diff --git a/.memory/worklog.json b/.memory/worklog.json
index ced2950..07fb61f 100644
--- a/.memory/worklog.json
+++ b/.memory/worklog.json
@@ -3309,6 +3309,19 @@
       "type": "session-heartbeat",
       "message": "Codex 会话活跃 · 最近命令：codex · 3 项未提交变更 · 最近提交：auto-save 2026-05-14 05:05 (~6)",
       "files_changed": 3
+    },
+    {
+      "ts": "2026-05-14T05:10:53+08:00",
+      "type": "commit",
+      "message": "auto-save 2026-05-14 05:10 (~3)",
+      "hash": "c7ca830",
+      "files_changed": 3
+    },
+    {
+      "ts": "2026-05-13T21:13:13Z",
+      "type": "session-heartbeat",
+      "message": "Claude 会话活跃 · 最近命令：claude · 2 项未提交变更 · 最近提交：auto-save 2026-05-14 05:10 (~3)",
+      "files_changed": 2
     }
   ]
 }
diff --git a/web/components/lightbox.tsx b/web/components/lightbox.tsx
index 7d00f36..10d5ce7 100644
--- a/web/components/lightbox.tsx
+++ b/web/components/lightbox.tsx
@@ -47,6 +47,15 @@ const LIVING_VIEW_OPTIONS = [
   ["action_use", "使用"],
 ]
 
+type LightboxTab = "clean" | "scene" | "subject" | "review"
+
+const LIGHTBOX_TABS: Array<{ key: LightboxTab; label: string }> = [
+  { key: "clean", label: "原图/清洗" },
+  { key: "scene", label: "场景图" },
+  { key: "subject", label: "主体包" },
+  { key: "review", label: "审核" },
+]
+
 export function FrameLightbox({ jobId, frames, activeIndex, selected, onClose, onChange, onToggleSelect, onJobUpdate, onSwitchPanel, onCopyImage, embedded = false }: Props) {
   const [describing, setDescribing] = useState(false)
   const [cleaning, setCleaning] = useState(false)
@@ -60,6 +69,7 @@ export function FrameLightbox({ jobId, frames, activeIndex, selected, onClose, o
   const [subjectKinds, setSubjectKinds] = useState<Record<string, SubjectKind>>({})
   const [subjectBackgrounds, setSubjectBackgrounds] = useState<Record<string, AssetBackground>>({})
   const [subjectViews, setSubjectViews] = useState<Record<string, string[]>>({})
+  const [activeTab, setActiveTab] = useState<LightboxTab>("clean")
   const [editingElement, setEditingElement] = useState<{
     id: string
     name_zh: string
@@ -119,6 +129,12 @@ export function FrameLightbox({ jobId, frames, activeIndex, selected, onClose, o
   const latestSceneAsset = f.scene_assets?.[f.scene_assets.length - 1] ?? null
   const selectedFrameIndices = Array.from(selected).sort((a, b) => a - b)
   const sharedSubjectFrameIndices = selectedFrameIndices.length > 1 ? selectedFrameIndices : [f.index]
+  const subjectAssetCount = elements.reduce((sum, item) => sum + (item.subject_assets?.length ?? 0), 0)
+  const cutoutCount = elements.reduce((sum, item) => sum + ((item.cutouts?.length ?? 0) || (item.cutout_id ? 1 : 0)), 0)
+  const qualityWarnings = [
+    ...(f.quality_report?.warnings ?? []),
+    ...(latestSceneAsset?.quality_report?.warnings ?? []),
+  ]
 
   const handleDescribe = async () => {
     setDescribing(true)
@@ -424,6 +440,28 @@ export function FrameLightbox({ jobId, frames, activeIndex, selected, onClose, o
         </div>
       )}
 
