From be06caaa9fd6cc716b65ffb6495209a86ea93ea8 Mon Sep 17 00:00:00 2001
From: GouJ <hefangjay@qq.com>
Date: Tue, 22 Sep 2026 19:56:12 +0800
Subject: [PATCH] =?UTF-8?q?feat:=20=E5=90=8C=E6=AD=A5=E6=9C=80=E6=96=B0?=
 =?UTF-8?q?=E7=94=9F=E5=9B=BE=E4=B8=8E=E8=BA=AB=E4=BB=BD=E8=83=BD=E5=8A=9B?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 docs/api-contract-sync.md                     |   7 ++
 docs/backend-coverage.md                      |  12 ++
 src/features/production/api.ts                |   4 +
 .../production/components/KeyframeDialog.vue  |  72 +++++++++++-
 src/features/production/types.ts              |  47 ++++++++
 .../subject-identity/SubjectIdentityPage.vue  |  79 ++++++++++++-
 src/features/subject-identity/api.ts          |  13 ++-
 src/features/subject-identity/types.ts        |  23 ++++
 .../subject-identity/useSubjectIdentity.ts    | 110 ++++++++++++------
 .../subject-images/SubjectImagesPage.vue      |  16 +++
 src/features/subject-images/api.ts            |   6 +-
 .../components/GenerateImageDialog.vue        |  56 ++++++++-
 src/features/subject-images/types.ts          |  20 ++++
 .../subject-images/useSubjectImages.ts        |  61 ++++++----
 tests/features/production/production.test.ts  |  55 +++++++++
 tests/features/subject-images/images.test.ts  |  35 ++++++
 tests/features/subject-images/prompts.test.ts |   8 ++
 tests/lib/http.test.ts                        |  28 ++++-
 18 files changed, 583 insertions(+), 69 deletions(-)

diff --git a/docs/api-contract-sync.md b/docs/api-contract-sync.md
index de4a0d4..5406faf 100644
--- a/docs/api-contract-sync.md
+++ b/docs/api-contract-sync.md
@@ -52,3 +52,10 @@
 - 素材公开字段包含 `publicUrl`、类型、MIME、扩展名、大小、分类、元数据和时间；页面只读取 `metadata.originalName` 作为辅助搜索，不依赖内部存储路径。
 - 视觉风格新增参考图可发送 `projectAssetId`，后端从所属项目素材解析 `imageUrl`；外部 `imageUrl` 入口继续兼容。资产关联关系不在公开风格图 DTO 中推断。
 - 删除仍由后端执行引用保护；前端不先删风格记录，也不将失败响应显示为已删除。
+
+## 2026-09-22 增量同步
+
+- 新增形态图和镜头首帧的 POST 预览接口；请求体默认 `{}`，只读取后端当前生成配置与参考图规划，不产生模型任务。
+- 首帧预览按 `referencePlan.selected`／`omitted` 展示 Provider 上限后的真实输入；不再从通用 `keyframe-spec` 推断最终选图。
+- 形态提示词、形态图片及主体身份文本批处理均可发送可选 `module`；省略时保持全项目行为。
+- 场景／道具身份母版使用 `/projects/:projectId/subject-identities/anchors/generate`，请求包含 `module`、`limit`、`concurrency`；成功项已由后端设为 Anchor 并锁定，不再由前端逐项补 PUT。
diff --git a/docs/backend-coverage.md b/docs/backend-coverage.md
index 52aebbc..72b57e8 100644
--- a/docs/backend-coverage.md
+++ b/docs/backend-coverage.md
@@ -83,3 +83,15 @@
 核对后端 dev `d9ca93ef54a4c5fbe3917c504531a45f31cb9966`。新增独立项目资产库页面，接入图片上传、列表筛选、详情预览、名称／分类编辑和删除；上传使用 multipart，前端与后端同时限制 JPEG、PNG、WebP 及单张 20MB。视觉风格页可按需读取资产库并以正式 `projectAssetId` 登记参考图，仍保留外部 URL 兼容入口。后端负责校验资产归属、图片类型和删除引用保护。
 
 启用的视觉风格参考图会由后端编译为形态生图的分类参考职责，身份母版继续承担主体一致性。前端不拼接 Provider Prompt，也不把资产库图片自动设为形态主图；已有形态图不会因资产或风格参考变化被客户端静默替换。
+
+## 2026-09-22 生图预览与主体模块批处理
+
+核对后端 dev `9d927cd3e9d7e01415ad953bf3a0957bc3825050`，补齐以下正式能力：
+
+- 形态生图弹窗接入 POST `/subject-forms/:subjectFormId/images/preview`，展示后端实际 Provider、模型、参考图顺序、生成配置与最终 Prompt；预览不创建生图任务。
+- 首帧弹窗接入 POST `/storyboard-shots/:shotId/keyframe/preview`，展示视觉风格、身份母版和形态主图经过去重与 Provider 上限规划后的实际选中／省略结果。
+- 形态提示词与形态图片批处理支持 `module=character|scene|prop`，前端提供统一批处理范围选择，不再只能执行全项目全部类型。
+- 身份文本批处理支持模块筛选；场景／道具新增稳定母版小批量入口，成功后由后端原子设为 Anchor 并锁定 Identity，前端保留部分失败回执。
+- 全局 GET `/assets` 已由现有资产库页面使用，无需新增重复页面。
+
+预览结果只用于提交前核对，不作为持久化任务或质量通过证明；真正生成仍需费用确认，并以刷新后的正式资产记录为准。
diff --git a/src/features/production/api.ts b/src/features/production/api.ts
index 408b6ee..045869f 100644
--- a/src/features/production/api.ts
+++ b/src/features/production/api.ts
@@ -6,6 +6,7 @@ import type {
     GenerateQualityVideosInput,
     KeyframeBatchResult,
     KeyframeReadiness,
+    KeyframeGenerationPreview,
     KeyframeSpec,
     ProjectVideoStatus,
     PromptReadiness,
@@ -106,6 +107,9 @@ export const productionApi = {
         request<ShotKeyframe[]>(`${shotPath(shotId)}/keyframes`, { signal }),
     keyframeSpec: (shotId: string, signal?: AbortSignal) =>
         request<KeyframeSpec>(`${shotPath(shotId)}/keyframe-spec`, { signal }),
+    /** 预览最终参考图规划与 Provider Prompt，不创建首帧任务。 */
+    previewKeyframe: (shotId: string) =>
+        request<KeyframeGenerationPreview>(`${shotPath(shotId)}/keyframe/preview`, { method: 'POST', body: {} }),
     generateKeyframe: (shotId: string, input: GenerateKeyframeInput) =>
         request<ShotKeyframe | null>(`${shotPath(shotId)}/keyframe`, {
             method: 'POST',
