@ninkch/dsh-infinite-canvas 0.17.9 → 0.17.10

package/README.md

@@ -17,11 +17,12 @@ Everything runs locally in your Harness. Bring your own generation APIs, or use
 ```mermaid
 flowchart LR
     A[Script node<br>AI screenplay] --> B[Storyboard node<br>AI shot list]
-    B --> C[Expand to canvas<br>1 shot = 1 node group]
-    C --> D[Image gen<br>multi-reference]
-    D --> E[Video gen<br>first frame + refs]
-    E --> F[Final cut<br>transitions · subtitles · voices]
-    F --> G[Export<br>WebM preview / FilmCraft MP4]
+    B --> C[Expand to canvas]
+    C --> D[Multi-shot video gen<br>all shots in ONE prompt]
+    D --> E[Final cut<br>transitions · subtitles · voices]
+    E --> F[Export<br>WebM preview / FilmCraft MP4]
+    A --> G[Reference image gen<br>characters · scenes · props]
+    G --> D
 ```
 
 ## Highlights
@@ -33,13 +34,16 @@ A per-project **series bible** (style baseline + character/scene/prop cards) is
 Reference images connect as an ordered, visible list — `[图1] 角色：Lin Ning`, `[图2] 场景：The Old House` — and that exact order is what reaches the model. Multi-character frames stay unambiguous, and Volcano Engine **Seedance**'s native `[图N]` convention is supported end to end (first frame, up to 4 reference images, official task API).
 
 ### From shots to a watchable cut
-The **Final Cut node** stitches shot videos in order — crossfade or black transitions, burned-in subtitles, original audio and TTS voices mixed in — and exports a preview cut instantly. For real post-production, one click writes a **FCP7 XML timeline + SRT** to disk and hands the render to [ArtCraft](#artcraft-bridge) (FilmCraft) for a true MP4.
+**One storyboard = one multi-shot video generation.** Expanding a storyboard wires it to a single video node: every shot (order + per-shot duration) is composed into one multi-shot prompt and your video model renders the whole clip in one call — the model makes the hard cuts, no per-shot stitching seams. The **Final Cut node** then stitches segments in order — crossfade or black transitions, burned-in subtitles, original audio and TTS voices mixed in — and exports a preview cut instantly. For real post-production, one click writes a **FCP7 XML timeline + SRT** to disk and hands the render to [ArtCraft](#artcraft-bridge) (FilmCraft) for a true MP4.
 
 ### An editor-grade canvas
 Pan, zoom, marquee select, alignment guides, groups that fold, a minimap, 60-step undo/redo, per-node media history (last 8 results), mask-based inpainting, and middle-click everything — the canvas part should feel like an editor, not a chat sidebar.
 
 ### Bring your own models
-Text models come from Harness itself (no extra keys). Image, video and TTS backends are **multi-config** — define as many endpoints as you want, mark a default, and pick per node. No backend configured? Demo mode synthesizes placeholder images and animated videos locally so the entire pipeline is testable offline.
+Text models come from Harness itself (no extra keys, no per-plugin parameters). Image, video and TTS backends are **multi-config** — define as many endpoints as you want, mark a default, and pick per node. Every gateway call is relayed through the local Harness host process, so **any vendor works** (no CORS headaches). No backend configured? Demo mode synthesizes placeholder images and animated videos locally so the entire pipeline is testable offline.
+
+### A portable data root
+Set a **data root folder** once and everything lands there, organized by **project name**: nodes and the series bible in `project.json`, the asset index in `assets.json`, media under `image/ video/ audio/ text/`, and Craft post-production under `craft/`. Move to a new machine by copying the folder and clicking **Import from data root**.
 