+      <div className="flex items-center gap-1 border-b border-white/10 bg-black/28 px-3 py-2">
+        {LIGHTBOX_TABS.map((tab) => (
+          <button
+            key={tab.key}
+            type="button"
+            onClick={() => setActiveTab(tab.key)}
+            className={`h-7 rounded-md px-2.5 text-[11px] font-medium transition ${
+              activeTab === tab.key
+                ? "bg-white text-black shadow"
+                : "bg-white/[0.06] text-white/58 hover:bg-white/[0.12] hover:text-white"
+            }`}
+          >
+            {tab.label}
+          </button>
+        ))}
+        <div className="ml-auto hidden items-center gap-2 text-[10px] text-white/42 sm:flex">
+          <span>{latestSceneAsset ? "场景已生成" : "场景待生成"}</span>
+          <span>·</span>
+          <span>{subjectAssetCount > 0 ? `${subjectAssetCount} 主体资产` : "主体待生成"}</span>
+        </div>
+      </div>
+
       {/* 主体 — 左：大图 + 清洗 / 选用；右：识别 + 元素清单 */}
       <div className="flex gap-3 p-3 overflow-hidden flex-1 min-h-0">
         {/* 左侧大图区 */}
@@ -485,6 +523,8 @@ export function FrameLightbox({ jobId, frames, activeIndex, selected, onClose, o
             )}
           </div>
 
+          {activeTab === "clean" && (
+            <>
           {/* 画框工具栏 */}
           {cropMode ? (
             extractNamePrompt ? (
@@ -623,7 +663,10 @@ export function FrameLightbox({ jobId, frames, activeIndex, selected, onClose, o
             {cleaning ? <Loader2 className="h-3.5 w-3.5 animate-spin" /> : <Sparkle className="h-3.5 w-3.5" />}
             {cleaning ? "清洗中…（5-15 秒）" : hasCleaned ? "重新清洗" : f.cleaned_applied ? "再次清洗" : "🧹 清洗水印"}
           </button>
+            </>
+          )}
 
+          {activeTab === "scene" && (
           <div className="rounded-lg border border-white/10 bg-white/[0.035] p-2">
             <div className="mb-2 flex items-center justify-between gap-2">
               <div className="text-[11.5px] font-semibold text-white">场景图</div>
@@ -673,6 +716,37 @@ export function FrameLightbox({ jobId, frames, activeIndex, selected, onClose, o
               {sceneGenerating ? "生成场景图中…" : latestSceneAsset ? "重新生成场景图" : "生成场景图"}
             </button>
           </div>
+          )}
+
+          {activeTab === "review" && (
+            <div className="space-y-2 rounded-lg border border-white/10 bg-white/[0.035] p-2">
+              <div className="text-[11.5px] font-semibold text-white">素材审核</div>
+              <div className="grid grid-cols-2 gap-1.5 text-[10.5px]">
+                <div className={`rounded border px-2 py-1 ${f.cleaned_applied || hasCleaned ? "border-emerald-300/30 bg-emerald-500/10 text-emerald-100" : "border-white/10 bg-black/25 text-white/55"}`}>
+                  {f.cleaned_applied ? "已应用清洗" : hasCleaned ? "清洗待确认" : "未清洗"}
+                </div>
+                <div className={`rounded border px-2 py-1 ${latestSceneAsset ? "border-emerald-300/30 bg-emerald-500/10 text-emerald-100" : "border-white/10 bg-black/25 text-white/55"}`}>
+                  {latestSceneAsset ? "场景图已生成" : "场景图未生成"}
+                </div>
+                <div className={`rounded border px-2 py-1 ${subjectAssetCount > 0 ? "border-violet-300/35 bg-violet-500/12 text-violet-100" : "border-white/10 bg-black/25 text-white/55"}`}>
+                  {subjectAssetCount > 0 ? `${subjectAssetCount} 张主体资产` : "主体包未生成"}
+                </div>
+                <div className={`rounded border px-2 py-1 ${qualityWarnings.length ? "border-amber-300/35 bg-amber-500/12 text-amber-100" : "border-emerald-300/30 bg-emerald-500/10 text-emerald-100"}`}>
+                  {qualityWarnings.length ? `${qualityWarnings.length} 个风险` : "质量可用"}
+                </div>
+              </div>
+              {qualityWarnings.length > 0 && (
+                <div className="rounded border border-amber-300/25 bg-amber-500/10 px-2 py-1.5 text-[10px] leading-snug text-amber-100/85">
+                  {qualityWarnings.slice(0, 3).map((warning, i) => (
+                    <div key={i}>{warning}</div>
+                  ))}
+                </div>
+              )}
+              <div className="text-[10px] leading-relaxed text-white/42">
+                审核通过后，把场景图和主体资产复制到分镜槽位；当前不会自动覆盖素材，避免模型误改细节。
+              </div>
+            </div>
+          )}
 