diff --git a/src/features/production/components/KeyframeDialog.vue b/src/features/production/components/KeyframeDialog.vue
index c5915c8..9047663 100644
--- a/src/features/production/components/KeyframeDialog.vue
+++ b/src/features/production/components/KeyframeDialog.vue
@@ -2,11 +2,13 @@
 import AppForm from '../../../components/ui/AppForm.vue'
 import { NFormItem } from 'naive-ui'
 import { sizeRules } from '../../../lib/form-rules'
-import { NAlert, NButton, NCheckbox, NInputNumber } from 'naive-ui'
+import { NAlert, NButton, NCheckbox, NCollapse, NCollapseItem, NInputNumber, NTag } from 'naive-ui'
 import { computed, ref, watch } from 'vue'
 import { AppDialog } from '../../../components/ui'
+import { errorMessage } from '../../../lib/http'
+import { productionApi } from '../api'
 import { validOptionalSize } from '../model'
-import type { GenerateKeyframeInput, ShotKeyframe } from '../types'
+import type { GenerateKeyframeInput, KeyframeGenerationPreview, ShotKeyframe } from '../types'
 
 /** 单镜头首帧表单；普通候选仍由用户明确选择，过期主首帧默认替换为新图。 */
 const props = defineProps<{
@@ -22,6 +24,9 @@ const width = ref<number | ''>('')
 const height = ref<number | ''>('')
 const setPrimary = ref(false)
 const confirmed = ref(false)
+const preview = ref<KeyframeGenerationPreview | null>(null)
+const previewError = ref('')
+const previewPending = ref(false)
 const hasPrimary = computed(() => props.keyframes.some(item => item.isPrimary && item.status === 'completed'))
 const sizeValid = computed(() => validOptionalSize(width.value, height.value))
 const canSubmit = computed(() => !props.disabled && sizeValid.value && confirmed.value)
@@ -32,6 +37,9 @@ function reset() {
     height.value = ''
     setPrimary.value = !!props.replacePrimary || !hasPrimary.value
     confirmed.value = false
+    preview.value = null
+    previewError.value = ''
+    previewPending.value = false
 }
 watch([() => props.shotId, open], reset, { immediate: true })
 
@@ -47,6 +55,23 @@ function submit() {
 }
 const formModel = computed(() => ({ width: width.value, height: height.value }))
 const rules = sizeRules(() => formModel.value)
+
+/** 读取后端最终参考图选择与 Prompt，不创建首帧候选。 */
+async function loadPreview() {
+    if (!props.shotId || previewPending.value) return
+    previewPending.value = true
+    previewError.value = ''
+    try {
+        const result = await productionApi.previewKeyframe(props.shotId)
+        if (!result || result.shotId !== props.shotId) throw new Error('预览结果与当前镜头不匹配。')
+        preview.value = result
+    } catch (cause) {
+        preview.value = null
+        previewError.value = errorMessage(cause)
+    } finally {
+        previewPending.value = false
+    }
+}
 </script>
 
 <template>
@@ -66,6 +91,49 @@ const rules = sizeRules(() => formModel.value)
             <NAlert v-if="replacePrimary" type="info" :show-icon="false" class="mt-5 text-xs">
                 当前主首帧使用的主体参考资产已经变化。新图成功后应设为主首帧，旧图会继续保留为历史候选。
             </NAlert>
+            <div class="mt-5 flex flex-wrap items-center gap-3">
+                <NButton :loading="previewPending" @click="loadPreview">预览实际生成输入</NButton>
+                <span class="text-xs text-muted">只读取参考图规划与最终 Prompt，不调用图片模型。</span>
+            </div>
+            <NAlert v-if="previewError" type="error" :show-icon="false" class="mt-3">{{ previewError }}</NAlert>
+            <div v-if="preview" class="panel mt-3 p-4 text-xs leading-6">
+                <div class="flex flex-wrap items-center gap-2">
+                    <strong>{{ preview.provider }} / {{ preview.model }}</strong>
+                    <NTag size="small" :bordered="false">
+                        已选 {{ preview.referencePlan.selected.length }} /
+                        {{ preview.referencePlan.totalCandidates }} 张
+                    </NTag>
+                    <NTag v-if="preview.referencePlan.omitted.length" size="small" type="warning" :bordered="false">
+                        省略 {{ preview.referencePlan.omitted.length }} 张
+                    </NTag>
+                </div>
+                <ul v-if="preview.referencePlan.selected.length" class="mt-3 space-y-1 text-muted">
+                    <li v-for="item in preview.referencePlan.selected" :key="`${item.role}:${item.imageId}`">
+                        {{ item.subjectRef }} · {{ item.role }} · <code>{{ item.imageId }}</code>
+                    </li>