 ## Quick start
 
@@ -57,28 +61,29 @@ Text models come from Harness itself (no extra keys). Image, video and TTS backe
 |---|---|---|
 | Text | Harness LLM service | Script & storyboard writers; pick any configured model, or override per node |
 | Image | OpenAI-compatible `/images/generations`, multi-reference via `/images/edits` | Size (1K/2K/4K) × ratio; mask inpainting via `edits` + PNG mask |
-| Video | Generic submit+poll, or **Volcano Seedance official tasks** | duration 4–10s, 480P/720P/1080P, ratio, first frame, up to 4 reference images |
+| Video | Generic submit+poll, or **Volcano Seedance official tasks** | duration up to 15 s, 480P/720P/1080P, ratio, first frame, up to 4 reference images |
 | TTS | OpenAI-compatible `/audio/speech` | Per-shot voices mixed into the final cut |
 
+All gateway calls are relayed by the local Harness host process (no browser CORS limits) — any vendor endpoint with an OpenAI-style shape works.
+
 ## ArtCraft bridge
 
-When the companion [`dsh-artcraft`](../dsh-artcraft/README.md) plugin is installed, the canvas talks to the local **PhotoCraft / FilmCraft / EffectCraft** toolchain over same-origin HTTP: assets sync to a per-project data folder, images get headless PhotoCraft grading (single or batch), and final cuts export as FCP7 XML + SRT for FilmCraft to render as MP4. Status lights, graceful degradation and a full contract live in [`docs/artcraft-handoff.md`](./docs/artcraft-handoff.md).
+When the companion [`dsh-artcraft`](../dsh-artcraft/README.md) plugin is installed, the canvas talks to the local **Craft family** (PhotoCraft, FilmCraft, EffectCraft, LightCraft, VectorCraft, DesignCraft, … — every bin the Craft suite reports, currently 12) over same-origin HTTP. Point ArtCraft's data root at the same folder as the canvas data root and all post-production lands in `projects/<project>/craft/` next to your other data: assets sync per project, images get headless PhotoCraft grading (single or batch), and final cuts export as FCP7 XML + SRT for FilmCraft to render as MP4. A status card in Settings shows one light per craft with an n/n ready summary.
 
 ## Publishing (maintainers)
 
-Testing funnel, beta tags, pack validation and the full release runbook live in [`docs/release.md`](docs/release.md).
+The maintainer runbook (red lines, branch model, versioning) lives in [`RELEASING.md`](RELEASING.md); the operations detail in [`docs/release.md`](docs/release.md); released changes in [`CHANGELOG.md`](CHANGELOG.md) — **GitHub Release notes are written in English** and copied from the changelog.
 
 Publishing runs through **npm Trusted Publishing** (OIDC) — no npm token exists anywhere:
 
-- **Release flow**: create a GitHub Release (tag `v0.17.0`, …) → Actions publishes that version automatically.
-- **Manual flow**: repo → Actions → *publish* → Run workflow → type a version.
-- The `publish.yml` workflow resolves the version from the release tag (or your input) and publishes with a `provenance` badge.
+- **Release flow**: dev tested & approved → merge to `main` → create a GitHub Release (tag `v0.17.9`, …) → Actions publishes that version automatically.
+- The `publish.yml` workflow guards on the release commit being on `main` and the tag matching `package.json`, then publishes with a `provenance` badge.
 
 ## Data & privacy
 
-- Canvas documents, projects, settings and the series bible live in browser `localStorage` (`dsh-ic:*`); assets live in IndexedDB (`dsh-infinite-canvas`) — **or on disk** once a data root is set in Settings (per-project folders under the configured root, served by the plugin backend with HTTP Range support for video seeking).
+- Canvas documents, projects and settings live in browser `localStorage` (`dsh-ic:*`); asset metadata lives in IndexedDB (`dsh-infinite-canvas`) while media files persist **on disk** once a data root is set (per-project folders by project name, served by the plugin backend with HTTP Range support for video seeking).
 - AI text generation runs through the Harness host process (your configured models, no extra keys).
-- Image/video/TTS calls go directly from the browser to the endpoints you configure — mind CORS and key handling.
+- Image/video/TTS gateway calls are relayed by the local Harness host process (avoids browser CORS limits); your endpoints and keys stay local.
 - Nothing is uploaded anywhere by the plugin itself.
 
