), 'utf8'), id))
+  }
+  return { items, rawNames: files }
+}
+
+function loadSkills() {
+  const dir = path.join(repoRoot, 'skills')
+  if (!fs.existsSync(dir)) return { items: [], rawNames: [] }
+  const dirs = fs
+    .readdirSync(dir, { withFileTypes: true })
+    .filter(d => d.isDirectory())
+    .sort((a, b) => a.name.localeCompare(b.name))
+
+  const items = []
+  for (const d of dirs) {
+    const f = path.join(dir, d.name, SKILL_DIR)
+    if (!fs.existsSync(f)) continue
+    const parsed = parseSkillMd(fs.readFileSync(f, 'utf8'), d.name)
+    if (parsed) items.push(parsed)
+  }
+  return { items, rawNames: dirs.map(d => d.name) }
+}
+
+function loadMcp() {
+  const dir = path.join(repoRoot, 'mcp')
+  if (!fs.existsSync(dir)) return { items: [], rawNames: [] }
+  const entries = fs
+    .readdirSync(dir, { withFileTypes: true })
+    .filter(e => e.name !== 'README.md')
+    .sort((a, b) => a.name.localeCompare(b.name))
+
+  const items = []
+  const rawNames = []
+  for (const e of entries) {
+    let manifest
+    let id
+    if (e.isFile() && e.name.endsWith('.toml')) {
+      manifest = path.join(dir, e.name)
+      id = e.name.replace(/\.toml$/, '')
+    } else if (e.isDirectory()) {
+      manifest = path.join(dir, e.name, 'MCP.toml')
+      id = e.name
+      if (!fs.existsSync(manifest)) continue
+    } else {
+      continue
+    }
+    items.push(parseToml(fs.readFileSync(manifest, 'utf8'), id))
+    rawNames.push(e.name)
+  }
+  return { items, rawNames }
+}
+
+const hands = loadDirCategory('hands', DIR_MANIFEST.hands)
+const agents = loadDirCategory('agents', DIR_MANIFEST.agents)
+const plugins = loadDirCategory('plugins', DIR_MANIFEST.plugins)
+const skills = loadSkills()
+const channels = loadFlatTomlCategory('channels')
+const providers = loadFlatTomlCategory('providers')
+const workflows = loadFlatTomlCategory('workflows')
+const mcp = loadMcp()
+
+// Signature mirrors the worker's per-category `name@sha`-style join, but
+// keyed off the file content hash so the registry-worker cron can detect
+// "no real change" identically to its current GitHub-Contents-API path.
+function categorySig(category, rawNames) {
+  return rawNames
+    .map(n => {
+      const p = path.join(repoRoot, category, n)
+      let bytes = ''
+      if (fs.existsSync(p)) {
+        if (fs.statSync(p).isDirectory()) {
+          // hash the directory's manifest for the signature
+          for (const m of ['HAND.toml', 'agent.toml', 'plugin.toml', 'SKILL.md', 'MCP.toml']) {
+            const inner = path.join(p, m)
+            if (fs.existsSync(inner)) {
+              bytes = fs.readFileSync(inner, 'utf8')
+              break
+            }
+          }
+        } else {
+          bytes = fs.readFileSync(p, 'utf8')
+        }
+      }
+      const sha = crypto.createHash('sha1').update(bytes).digest('hex')
+      return `${n}@${sha}`
+    })
+    .join(',')
+}
+
+const signature = [
+  'hands', 'channels', 'providers', 'workflows',
+  'agents', 'plugins', 'skills', 'mcp',
+]
+  .map((c, _) => {
+    const raw = ({ hands, channels, providers, workflows, agents, plugins, skills, mcp }[c]).rawNames
+    return `${c}=${categorySig(c, raw)}`
+  })
+  .join('|')
+
+const result = {
+  hands: hands.items,
+  channels: channels.items,
+  providers: providers.items,
+  workflows: workflows.items,
+  agents: agents.items,
+  plugins: plugins.items,
+  skills: skills.items,
+  mcp: mcp.items,
+  handsCount: hands.rawNames.length,
+  channelsCount: channels.rawNames.length,
+  providersCount: providers.rawNames.length,
+  workflowsCount: workflows.rawNames.length,
+  agentsCount: agents.rawNames.length,
+  pluginsCount: plugins.rawNames.length,
+  skillsCount: skills.rawNames.length,
+  mcpCount: mcp.rawNames.length,
+  // fetchedAt intentionally OMITTED — including a wall-clock timestamp
+  // here would make every CI run produce a diff and an empty commit
+  // even when the registry hasn't changed. The worker stamps an
+  // updated_at into D1 when it ingests the file, which is what
+  // dashboards should display anyway.
+  signature,
+}
+
+const json = JSON.stringify(result)
+const prev = fs.existsSync(outPath) ? fs.readFileSync(outPath, 'utf8') : null
+if (prev === json) {
+  console.log(
+    `registry-index.json unchanged (${plugins.rawNames.length} plugins, ` +
+    `${hands.rawNames.length} hands, ${agents.rawNames.length} agents, ` +
+    `${skills.rawNames.length} skills, ${channels.rawNames.length} channels, ` +
+    `${providers.rawNames.length} providers, ${workflows.rawNames.length} workflows, ` +
+    `${mcp.rawNames.length} mcp)`,
+  )
+  process.exit(0)
+}
+fs.writeFileSync(outPath, json)
+console.log(
+  `registry-index.json updated: ${plugins.rawNames.length}p ${hands.rawNames.length}h ` +
+  `${agents.rawNames.length}a ${skills.rawNames.length}s ${channels.rawNames.length}c ` +
+  `${providers.rawNames.length}pr ${workflows.rawNames.length}w ${mcp.rawNames.length}mcp`,
+)