           <button
             onClick={() => onToggleSelect(f.index)}
diff --git a/web/components/nodes/index.tsx b/web/components/nodes/index.tsx
index a84cb01..b90d5ac 100644
--- a/web/components/nodes/index.tsx
+++ b/web/components/nodes/index.tsx
@@ -11,6 +11,7 @@ import {
   Link2, Upload, Download, Scissors, Image as ImageIcon,
   Mic, Languages, FileEdit, Film, FileVideo, Loader2, Plus, X, LayoutGrid, Maximize2,
   Copy, Trash2, Move, PanelLeft, PanelRight, PanelBottom, ChevronLeft, ChevronRight, SlidersHorizontal,
+  CheckCircle2, AlertTriangle, Sparkles, Package,
 } from "lucide-react"
 import { toast } from "sonner"
 import { NodeShell, type NodeStatus, type NodeKind } from "./node-shell"
@@ -1129,6 +1130,8 @@ const THUMB_W = 64
 const THUMB_GAP = 6
 
 type ElementPreview = { frameIdx: number; elementId: string; name: string; src: string; cid: string; timestamp: number }
+type SceneAssetPreview = { frameIdx: number; assetId: string; label: string; src: string; width: number; height: number; risk?: string }
+type SubjectAssetPreview = { frameIdx: number; elementId: string; assetId: string; label: string; src: string; width: number; height: number; view: string }
 
 function collectElementCrops(job: Job | null): ElementPreview[] {
   return job
@@ -1154,6 +1157,41 @@ function collectElementCrops(job: Job | null): ElementPreview[] {
     : []
 }
 
+function collectSceneAssets(job: Job | null): SceneAssetPreview[] {
+  return job
+    ? job.frames.flatMap((f) =>
+        (f.scene_assets ?? []).map((asset) => ({
+          frameIdx: f.index,
+          assetId: asset.id,
+          label: asset.label || `分镜 ${f.index + 1} 场景图`,
+          src: apiAssetUrl(asset.url),
+          width: asset.width,
+          height: asset.height,
+          risk: asset.quality_report?.risk,
+        })).filter((p) => p.src),
+      )
+    : []
+}
+
+function collectSubjectAssets(job: Job | null): SubjectAssetPreview[] {
+  return job
+    ? job.frames.flatMap((f) =>
+        (f.elements ?? []).flatMap((element) =>
+          (element.subject_assets ?? []).map((asset) => ({
+            frameIdx: f.index,
+            elementId: element.id,
+            assetId: asset.id,
+            label: asset.label || `${element.name_zh} · ${asset.view}`,
+            src: apiAssetUrl(asset.url),
+            width: asset.width,
+            height: asset.height,
+            view: asset.view,
+          })),
+        ).filter((p) => p.src),
+      )
+    : []
+}
+
 function videoModelLabel(model: string) {
   const m = model.toLowerCase()
   if (m.includes("kling")) return "Kling"
@@ -1175,10 +1213,18 @@ export function VisualLabNode({ data, selected }: any) {
     ? `${job.width}/${job.height}`
     : "9/16"
   const elementCrops = collectElementCrops(job)
+  const sceneAssets = collectSceneAssets(job)
+  const subjectAssets = collectSubjectAssets(job)
   const cleanedCount = frames.filter((x) => x.cleaned_url).length
   const cutoutCount = frames.reduce((s, x) => s + (x.elements?.filter((e) => hasCutout(e)).length ?? 0), 0)
-  const sceneAssetCount = frames.reduce((s, x) => s + (x.scene_assets?.length ?? 0), 0)
-  const subjectAssetCount = frames.reduce((s, x) => s + (x.elements?.reduce((n, e) => n + (e.subject_assets?.length ?? 0), 0) ?? 0), 0)
+  const sceneAssetCount = sceneAssets.length
+  const subjectAssetCount = subjectAssets.length
+  const selectedFrameCount = frames.filter((f) => d.selectedFrames.has(f.index)).length
+  const targetFrameCount = selectedFrameCount || frames.length
+  const qualityRiskCount = frames.filter((f) => f.quality_report?.risk && f.quality_report.risk !== "ok").length
+  const preparedUnits = Math.min(targetFrameCount, sceneAssetCount) + (subjectAssetCount > 0 ? 1 : 0)
+  const totalUnits = Math.max(1, targetFrameCount + 1)
+  const prepPct = Math.min(100, Math.round((preparedUnits / totalUnits) * 100))
   const runningVideo = videos.some((v) => v.status === "queued" || v.status === "in_progress")
   const completedVideos = videos.filter((v) => v.status === "completed" && v.url)
   const failedVideo = videos.some((v) => v.status === "failed")
@@ -1191,9 +1237,11 @@ export function VisualLabNode({ data, selected }: any) {
     : keyframeStatus(job)
 