 ## Requirements & limitations

package/assets/canvas-studio/SKILL.md

@@ -35,7 +35,12 @@ description: 用无限画布（Infinite Canvas）做短剧/视频创作时使用
 6. **删除是破坏性操作。** 先 `canvas_list_projects` 确认 id，且仅在用户明确要求时执行
    （`canvas_delete_node` / `canvas_delete_project`）。
 7. **完成后让用户看到结果**：调用 `canvas_open({ projectId })`，画布会在数秒内自动跳转。
-8. **汇报固定要素**：项目名 + 项目 id + 节点名/id + 成稿摘要（若已生成）；有错误如实说明。
+8. **一致性告警必须闭环。** 分镜节点生成/更新后，`canvas_node_status` 会返回 `warnings`——
+   分镜中出现但**未在设定集登记**或**已登记但描述为空**的实体清单（含画面/动作文本里的 @提及）。
+   只要 `warnings` 非空：**逐条** `canvas_update_bible` 补全（名称与告警**逐字一致**、描述写实可生成），
+   再生成参考图；**绝不忽略、绝不跳过、绝不只补一部分**。设定集更新后告警会自动重算
+   （几秒内随文档同步刷新），以最新一次 `canvas_node_status` 为准。
+9. **汇报固定要素**：项目名 + 项目 id + 节点名/id + 成稿摘要（若已生成）；有错误如实说明。
 