+                </ul>
+                <NCollapse class="mt-3" :default-expanded-names="['prompt']">
+                    <NCollapseItem name="prompt" title="最终 Provider Prompt">
+                        <p class="mt-3 whitespace-pre-wrap">{{ preview.providerPrompt }}</p>
+                    </NCollapseItem>
+                    <NCollapseItem
+                        v-if="preview.referencePlan.omitted.length"
+                        name="omitted"
+                        title="未进入模型的参考图"
+                    >
+                        <ul class="mt-3 space-y-1 text-muted">
+                            <li v-for="item in preview.referencePlan.omitted" :key="`${item.role}:${item.imageId}`">
+                                {{ item.subjectRef }} · {{ item.role }} · <code>{{ item.imageId }}</code>
+                            </li>
+                        </ul>
+                    </NCollapseItem>
+                    <NCollapseItem name="config" title="生成配置">
+                        <pre class="mt-3 whitespace-pre-wrap">{{
+                            JSON.stringify(preview.generationConfig, null, 2)
+                        }}</pre>
+                    </NCollapseItem>
+                </NCollapse>
+            </div>
             <div class="mt-5 grid grid-cols-2 gap-4">
                 <NFormItem path="width" label="宽度（可选）"
                     ><NInputNumber
diff --git a/src/features/production/types.ts b/src/features/production/types.ts
index ec5da29..f163bb3 100644
--- a/src/features/production/types.ts
+++ b/src/features/production/types.ts
@@ -146,6 +146,53 @@ export interface KeyframeSpec {
     references: KeyframeReference[]
 }
 
+export type ImageReferenceRole =
+    | 'style-overall'
+    | 'style-character'
+    | 'style-scene'
+    | 'style-prop'
+    | 'identity-anchor'
+    | 'form-primary'
+
+export interface KeyframePreviewReference {
+    originalImageNumber: number
+    subjectRef: string
+    module: string
+    role: ImageReferenceRole
+    imageId: string
+    imageUrl: string
+}
+
+export interface KeyframeReferencePlanItem {
+    subjectId?: string
+    subjectRef: string
+    module: string
+    role: ImageReferenceRole
+    imageId: string
+    subjectFormId?: string
+}
+
+/** 首帧生成前的无费用参考图规划与 Provider Prompt 预览。 */
+export interface KeyframeGenerationPreview {
+    shotId: string
+    episodeNo: number
+    beatNo: number
+    shotNo: number
+    provider: string
+    model: string
+    generationConfig: Record<string, unknown>
+    references: KeyframePreviewReference[]
+    referencePlan: {
+        provider: string
+        maxReferenceImages: number | null
+        totalCandidates: number
+        strategy: 'all' | 'subject-coverage' | 'repair-aware'
+        selected: KeyframeReferencePlanItem[]
+        omitted: KeyframeReferencePlanItem[]
+    }
+    providerPrompt: string
+}
+
 /** 视频生成规格中的主体参考图。 */
 export interface VideoReference {
     /** 当前目标视频模型的授权素材，不拼接模型私有地址。 */
diff --git a/src/features/subject-identity/SubjectIdentityPage.vue b/src/features/subject-identity/SubjectIdentityPage.vue
index dcbab8f..a1ca4d4 100644
--- a/src/features/subject-identity/SubjectIdentityPage.vue
+++ b/src/features/subject-identity/SubjectIdentityPage.vue
@@ -64,6 +64,7 @@ const {
     save,
     generate,
     generateProject,
+    generateStableAnchors,
     generateCastingCandidates,
     generateImage,
     generateCastingCandidate,
@@ -72,6 +73,7 @@ const {
 } = useSubjectIdentity()
 const search = ref('')
 const module = ref('all')
+const identityBatchModule = ref<'all' | 'character' | 'scene' | 'prop'>('all')
 const imageOpen = ref(false)
 const castingOpen = ref(false)
 const pendingSelection = ref('')
@@ -301,6 +303,7 @@ const rules = { concurrency: integerRule('并发数'), candidateLimit: integerRu
                                                             blocked ||
                                                             !hasStyle ||
                                                             !batchValid ||
+                                                            !candidateLimitValid ||
                                                             dirty ||
                                                             !casting.total
                                                         "
@@ -343,6 +346,46 @@ const rules = { concurrency: integerRule('并发数'), candidateLimit: integerRu