   type VisualPreview =
-    | { id: string; kind: "frame"; frameIdx: number; src: string; label: string; caption: string; borderClass: string }
-    | { id: string; kind: "cutout"; frameIdx: number; elementId: string; cutoutId: string; src: string; label: string; caption: string; borderClass: string }
-    | { id: string; kind: "video"; videoId: string; videoSrc?: string; posterSrc?: string; label: string; caption: string; borderClass: string }
+    | { id: string; kind: "frame"; group: string; frameIdx: number; src: string; label: string; caption: string; borderClass: string; aspect: string }
+    | { id: string; kind: "scene"; group: string; frameIdx: number; assetId: string; src: string; label: string; caption: string; borderClass: string; aspect: string }
+    | { id: string; kind: "subject"; group: string; frameIdx: number; assetId: string; src: string; label: string; caption: string; borderClass: string; aspect: string }
+    | { id: string; kind: "cutout"; group: string; frameIdx: number; elementId: string; cutoutId: string; src: string; label: string; caption: string; borderClass: string; aspect: string }
+    | { id: string; kind: "video"; group: string; videoId: string; videoSrc?: string; posterSrc?: string; label: string; caption: string; borderClass: string; aspect: string }
 
   const [hoverPreview, setHoverPreview] = useState<PreviewAnchor<string> | null>(null)
   const [pinnedPreview, setPinnedPreview] = useState<PreviewAnchor<string> | null>(null)
@@ -1204,15 +1252,42 @@ export function VisualLabNode({ data, selected }: any) {
     ...(job && jobId ? frames.map((f) => ({
       id: `frame:${f.index}`,
       kind: "frame" as const,
+      group: "关键帧",
       frameIdx: f.index,
       src: effectiveFrameUrl(jobId, f),
       label: `分镜 ${f.index + 1}`,
-      caption: `${f.timestamp.toFixed(2)}s`,
-      borderClass: "border-orange-300/50",
+      caption: `${f.timestamp.toFixed(2)}s${f.quality_report?.risk && f.quality_report.risk !== "ok" ? " · 风险" : ""}`,
+      borderClass: f.quality_report?.risk === "bad" ? "border-rose-300/70" : f.quality_report?.risk === "warn" ? "border-amber-300/70" : "border-orange-300/50",
+      aspect,
     })) : []),
+    ...sceneAssets.map((p) => ({
+      id: `scene:${p.frameIdx}:${p.assetId}`,
+      kind: "scene" as const,
+      group: "场景图",
+      frameIdx: p.frameIdx,
+      assetId: p.assetId,
+      src: p.src,
+      label: p.label,
+      caption: `${p.width}×${p.height}`,
+      borderClass: p.risk === "bad" ? "border-rose-300/70" : p.risk === "warn" ? "border-amber-300/70" : "border-emerald-300/60",
+      aspect: p.width && p.height ? `${p.width}/${p.height}` : aspect,
+    })),
+    ...subjectAssets.map((p) => ({
+      id: `subject:${p.frameIdx}:${p.assetId}`,
+      kind: "subject" as const,
+      group: "主体包",
+      frameIdx: p.frameIdx,
+      assetId: p.assetId,
+      src: p.src,
+      label: p.label,
+      caption: `${p.width}×${p.height}`,
+      borderClass: "border-violet-300/65",
+      aspect: p.width && p.height ? `${p.width}/${p.height}` : "1/1",
+    })),
     ...elementCrops.map((p) => ({
       id: `cutout:${p.frameIdx}:${p.elementId}:${p.cid}`,
       kind: "cutout" as const,
+      group: "普通抠图",
       frameIdx: p.frameIdx,
       elementId: p.elementId,
       cutoutId: p.cid,
@@ -1220,6 +1295,7 @@ export function VisualLabNode({ data, selected }: any) {
       label: p.name,
       caption: `分镜 ${p.frameIdx + 1}`,
       borderClass: "border-violet-300/60",
+      aspect: "1/1",
     })),
     ...videos.map((v, i) => {
       const videoSrc = apiAssetUrl(v.url)
@@ -1227,12 +1303,14 @@ export function VisualLabNode({ data, selected }: any) {
       return {
         id: `video:${v.id}`,
         kind: "video" as const,
+        group: "视频任务",
         videoId: v.id,
         videoSrc: v.status === "completed" && videoSrc ? videoSrc : undefined,
         posterSrc: posterSrc || undefined,
         label: `视频 ${i + 1}`,
         caption: `${videoModelLabel(v.model)} · ${v.status}`,