 ## 二、标准工作流（对话填提示词+连线+触发，生成由节点完成）
 

package/client.js

@@ -299,6 +299,9 @@ window.__ModuleLoader__.load({
       'shotsUnit': '镜',
       'savedOk': '已存为资产',
       'warnUnregistered': '未在设定集/资产中登记',
+      'warnEmptyDesc': '已登记但描述为空（待补全）',
+      'catMention': '提及',
+      'assetCategoriesFixed': '已按命名规范纠正 {n} 个资产的分类',
       'consistencyTitle': '一致性检查',
       'restartHint': '插件后台尚未加载最新版本：请重启一次 DeepSeek Harness 后重试（仅首次更新需要）',
       'rpcFail': '请求失败',
@@ -636,6 +639,9 @@ window.__ModuleLoader__.load({
       'shotsUnit': 'shots',
       'savedOk': 'Saved as asset',
       'warnUnregistered': 'not in bible/assets',
+      'warnEmptyDesc': 'registered but description is empty (fill it in)',
+      'catMention': 'Mention',
+      'assetCategoriesFixed': 'Re-categorized {n} assets by naming convention',
       'consistencyTitle': 'Consistency check',
       'restartHint': 'The plugin backend is not loaded yet: restart DeepSeek Harness once and retry (first update only)',
       'rpcFail': 'Request failed',
@@ -1270,6 +1276,17 @@ window.__ModuleLoader__.load({
     function entityCategoryOf(kind) {
       return kind === 'character' ? 'character' : kind === 'scene' ? 'scene' : kind === 'prop' ? 'prop' : 'general';
     }
+    /** 按命名规范前缀推导资产品类：角色-/场景-/道具-/剧本-/分镜- → 对应分类（无前缀返回 ''）。
+     *  用于 @引用未命中时（如对话侧直接写角色名）的兜底分类。 */
+    function categoryFromName(name) {
+      const n = String(name || '').trim();
+      if (/^角色[-_／/]/.test(n)) return 'character';
+      if (/^场景[-_／/]/.test(n)) return 'scene';
+      if (/^道具[-_／/]/.test(n)) return 'prop';
+      if (/^剧本[-_／/]/.test(n)) return 'script';
+      if (/^分镜(表)?[-_／/]/.test(n)) return 'storyboard';
+      return '';
+    }
     /** 在设定集里按名称找实体（角色/场景/道具），携带其链接的资产 id。 */
     function matchBibleEntity(name, bible) {
       const want = String(name || '').trim().toLowerCase();
@@ -1610,6 +1627,27 @@ window.__ModuleLoader__.load({
       }
       return { migrated: done, total: targets.length };
     }
+    /** 启动自检：把分类为「通用」但命名符合规范前缀（角色-/场景-/道具-/剧本-/分镜-）的存量资产纠正分类。 */
+    async function fixGeneralCategories() {
+      try {
+        const all = await idbAll();
+        let fixed = 0;
+        for (const rec of all) {
+          if (!rec || !rec.id || (rec.kind !== 'image' && rec.kind !== 'video')) continue;
+          if ((rec.category || 'general') !== 'general') continue;
+          const c = categoryFromName(rec.name);
+          if (!c) continue;
+          const next = { ...rec, category: c };
+          await idbPut(next);
+          mirrorAssetIndex(next);
+          fixed++;
+        }
+        if (fixed) {
+          update((s) => { s.assetsTick++; });
+          toast(T('assetCategoriesFixed').replace('{n}', String(fixed)));
+        }
+      } catch { /* 静默：失败不影响主流程，下次启动再试 */ }
+    }
     /** 数据根目录就绪后自动把浏览器 IndexedDB 里的存量资产搬到磁盘（替代原"迁移历史资产"按钮，全自动无需手动触发）。 */
     async function autoMigrateAssetsToRoot() {
       if (!S.dataRoot) return;
@@ -3059,34 +3097,46 @@ window.__ModuleLoader__.load({
         duration: Number.isFinite(Number(s.duration)) ? clamp(Math.round(Number(s.duration)), 1, 15) : 3,
       }));
     }
-    /** 本地一致性检查：分镜里出现的角色/场景/道具名是否都能在设定集或图片资产中找到。 */
+    /** 本地一致性检查：分镜里出现的角色/场景/道具（含 visual/motion 文本里的 @提及）是否都在
+     *  设定集/资产中登记；两级告警——「未登记」与「已登记但描述为空（待补全）」。 */
     async function checkShotConsistency(shots) {
       const bible = loadBible(S.activeId);
-      const known = new Set();
-      const addName = (n) => { const q = String(n || '').trim().toLowerCase(); if (q) known.add(q); };
-      (bible.characters || []).forEach((c) => addName(c.name));
-      (bible.scenes || []).forEach((c) => addName(c.name));
-      (bible.props || []).forEach((c) => addName(c.name));
-      try {
-        const recs = (await idbByProject(S.activeId)).concat(await idbByProject(GENERAL_PID));
-        recs.forEach((r) => addName(r.name));
-      } catch { /* 忽略 */ }
+      const entries = [];
+      const addEntry = (kind, name, desc) => {
+        const nm = String(name || '').trim();
+        if (nm) entries.push({ kind, name: nm, desc: String(desc || ''), lower: nm.toLowerCase() });
+      };
+      (bible.characters || []).forEach((c) => addEntry(T('catCharacter'), c && c.name, c && c.desc));
+      (bible.scenes || []).forEach((c) => addEntry(T('catScene'), c && c.name, c && c.desc));
+      (bible.props || []).forEach((c) => addEntry(T('catProp'), c && c.name, c && c.desc));
+      const known = new Set(entries.map((e) => e.lower));
+      try { ((await idbByProject(S.activeId)).concat(await idbByProject(GENERAL_PID))).forEach((r) => { const q = String(r && r.name || '').trim().toLowerCase(); if (q) known.add(q); }); } catch { /* 忽略 */ }
       const warnings = [];
+      const seen = new Set();
+      const warn = (line) => { if (!seen.has(line) && warnings.length < 12) { seen.add(line); warnings.push(line); } };