                                                 /></NButton>
                                             </div>
                                         </section>
+                                        <section class="panel mb-5 p-5" aria-label="场景与道具身份母版">
+                                            <div class="flex flex-wrap items-start justify-between gap-4">
+                                                <div>
+                                                    <h3 class="text-sm font-semibold">场景与道具稳定母版</h3>
+                                                    <p class="mt-2 text-xs leading-6 text-muted">
+                                                        为已有身份提示词但缺少有效母版的场景或道具小批量生图。成功后后端会自动设为
+                                                        Anchor 并锁定 Identity。
+                                                    </p>
+                                                </div>
+                                                <div class="flex flex-wrap gap-3">
+                                                    <ConfirmAction
+                                                        label="生成场景母版"
+                                                        :disabled="
+                                                            blocked ||
+                                                            !hasStyle ||
+                                                            !batchValid ||
+                                                            !candidateLimitValid ||
+                                                            dirty ||
+                                                            !subjects.length
+                                                        "
+                                                        acknowledgement
+                                                        :description="`最多处理 ${candidateLimit} 个场景，生成成功后自动设为身份母版并锁定。可能产生图片模型费用。`"
+                                                        @confirm="generateStableAnchors('scene')"
+                                                    />
+                                                    <ConfirmAction
+                                                        label="生成道具母版"
+                                                        :disabled="
+                                                            blocked ||
+                                                            !hasStyle ||
+                                                            !batchValid ||
+                                                            dirty ||
+                                                            !subjects.length
+                                                        "
+                                                        acknowledgement
+                                                        :description="`最多处理 ${candidateLimit} 个道具，生成成功后自动设为身份母版并锁定。可能产生图片模型费用。`"
+                                                        @confirm="generateStableAnchors('prop')"
+                                                    />
+                                                </div>
+                                            </div>
+                                        </section>
                                         <NCollapse class="panel mb-5 p-5" :default-expanded-names="['details']"
                                             ><NCollapseItem name="details"
                                                 ><template #header>高级：全部主体身份文本</template>
@@ -350,6 +393,16 @@ const rules = { concurrency: integerRule('并发数'), candidateLimit: integerRu
                                                     同时处理 Character、Scene 和
                                                     Prop；角色选角建议优先使用上方专用入口。已锁定身份始终跳过。
                                                 </p>
+                                                <NSelect
+                                                    v-model:value="identityBatchModule"
+                                                    class="mt-4 w-48"
+                                                    aria-label="身份文本批处理范围"
+                                                    :options="[
+                                                        { label: String('全部类型'), value: 'all' },
+                                                        { label: String('仅人物'), value: 'character' },