         borderClass: v.status === "completed" ? "border-emerald-300/60" : v.status === "failed" ? "border-rose-300/70" : "border-violet-300/55",
+        aspect,
       }
     }),
   ]
@@ -1266,11 +1344,14 @@ export function VisualLabNode({ data, selected }: any) {
                   p.kind === "frame"
                     ? isSelected ? "border-emerald-400 ring-2 ring-emerald-400/60" : "border-white/30 dark:border-white/20"
                     : p.borderClass
-                } ${p.kind === "cutout" ? "bg-white" : "bg-black"}`}
-                style={{ height: THUMBNAIL_HEIGHT, aspectRatio: aspect }}
+                } ${p.kind === "cutout" || p.kind === "subject" ? "bg-white" : "bg-black"}`}
+                style={{ height: THUMBNAIL_HEIGHT, aspectRatio: p.aspect }}
                 onMouseEnter={(e) => setHoverPreview({ id: p.id, ...canvasThumbnailAnchor(rootRef.current, e.currentTarget) })}
                 onMouseLeave={() => setHoverPreview(null)}
               >
+                <div className="absolute -top-6 left-0 z-[68] rounded bg-black/75 px-1.5 py-0.5 text-[9px] font-medium text-white/80 backdrop-blur">
+                  {p.group}
+                </div>
                 <button
                   type="button"
                   onClick={(e) => {
@@ -1279,6 +1360,16 @@ export function VisualLabNode({ data, selected }: any) {
                     setPinnedPreview((prev) => (prev?.id === p.id ? null : { id: p.id, ...anchor }))
                     if (p.kind === "frame") {
                       ;(d.onOpenFramePanel ?? d.onExpandFrame)(p.frameIdx)
+                    } else if (p.kind === "scene" || p.kind === "subject") {
+                      d.onCopyImage?.({
+                        kind: "asset",
+                        frame_idx: p.frameIdx,
+                        element_id: p.assetId,
+                        cutout_id: p.assetId,
+                        label: p.label,
+                      })
+                      if (!d.selectedFrames.has(p.frameIdx)) d.onToggleFrame(p.frameIdx)
+                      d.onOpenWorkbench?.(p.frameIdx)
                     } else if (p.kind === "cutout") {
                       if (!d.selectedFrames.has(p.frameIdx)) d.onToggleFrame(p.frameIdx)
                       d.onOpenStoryboard?.(p.frameIdx)
@@ -1291,7 +1382,7 @@ export function VisualLabNode({ data, selected }: any) {
                       }
                     }
                   }}
-                  title={`${p.label} · 单击钉住预览${p.kind === "frame" ? " / 打开镜头处理" : p.kind === "cutout" ? " / 进入分镜编排" : " / 复制 prompt"}`}
+                  title={`${p.label} · 单击钉住预览${p.kind === "frame" ? " / 打开素材审核" : p.kind === "video" ? " / 复制 prompt" : " / 复制到分镜编排"}`}
                   className="absolute inset-0 h-full w-full overflow-hidden rounded-md"
                 >
                   {p.kind === "video" ? (
@@ -1303,13 +1394,13 @@ export function VisualLabNode({ data, selected }: any) {
                       <div className="absolute inset-0 bg-violet-950/50" />
                     )
                   ) : (
-                    <img src={p.src} alt={p.label} className={`absolute inset-0 h-full w-full ${p.kind === "cutout" ? "object-contain" : "object-cover"}`} />
+                    <img src={p.src} alt={p.label} className={`absolute inset-0 h-full w-full ${p.kind === "frame" ? "object-cover" : "object-contain"}`} />
                   )}
                   {p.kind === "frame" && isSelected && (
                     <div className="absolute inset-0 bg-emerald-400/15 rounded-md pointer-events-none" />
                   )}
                   <div className="absolute bottom-0 right-0 bg-black/70 px-1 py-0.5 text-[8.5px] font-mono leading-none text-white rounded-bl rounded-br-md">
-                    {p.kind === "frame" ? p.caption.replace("s", "") + "s" : p.kind === "cutout" ? "元素" : "视频"}
+                    {p.kind === "frame" ? p.caption.replace("s", "") + "s" : p.kind === "scene" ? "场景" : p.kind === "subject" ? "主体" : p.kind === "cutout" ? "抠图" : "视频"}
                   </div>
                 </button>
 