+      const check = (title, kind, name) => {
+        const q = String(name || '').trim().toLowerCase();
+        if (!q) return;
+        const exact = entries.find((e) => e.lower === q);
+        if (exact) {
+          if (!exact.desc.trim()) warn(`${title} · ${kind}「${exact.name}」${T('warnEmptyDesc')}`);
+          return;
+        }
+        for (const k of known) { if (k.includes(q) || q.includes(k)) return; }
+        warn(`${title} · ${kind}「${String(name).trim()}」${T('warnUnregistered')}`);
+      };
       for (const sh of shots) {
-        const items = [
-          ...sh.characters.map((n) => [n, T('catCharacter')]),
-          [sh.scene, T('catScene')],
-          ...sh.props.map((n) => [n, T('catProp')]),
-        ];
-        for (const [nm, kind] of items) {
-          const q = String(nm || '').trim().toLowerCase();
-          if (!q) continue;
-          if (![...known].some((k) => k.includes(q) || q.includes(k))) {
-            warnings.push(`${sh.title} · ${kind}「${nm}」${T('warnUnregistered')}`);
-          }
+        const title = String(sh.title || '').trim() || T('shotPrefix');
+        for (const nm of (sh.characters || [])) check(title, T('catCharacter'), nm);
+        check(title, T('catScene'), sh.scene);
+        for (const nm of (sh.props || [])) check(title, T('catProp'), nm);
+        // visual/motion 文本里的 @提及 同样纳入检查（数组外的实体也要登记，杜绝检查漏）
+        const text = String(sh.visual || '') + ' ' + String(sh.motion || '');
+        for (const m of (text.match(/@[^\s@，。；、！？:：（）"'）)]+/g) || [])) {
+          check(title, T('catMention'), m.slice(1).replace(/[）)]+$/, ''));
         }
       }
-      return warnings.slice(0, 12);
+      return warnings;
     }
     /** 分镜生成后：把设定集里没有的角色/场景/道具自动补进设定集（描述留空待用户补充）。 */
     async function autoRegisterBibleEntries(shots) {
@@ -3710,14 +3760,14 @@ window.__ModuleLoader__.load({
       flushSave();
       try {
         const shots = await aiStoryboardShots(st, n.data.requirement, n.data.count || 6, n.data.model, (kind, d) => aiStreamAppend(n.id, kind, d));
+        // 先自动补录缺失条目（空描述），再算一致性告警：否则告警标志的条目其实马上就被自动补录了，"未登记"语义会误导用户与对话侧
+        try { await autoRegisterBibleEntries(shots); } catch { /* 补录失败不影响分镜 */ }
         const warnings = await checkShotConsistency(shots);
         update((s) => {
           const nd = s.graph.nodes[n.id];
           if (nd) { nd.data.shots = shots; nd.data.aiStatus = 'idle'; nd.data.warnings = warnings; nd.data.status = 'done'; nd.data.error = null; }
           s.assetsTick++;
         });
-        // 分镜里出现的新角色/场景/道具 → 自动补进设定集（描述留空待补充），保证设定集始终是唯一事实源
-        try { await autoRegisterBibleEntries(shots); } catch { /* 补录失败不影响分镜 */ }
         await saveTextAssetRecord(S.activeId, {
           name: (String(n.data.label || '').trim() || T('storyboardAssetPrefix')) + ' ' + nowStamp(),
           text: storyboardAssetText(shots),
@@ -3824,7 +3874,8 @@ window.__ModuleLoader__.load({
             entity: entities.length ? entities[0].name : undefined,
             entities: entities.map((e) => e.name),
             nodeLabel: (n.data && n.data.label) || '',
-            category: entities.length ? entityCategoryOf(entities[0].kind) : 'general',
+            // 分类：@引用命中的实体品类优先；无引用时按命名规范前缀（角色-/场景-/道具-）兜底
+            category: entities.length ? entityCategoryOf(entities[0].kind) : (categoryFromName(n.data.label) || 'general'),
           });
           commitGenResult(n.id, rec.id, demo, 'image');
         } else {
@@ -3842,7 +3893,7 @@ window.__ModuleLoader__.load({
             entity: entities.length ? entities[0].name : undefined,
             entities: entities.map((e) => e.name),
             nodeLabel: (n.data && n.data.label) || '',
-            category: entities.length ? entityCategoryOf(entities[0].kind) : 'general',
+            category: entities.length ? entityCategoryOf(entities[0].kind) : (categoryFromName(n.data.label) || 'general'),
           });
           commitGenResult(n.id, rec.id, demo, 'video');
           // 尾帧提取（通用）：供 last 端口输出与链式下一镜使用
@@ -7249,8 +7300,8 @@ window.__ModuleLoader__.load({
           probeCraft();
           // 检查更新（Host RPC；Host 旧版未带此方法时静默）
           checkUpdate(true);
-          // 数据根目录（Host 旧版静默降级为 IndexedDB 模式）；就绪后自动搬存量资产到磁盘，再对齐索引
-          refreshDataRoot().then(() => autoMigrateAssetsToRoot()).then(() => syncAssetIndexToHost()).catch(() => { /* 静默 */ });
+          // 数据根目录（Host 旧版静默降级为 IndexedDB 模式）；就绪后自动搬存量资产到磁盘、纠正分类，再对齐索引
+          refreshDataRoot().then(() => autoMigrateAssetsToRoot()).then(() => fixGeneralCategories()).then(() => syncAssetIndexToHost()).catch(() => { /* 静默 */ });
           // 对话 ↔ 画布同步（启动一次 + 3s 轮询；旧 Host 自动停轮询）
           hostSyncTick().then(() => pushAllProjectsToHost());
           const hostSyncTimer = setInterval(() => hostSyncTick(), 3000);