+                                                        { label: String('仅场景'), value: 'scene' },
+                                                        { label: String('仅道具'), value: 'prop' }
+                                                    ]" />
                                                 <ConfirmAction
                                                     class="mt-4"
                                                     :label="force ? '重生成全部未锁定身份' : '补齐全部主体身份文本'"
@@ -358,7 +411,17 @@ const rules = { concurrency: integerRule('并发数'), candidateLimit: integerRu
                                                     "
                                                     acknowledgement
                                                     description="生成全项目主体身份描述与提示词，不生成图片。覆盖不会自动更新旧身份图、形态图、首帧或视频。"
-                                                    @confirm="generateProject(false)" /></NCollapseItem
+                                                    @confirm="
+                                                        identityBatchModule === 'character'
+                                                            ? generateProject(true)
+                                                            : generateProject(
+                                                                  false,
+                                                                  identityBatchModule === 'scene' ||
+                                                                      identityBatchModule === 'prop'
+                                                                      ? identityBatchModule
+                                                                      : undefined
+                                                              )
+                                                    " /></NCollapseItem
                                         ></NCollapse>
                                         <section
                                             v-if="session.receipt"
@@ -382,13 +445,25 @@ const rules = { concurrency: integerRule('并发数'), candidateLimit: integerRu
                                                 {{ session.receipt.result.skippedLocked }}）· 失败
                                                 {{ session.receipt.result.failed }}
                                             </p>
-                                            <p v-else class="mt-3 text-xs text-muted">
+                                            <p
+                                                v-else-if="session.receipt.kind === 'casting'"
+                                                class="mt-3 text-xs text-muted"
+                                            >
                                                 共 {{ session.receipt.result.totalCharacters }} 个角色 · 缺候选
                                                 {{ session.receipt.result.missingAnchor }} · 本批目标
                                                 {{ session.receipt.result.targetCount }} · 已生成
                                                 {{ session.receipt.result.generated }} · 失败
                                                 {{ session.receipt.result.failed }}
                                             </p>
+                                            <p v-else class="mt-3 text-xs text-muted">
+                                                共 {{ session.receipt.result.total }} 个{{
+                                                    session.receipt.result.module === 'scene' ? '场景' : '道具'
+                                                }}
+                                                · 符合条件 {{ session.receipt.result.eligibleCount }} · 本批目标
+                                                {{ session.receipt.result.targetCount }} · 已生成并锁定
+                                                {{ session.receipt.result.generated }} · 失败
+                                                {{ session.receipt.result.failed }}
+                                            </p>
                                             <NAlert
                                                 v-if="session.receipt.result.failed"
                                                 role="alert"
diff --git a/src/features/subject-identity/api.ts b/src/features/subject-identity/api.ts
index 0d76b0c..da3a918 100644
--- a/src/features/subject-identity/api.ts
+++ b/src/features/subject-identity/api.ts