@@ -1327,6 +1418,26 @@ export function VisualLabNode({ data, selected }: any) {
                   </button>
                 )}
 
+                {(p.kind === "scene" || p.kind === "subject") && d.onCopyImage && (
+                  <button
+                    type="button"
+                    onClick={(e) => {
+                      e.stopPropagation()
+                      d.onCopyImage?.({
+                        kind: "asset",
+                        frame_idx: p.frameIdx,
+                        element_id: p.assetId,
+                        cutout_id: p.assetId,
+                        label: p.label,
+                      })
+                    }}
+                    title="复制此素材（到分镜头编排工作台插槽粘贴）"
+                    className="absolute top-1.5 left-1.5 z-[70] inline-flex h-7 w-7 items-center justify-center rounded-full bg-violet-500/95 text-white shadow-lg backdrop-blur transition hover:scale-110 hover:bg-violet-400"
+                  >
+                    <Copy className="h-3.5 w-3.5" />
+                  </button>
+                )}
+
                 {p.kind === "cutout" && d.onCopyImage && (
                   <button
                     type="button"
@@ -1435,7 +1546,7 @@ export function VisualLabNode({ data, selected }: any) {
             imgSrc={p.kind === "video" ? p.posterSrc : p.src}
             videoSrc={p.kind === "video" ? p.videoSrc : undefined}
             poster={p.kind === "video" ? p.posterSrc : undefined}
-            aspect={aspect}
+            aspect={p.aspect}
             label={p.label}
             caption={p.caption}
             borderClass={p.borderClass}
@@ -1458,39 +1569,94 @@ export function VisualLabNode({ data, selected }: any) {
         pinned={d.pinnedNodes?.has("visual")}
         onTogglePin={() => d.onToggleNodePin?.("visual")}
       >
-        <div className="grid grid-cols-3 gap-2 text-[10.5px] text-[var(--text-soft)]">
+        <div className="rounded-md border border-white/10 bg-black/20 p-2">
+          <div className="mb-1.5 flex items-center justify-between gap-2">
+            <div className="text-[11px] font-semibold text-[var(--text-strong)]">素材准备进度</div>
+            <div className="font-mono text-[10px] text-[var(--text-faint)]">{prepPct}%</div>
+          </div>
+          <div className="h-2 overflow-hidden rounded-full bg-white/10">
+            <div
+              className="h-full rounded-full bg-emerald-400 transition-all"
+              style={{ width: `${prepPct}%` }}
+            />
+          </div>
+          <div className="mt-1.5 flex items-center justify-between gap-2 text-[9.5px] text-[var(--text-faint)]">
+            <span>{targetFrameCount} 张目标帧</span>
+            {qualityRiskCount > 0 ? (
+              <span className="inline-flex items-center gap-1 text-amber-300/85">
+                <AlertTriangle className="h-2.5 w-2.5" />
+                {qualityRiskCount} 个质量风险
+              </span>
+            ) : (
+              <span className="inline-flex items-center gap-1 text-emerald-300/80">
+                <CheckCircle2 className="h-2.5 w-2.5" />
+                无质量风险
+              </span>
+            )}
+          </div>
+        </div>
+
+        <div className="mt-2 grid grid-cols-2 gap-2 text-[10.5px] text-[var(--text-soft)]">
           <button
             type="button"
             onClick={(e) => { e.stopPropagation(); openFirstFrame() }}
             disabled={frames.length === 0}
-            className="rounded-md border border-white/10 px-2 py-1.5 text-left transition hover:border-orange-300/50 hover:bg-orange-400/10 disabled:opacity-35"
-            title="打开镜头处理面板"
+            className="min-h-14 rounded-md border border-white/10 px-2 py-2 text-left transition hover:border-orange-300/50 hover:bg-orange-400/10 disabled:opacity-35"