package/index.js

@@ -602,7 +602,7 @@ async function projectsPut(args) {
     tombstoneAt: pruneStamp(tombstoneAt, keptTomb),
     edgeTombstoneAt: pruneStamp(edgeTombstoneAt, keptEdgeTomb),
   };
-  // 设定集（世界观/风格/角色/场景/道具）：随文档一起同步，对话侧可读写
+  // 设定集随文档一起同步（世界观/风格/角色/场景/道具）：对话侧可读写
   if (doc.bible !== undefined) {
     out.bible = normalizeBible(doc.bible);
     out.bibleUpdatedAt = Date.now();
@@ -610,6 +610,18 @@ async function projectsPut(args) {
     out.bible = prev.bible;
     out.bibleUpdatedAt = prev.bibleUpdatedAt || 0;
   }
+  // 设定集有更新时重算全部分镜表的一致性告警（客户端侧设定集编辑也走这条路径）
+  if (out.bible !== undefined) {
+    try {
+      const assetNames = readAssetIndexNames(id);
+      for (const nd of Object.values(merged)) {
+        if (nd && nd.type === 'storyboard' && Array.isArray(nd.data && nd.data.shots) && nd.data.shots.length) {
+          const nextW = storyboardConsistencyWarnings(out.bible, nd.data.shots, assetNames);
+          if (JSON.stringify(nextW) !== JSON.stringify(nd.data.warnings || [])) nd.data.warnings = nextW;
+        }
+      }
+    } catch { /* 重算失败不影响保存 */ }
+  }
   mkdirSync(PROJECTS_DIR, { recursive: true });
   writeIcDoc(id, out); // 统一走原子写 + 数据目录镜像（此前直写绕过了镜像与原子性）
   return { ok: true, id, updatedAt: out.updatedAt };
@@ -847,6 +859,8 @@ function canvasWriteNode(args) {
         }));
         node.data.aiStatus = 'idle';
         node.data.aiError = null;
+        // 写入分镜时同步计算一致性告警：node_status 能立刻返回 warnings（对话侧据此补全设定集）
+        try { node.data.warnings = storyboardConsistencyWarnings(doc.bible || {}, node.data.shots, readAssetIndexNames(id)); } catch { node.data.warnings = []; }
       } else {
         // 二选一强约束（不可用 overwrite 绕过）：已连线到上游剧本的分镜表，内容只从连线取——
         // 两套来源并存会让生成结果不可预期。如需改用粘贴，请先删除该连线或重建分镜表节点。
@@ -983,6 +997,16 @@ async function canvasUpdateBible(args) {
   doc.bible = normalizeBible(next);
   doc.bibleUpdatedAt = Date.now();
   doc.updatedAt = Date.now();
+  // 设定集更新后重算全部分镜表的一致性告警：否则节点上挂着过期告警，用户与对话侧都会被误导
+  try {
+    const assetNames = readAssetIndexNames(id);
+    for (const nd of Object.values(doc.nodes || {})) {
+      if (nd && nd.type === 'storyboard' && Array.isArray(nd.data && nd.data.shots) && nd.data.shots.length) {
+        const nextW = storyboardConsistencyWarnings(doc.bible, nd.data.shots, assetNames);
+        if (JSON.stringify(nextW) !== JSON.stringify(nd.data.warnings || [])) nd.data.warnings = nextW;
+      }
+    }
+  } catch { /* 重算失败不影响设定集写入 */ }
   writeIcDoc(id, doc);
   markDocChanged(id);
   return { projectId: id, bible: doc.bible, bibleUpdatedAt: doc.bibleUpdatedAt };
@@ -1108,6 +1132,57 @@ async function canvasRunNode(args) {
   return { ok: true, projectId: pid, nodeId, type: node.type, label: (node.data && node.data.label) || '' };
 }
 /** 读取节点运行状态与产物（画布执行完成后会把结果推回 Host 文档）。 */
+/** 分镜一致性检查（与客户端 checkShotConsistency 同规则，Host 供状态查询/设定集更新后重算）：
+ *  检查 shots 的角色/场景/道具数组 **以及 visual/motion 文本里的 @提及** 是否都在设定集登记；
+ *  两级告警：「未登记」（设定集/资产都没有）与「已登记但描述为空」（待补全）。 */
+function storyboardConsistencyWarnings(bible, shots, extraNames) {
+  const entries = [];
+  const addEntry = (kind, name, desc) => {
+    const nm = String(name || '').trim();
+    if (nm) entries.push({ kind, name: nm, desc: String(desc || ''), lower: nm.toLowerCase() });
+  };
+  (bible.characters || []).forEach((c) => addEntry('角色', c && c.name, c && c.desc));
+  (bible.scenes || []).forEach((c) => addEntry('场景', c && c.name, c && c.desc));
+  (bible.props || []).forEach((c) => addEntry('道具', c && c.name, c && c.desc));
+  const known = new Set(entries.map((e) => e.lower).concat((extraNames || []).map((n) => String(n).trim().toLowerCase()).filter(Boolean)));
+  const warnings = [];
+  const seen = new Set();
+  const warn = (line) => { if (!seen.has(line) && warnings.length < 12) { seen.add(line); warnings.push(line); } };
+  const check = (title, kind, name) => {
+    const q = String(name || '').trim().toLowerCase();
+    if (!q) return;
+    const exact = entries.find((e) => e.lower === q);
+    if (exact) {
+      if (!exact.desc.trim()) warn(`${title} · ${kind}「${exact.name}」已登记但描述为空（待补全）`);
+      return;
+    }
+    for (const k of known) { if (k.includes(q) || q.includes(k)) return; }
+    warn(`${title} · ${kind}「${String(name).trim()}」未在设定集中登记`);
+  };
+  for (const sh of shots || []) {
+    const title = String((sh && sh.title) || '').trim() || '分镜';
+    for (const nm of (sh.characters || [])) check(title, '角色', nm);
+    check(title, '场景', sh && sh.scene);
+    for (const nm of (sh.props || [])) check(title, '道具', nm);
+    // visual/motion 文本里的 @提及 同样纳入检查（数组外的实体也要登记，杜绝检查漏）
+    const text = String((sh && sh.visual) || '') + ' ' + String((sh && sh.motion) || '');
+    for (const m of (text.match(/@[^\s@，。；、！？:：（）"'）)]+/g) || [])) {
+      check(title, '提及', m.slice(1).replace(/[）)]+$/, ''));
+    }
+  }
+  return warnings;
+}
+/** 从数据目录的项目资产索引里取资产名（与客户端"设定集/资产中登记"的判定口径对齐）。 */
+function readAssetIndexNames(pid) {
+  const root = dataRoot();
+  if (!root || !pid) return [];
+  try {
+    const entry = readDataRootMeta().projects.find((p) => p.id === pid);
+    if (!entry) return [];
+    const list = JSON.parse(readFileSync(path.join(root, 'projects', entry.folder, 'assets.json'), 'utf8'));
+    return Array.isArray(list) ? list.map((x) => x && x.name).filter(Boolean) : [];
+  } catch { return []; }
+}
 async function canvasNodeStatus(args) {
   const pid = String((args && args.projectId) || '');
   const nodeId = String((args && args.nodeId) || '');
@@ -1122,6 +1197,8 @@ async function canvasNodeStatus(args) {
     output: String((n.data && n.data.output) || '').slice(0, 6000),
     idea: String((n.data && n.data.idea) || '').slice(0, 2000),
     assetId: (n.data && n.data.assetId) || null,
+    // 分镜表：一致性告警随状态返回——对话侧据此逐条补全设定集（缺失则要自动补充，绝不忽略）
+    warnings: Array.isArray(n.data && n.data.warnings) ? n.data.warnings : [],
   };
 }
 async function canvasOpen(args) {
@@ -1327,7 +1404,7 @@ async function updateInstall() {
 export const inject = ['webServer', 'tools'];
 