@@ -8,7 +8,9 @@ import type {
     GenerateIdentityImageInput,
     IdentityBatchResult,
     IdentityImage,
+    IdentityModule,
     SaveIdentityInput,
+    StableAnchorBatchResult,
     SubjectIdentity
 } from './types'
 
@@ -29,7 +31,7 @@ export const subjectIdentityApi = {
             body: { force },
             timeoutMs: 0
         }),
-    generateProject: (projectId: string, input: { force: boolean; concurrency: number }) =>
+    generateProject: (projectId: string, input: { force: boolean; concurrency: number; module?: IdentityModule }) =>
         request<IdentityBatchResult>(`/projects/${encodeURIComponent(projectId)}/subject-identities/generate`, {
             method: 'POST',
             body: input,
@@ -56,6 +58,15 @@ export const subjectIdentityApi = {
             `/projects/${encodeURIComponent(projectId)}/character-casting/candidates/generate`,
             { method: 'POST', body: input, timeoutMs: 0 }
         ),
+    /** 为场景或道具生成稳定身份母版，并由后端原子设为 Anchor、锁定 Identity。 */
+    generateStableAnchors: (
+        projectId: string,
+        input: { module: 'scene' | 'prop'; limit: number; concurrency: number }
+    ) =>
+        request<StableAnchorBatchResult>(
+            `/projects/${encodeURIComponent(projectId)}/subject-identities/anchors/generate`,
+            { method: 'POST', body: input, timeoutMs: 0 }
+        ),
     listImages: (subjectId: string, signal?: AbortSignal) =>
         request<IdentityImage[]>(`${identityPath(subjectId)}/images`, { signal }),
     generateImage: (subjectId: string, input: GenerateIdentityImageInput) =>
diff --git a/src/features/subject-identity/types.ts b/src/features/subject-identity/types.ts
index 54fb93a..1946f0b 100644
--- a/src/features/subject-identity/types.ts
+++ b/src/features/subject-identity/types.ts
@@ -67,6 +67,28 @@ export interface IdentityBatchResult {
     failures: { subjectId: string; subjectRef: string; error: string }[]
 }
 
+export type IdentityModule = 'character' | 'scene' | 'prop'
+
+export interface StableAnchorBatchResult {
+    module: 'scene' | 'prop'
+    total: number
+    eligibleCount: number
+    targetCount: number
+    generated: number
+    failed: number
+    failures: { subjectId: string; subjectRef: string; subjectName: string; error: string }[]
+    results: Array<{
+        subjectId: string
+        subjectRef: string
+        subjectName: string
+        success: boolean
+        imageId?: string
+        imageUrl?: string | null
+        identityLocked?: boolean
+        error?: string
+    }>
+}
+
 /** Character 的正式选角阶段状态。 */
 export type CharacterCastingStatus = 'missing_identity' | 'missing_anchor' | 'candidate_pending' | 'unlocked' | 'ready'
 
@@ -144,3 +166,4 @@ export interface ConfirmCastingResult {
 export type IdentityReceipt =
     | { kind: 'identities'; title: string; result: IdentityBatchResult }
     | { kind: 'casting'; title: string; result: CastingBatchResult }
+    | { kind: 'stable-anchors'; title: string; result: StableAnchorBatchResult }
diff --git a/src/features/subject-identity/useSubjectIdentity.ts b/src/features/subject-identity/useSubjectIdentity.ts
index a450445..ca15bd1 100644
--- a/src/features/subject-identity/useSubjectIdentity.ts
+++ b/src/features/subject-identity/useSubjectIdentity.ts
@@ -1,5 +1,5 @@
 import { computed, ref, watch } from 'vue'
-import { useQuery } from '../../composables/useQuery'
+import { usePolling } from '../../composables/usePolling'
 import { visualStyleApi } from '../visual-style'
 import { subjectImagesApi } from '../subject-images/api'
 import { hasRunningImages } from '../subject-images/model'
@@ -30,42 +30,58 @@ export function useSubjectIdentity() {
     const candidateLimit = ref(3)
     const force = ref(false)
     /** 目录与身份已有刷新按钮，查询只在进入、切换和手动刷新时读取。 */
-    const catalog = useQuery(projectId, async (id, signal) => {
-        const forms = await subjectImagesApi.listForms(id, signal)
-        if (
-            forms.some(
-                form => form.subject.projectId !== id || form.images.some(image => image.subjectFormId !== form.id)
+    const catalog = usePolling(
+        projectId,
+        async (id, signal) => {
+            const forms = await subjectImagesApi.listForms(id, signal)
+            if (
+                forms.some(
+                    form => form.subject.projectId !== id || form.images.some(image => image.subjectFormId !== form.id)
