fix(canvas): 分镜脚本 - 补齐拆镜输出契约并让镜头数与每镜时长生效

web/src/lib/canvas/canvas-project-domain.ts

@@ -120,6 +120,22 @@ export function storyboardRowsOutputContract(requirement: string) {
     ].join("\n");
 }
 
+// 分镜文本任务实际发出的 prompt：技能上下文只描述工作流、不约束输出结构，画布节点和章节分镜又都没有
+// metadata.promptTemplateOperation（后端只在有该字段时编译服务端模板），所以契约必须由调用方显式拼上，
+// 否则模型按自然语言习惯返回 Markdown 表格或直接续写正文，解析阶段拿不到 rows。
+// 镜头数与每镜时长只能写进契约文本：后端 canvasGenerationInput 不读 input.shotCount / shotDurationSeconds，
+// 仅放进 input 时节点上的「N 镜」「每镜 N 秒」对模型不可见。
+export function withStoryboardOutputContract(prompt: string, options: { shotCount?: number; shotDurationSeconds?: number } = {}) {
+    const shotCount = Math.trunc(Number(options.shotCount) || 0);
+    const shotDurationSeconds = Math.trunc(Number(options.shotDurationSeconds) || 0);
+    const requirement = [
+        shotCount > 0 ? `全片共 ${shotCount} 个镜头` : "",
+        shotDurationSeconds > 0 ? `每镜约 ${shotDurationSeconds} 秒，durationSeconds 填写该值` : "",
+        "每个镜头必须能独立用于生成首帧图片和镜头视频。",
+    ].filter(Boolean).join("；");
+    return [prompt, storyboardRowsOutputContract(requirement)].join("\n\n");
+}
+
 // 文本任务的 result_json 外层是 {mode, reasoning, text} 包装，分镜表在 text 字段的内层 JSON 字符串里；
 // 历史模板路径可能直接落 {title, rows}。两种结构都要能解析，内层再做代码块剥离与首尾大括号截取兜底，
 // 模型偶发违反契约（包 ```json 或混入解释文字）时仍能取出分镜表。

web/src/pages/canvas/use-canvas-storyboard.ts

@@ -17,6 +17,7 @@ import {
     expandStoryboardTextMentions,
     storyboardRowsFromTask,
     storyboardPromptTemplateMetadata,
+    withStoryboardOutputContract,
 } from "@/lib/canvas/canvas-project-domain";
 import { buildNodeMentionReferences } from "@/lib/canvas/canvas-resource-references";
 import { buildStoryboardAssetCatalog } from "@/lib/canvas/canvas-storyboard-assets";
@@ -168,16 +169,18 @@ export function useCanvasStoryboard({
             assertCurrent();
             setNodes((current) => current.map((node) => node.id === nodeId ? { ...node, metadata: { ...node.metadata, composerContent: prompt, status: NODE_STATUS_LOADING, taskStage: "正在创建任务", taskProgress: 0, errorDetails: undefined, ...skillExecution.metadata } } : node));
             assertCurrent();
+            // 输入框内容只是剧情，拆镜要求与镜头数、每镜时长都必须随契约一起发出去。
+            const storyboardPrompt = withStoryboardOutputContract(skillExecution.prompt, { shotCount: requestedShotCount, shotDurationSeconds });
             const request = {
                 projectId,
                 type: "canvas_text",
                 operation: "storyboard",
-                prompt: skillExecution.prompt,
+                prompt: storyboardPrompt,
                 model: generationConfig.model,
                 ...(logicalModelIDForConfig(generationConfig) ? { logicalModelId: logicalModelIDForConfig(generationConfig) } : {}),
                 input: {
                     mode: "text",
-                    prompt: skillExecution.prompt,
+                    prompt: storyboardPrompt,
                     canvasAssets: buildStoryboardAssetCatalog(nodesRef.current),
                     requirements: "输出可直接编辑并用于批量生成图片和视频的分镜表。",
                     projectStyle: storyboardContext.projectStyle,

web/src/pages/projects/detail/project-chapter-ai.ts

@@ -1,6 +1,6 @@
 import { runBackendCanvasGenerationTask } from "@/lib/canvas/canvas-project-generation";
 import { PromptTemplateOperation, promptTemplateTaskPlaceholder } from "@/lib/prompts";
-import { storyboardRowsFromTask, storyboardRowsOutputContract } from "@/lib/canvas/canvas-project-domain";
+import { storyboardRowsFromTask, withStoryboardOutputContract } from "@/lib/canvas/canvas-project-domain";
 import { parseChapterAssetBreakdown, type ChapterAssetBreakdown } from "@/lib/canvas/chapter-asset-breakdown";
 import { parseCharacterBreakdown } from "@/lib/canvas/canvas-character-reference";
 import { backendProviderConfig, parseBackendGenerationResult } from "@/services/api/generation-task";
@@ -91,10 +91,7 @@ export async function generateChapterStoryboard(input: ChapterStoryboardGenerati
         selectedSkillIds: input.selectedSkillIds,
     });
     // 技能上下文只描述工作流，不会约束输出结构；必须显式带上输出契约，否则模型返回 Markdown 表格。
-    const prompt = [
-        skillExecution.prompt,
-        storyboardRowsOutputContract("每个镜头必须能独立用于生成首帧图片和镜头视频。"),
-    ].join("\n\n");
+    const prompt = withStoryboardOutputContract(skillExecution.prompt);
     const task = await createGenerationTask({
         projectId: input.projectId,
         type: "canvas_text",