 // 仅供诊断 / 自动化测试（不属于 RPC 面，运行时行为不受影响）
-export const _internals = { updateCheck, updateInstall, npmLatestVersion, cmpVer, findPnpm, getProfileDir: () => PROFILE_DIR, getVersion: () => VERSION, dataRoot, setDataRoot, dataRootMigrate, dataRootImport, ensureProjectFolder, physicalRel, mirrorProjectToDataRoot, writeAssetIndex, removeAssetIndexEntry, readDataRootMeta, removeDataRootProject, deletedProjectIds, addDeletedProjectId, clearDeletedProjectId, fetchProxyForward, safeJoin, mimeOf, pipeFile, getIcConfigPath: () => CONFIG_PATH, canvasCreateProject, canvasWriteNode, canvasConnect, canvasDeleteProject, canvasDeleteNode, canvasRunNode, canvasNodeStatus, canvasGetBible, canvasUpdateBible, canvasCreateNode, canvasOpen, projectsList, projectsPut, projectsGet, projectsDelete, projectsSync };
+export const _internals = { updateCheck, updateInstall, npmLatestVersion, cmpVer, findPnpm, getProfileDir: () => PROFILE_DIR, getVersion: () => VERSION, dataRoot, setDataRoot, dataRootMigrate, dataRootImport, ensureProjectFolder, physicalRel, mirrorProjectToDataRoot, writeAssetIndex, removeAssetIndexEntry, readDataRootMeta, removeDataRootProject, deletedProjectIds, addDeletedProjectId, clearDeletedProjectId, fetchProxyForward, storyboardConsistencyWarnings, readAssetIndexNames, safeJoin, mimeOf, pipeFile, getIcConfigPath: () => CONFIG_PATH, canvasCreateProject, canvasWriteNode, canvasConnect, canvasDeleteProject, canvasDeleteNode, canvasRunNode, canvasNodeStatus, canvasGetBible, canvasUpdateBible, canvasCreateNode, canvasOpen, projectsList, projectsPut, projectsGet, projectsDelete, projectsSync };
 