+            title="打开素材准备 / 审核面板"
           >
-            <div className="text-[var(--text-strong)] text-[12px] font-semibold">{frames.length}</div>
-            <div>关键帧</div>
+            <div className="mb-1 flex items-center gap-1 text-[var(--text-strong)] text-[12px] font-semibold">
+              <ImageIcon className="h-3 w-3 text-orange-300" />
+              {targetFrameCount}/{frames.length}
+            </div>
+            <div>目标关键帧</div>
+          </button>
+          <button
+            type="button"
+            onClick={(e) => { e.stopPropagation(); openFirstFrame() }}
+            disabled={!job || frames.length === 0}
+            className="min-h-14 rounded-md border border-white/10 px-2 py-2 text-left transition hover:border-emerald-300/50 hover:bg-emerald-400/10 disabled:opacity-35"
+            title="生成 / 审核每张关键帧的场景图"
+          >
+            <div className="mb-1 flex items-center gap-1 text-[var(--text-strong)] text-[12px] font-semibold">
+              <Sparkles className="h-3 w-3 text-emerald-300" />
+              {sceneAssetCount}/{targetFrameCount}
+            </div>
+            <div>场景图</div>
+          </button>
+          <button
+            type="button"
+            onClick={(e) => { e.stopPropagation(); openFirstFrame() }}
+            disabled={!job || frames.length === 0}
+            className="min-h-14 rounded-md border border-white/10 px-2 py-2 text-left transition hover:border-violet-300/50 hover:bg-violet-400/10 disabled:opacity-35"
+            title="生成主体多视角 / 动作 / 表情资产包"
+          >
+            <div className="mb-1 flex items-center gap-1 text-[var(--text-strong)] text-[12px] font-semibold">
+              <Package className="h-3 w-3 text-violet-300" />
+              {subjectAssetCount || cutoutCount}
+            </div>
+            <div>主体资产</div>
           </button>
           <button
             type="button"
             onClick={(e) => { e.stopPropagation(); d.onOpenWorkbench?.(frames.find((f) => d.selectedFrames.has(f.index))?.index ?? frames[0]?.index) }}
             disabled={!job || frames.length === 0}
-            className="rounded-md border border-white/10 px-2 py-1.5 text-left transition hover:border-violet-300/50 hover:bg-violet-400/10 disabled:opacity-35"
-            title="进入分镜编排"
+            className="min-h-14 rounded-md border border-white/10 px-2 py-2 text-left transition hover:border-pink-300/50 hover:bg-pink-400/10 disabled:opacity-35"
+            title="进入分镜编排和视频生成"
           >
-            <div className="text-[var(--text-strong)] text-[12px] font-semibold">{elementCrops.length}</div>
-            <div>元素 / 编排</div>
+            <div className="mb-1 flex items-center gap-1 text-[var(--text-strong)] text-[12px] font-semibold">
+              <Film className="h-3 w-3 text-pink-300" />
+              {videos.length}
+            </div>
+            <div>分镜 / 视频</div>
           </button>
-          <div className="rounded-md border border-white/10 px-2 py-1.5">
-            <div className="text-[var(--text-strong)] text-[12px] font-semibold">{videos.length}</div>
-            <div>视频任务</div>
-          </div>
         </div>
         <div className="mt-2 text-[10.5px] leading-snug text-[var(--text-faint)]">
           {frames.length > 0 ? (
             <>
-              {cleanedCount} 已清洗 · {sceneAssetCount} 场景图 · {subjectAssetCount || cutoutCount} 主体素材 · {d.selectedFrames.size}/{frames.length} 入编排 · {completedVideos.length} 已完成
+              {cleanedCount} 已清洗 · {sceneAssetCount} 场景图 · {subjectAssetCount || cutoutCount} 主体素材 · {selectedFrameCount}/{frames.length} 入编排 · {completedVideos.length} 已完成
             </>
           ) : (
-            "解析后这里展示关键帧、元素和视频任务；具体处理仍在点击后的工作台完成。"
+            "解析后这里变成素材准备看板：先审关键帧，再生成场景图和主体资产包。"
           )}
         </div>
       </NodeShell>