 function resolveModel(providers, want) {
   const byProvider = want && want.provider
@@ -1730,7 +1807,7 @@ export function apply(ctx) {
         });
         reg({
           name: 'canvas_node_status',
-          description: '查询画布节点的运行状态与产物（状态、错误、剧本成稿内容、生成出的资产 id）。用于确认 canvas_run_node 是否完成。',
+          description: '查询画布节点的运行状态与产物（状态、错误、剧本成稿内容、生成出的资产 id）。分镜表节点还会返回 warnings（一致性告警：未登记/描述为空的实体清单）——非空时必须逐条 canvas_update_bible 补全，绝不忽略。用于确认 canvas_run_node 是否完成。',
           parameters: {
             type: 'object',
             properties: {
@@ -1741,7 +1818,7 @@ export function apply(ctx) {
             additionalProperties: false,
           },
           output: {
-            schema: { type: 'object', properties: { ok: { type: 'boolean' }, text: { type: 'string' }, status: { type: 'string' }, assetId: { type: 'string' } }, required: ['ok', 'text'] },
+            schema: { type: 'object', properties: { ok: { type: 'boolean' }, text: { type: 'string' }, status: { type: 'string' }, assetId: { type: 'string' }, warnings: { type: 'array', items: { type: 'string' } } }, required: ['ok', 'text'] },
             render: (args, value) => [{ type: 'text', text: String(value.text || '') }],
           },
           timeoutMs: 30000,
@@ -1752,7 +1829,13 @@ export function apply(ctx) {
             if (s.assetId) lines.push('产物资产: ' + s.assetId);
             if (s.idea) lines.push('提示词: ' + s.idea.slice(0, 400));
             if (s.output) lines.push('成稿:\n' + s.output);
-            return { ok: true, status: s.status, assetId: s.assetId || undefined, text: lines.join('\n') };
+            // 一致性告警：结构化返回 + 渲染进文本——对话侧据此逐条补全设定集，绝不忽略
+            const warnings = Array.isArray(s.warnings) ? s.warnings : [];
+            if (warnings.length) {
+              lines.push('一致性告警（' + warnings.length + ' 条，须逐条 canvas_update_bible 补全，名称逐字一致）:');
+              warnings.forEach((w) => lines.push(' · ' + w));
+            }
+            return { ok: true, status: s.status, assetId: s.assetId || undefined, warnings: warnings.length ? warnings : undefined, text: lines.join('\n') };
           },
         });
         reg({

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ninkch/dsh-infinite-canvas",
-  "version": "0.17.9",
+  "version": "0.17.10",
   "description": "Node-based short-drama studio for DeepSeek Harness: script → storyboard → assets → shots → voiced, subtitled final cut on an infinite canvas.",
   "keywords": [
     "dsh",