+                )
             )
-        )
-            throw new Error('主体目录与当前项目不匹配，请刷新后重试。')
-        return {
-            subjects: groupIdentitySubjects(forms),
-            running: forms.some(form => hasRunningImages(form.images))
-        }
-    })
-    const styleQuery = useQuery(projectId, async (id, signal) => {
-        const style = await visualStyleApi.get(id, signal)
-        if (style && style.projectId !== id) throw new Error('视觉风格与当前项目不匹配。')
-        return { style }
-    })
-    const castingQuery = useQuery(projectId, async (id, signal) => {
-        const readiness = await subjectIdentityApi.castingReadiness(id, signal)
-        return readiness
-    })
+                throw new Error('主体目录与当前项目不匹配，请刷新后重试。')
+            return {
+                subjects: groupIdentitySubjects(forms),
+                running: forms.some(form => hasRunningImages(form.images))
+            }
+        },
+        false
+    )
+    const styleQuery = usePolling(
+        projectId,
+        async (id, signal) => {
+            const style = await visualStyleApi.get(id, signal)
+            if (style && style.projectId !== id) throw new Error('视觉风格与当前项目不匹配。')
+            return { style }
+        },
+        false
+    )
+    const castingQuery = usePolling(
+        projectId,
+        async (id, signal) => {
+            const readiness = await subjectIdentityApi.castingReadiness(id, signal)
+            return readiness
+        },
+        false
+    )
     const subjects = computed(() =>
         mergeCastingSubjects(catalog.data.value?.subjects ?? [], castingQuery.data.value?.items ?? [], projectId.value)
     )
     const subject = computed(() => subjects.value.find(item => item.id === selectedId.value))
     const selectionKey = computed(() => subject.value?.id ?? '')
-    const detail = useQuery(selectionKey, async (id, signal) => {
-        if (!id) return null
-        const identity = await subjectIdentityApi.get(id, signal)
-        if (!identity) return { identity: null, images: [] as IdentityImage[] }
-        assertIdentity(identity, id)
-        const images = await subjectIdentityApi.listImages(id, signal)
-        if (images.some(image => image.identityId !== identity.id)) throw new Error('身份图片与当前主体不匹配。')
-        return { identity, images }
-    })
+    const detail = usePolling(
+        selectionKey,
+        async (id, signal) => {
+            if (!id) return null
+            const identity = await subjectIdentityApi.get(id, signal)
+            if (!identity) return { identity: null, images: [] as IdentityImage[] }
+            assertIdentity(identity, id)
+            const images = await subjectIdentityApi.listImages(id, signal)
+            if (images.some(image => image.identityId !== identity.id)) throw new Error('身份图片与当前主体不匹配。')
+            return { identity, images }
+        },
+        false
+    )
     const identity = computed(() => detail.data.value?.identity ?? null)
     const images = computed(() => detail.data.value?.images ?? [])
     const castingItem = computed(() =>
@@ -148,13 +164,17 @@ export function useSubjectIdentity() {
     }
 
     /** 批量仅生成身份文本，已锁定项由后端跳过，回执保留部分失败。 */
-    async function generateProject(charactersOnly = false) {
+    async function generateProject(charactersOnly = false, module?: 'scene' | 'prop') {
         if (blocked.value || !hasStyle.value || !batchValid.value || dirty.value || !subjects.value.length) return
         const id = projectId.value
         const target = getIdentitySession(id)
-        const input = { concurrency: concurrency.value, force: force.value }
+        const input = { concurrency: concurrency.value, force: force.value, ...(module ? { module } : {}) }
         target.receipt = null
-        const title = charactersOnly ? '批量生成角色身份文本' : '批量生成全部主体身份文本'
+        const title = charactersOnly
+            ? '批量生成角色身份文本'
+            : module
+              ? `批量生成${module === 'scene' ? '场景' : '道具'}身份文本`
+              : '批量生成全部主体身份文本'
         await runOperation(id, title, async () => {
             target.receipt = {
                 kind: 'identities',
@@ -167,6 +187,27 @@ export function useSubjectIdentity() {
         await Promise.all([detail.refresh(), castingQuery.refresh()])
     }
 
+    /** 场景／道具身份母版由后端独立生成、设为 Anchor 并锁定，保留逐主体失败回执。 */
+    async function generateStableAnchors(module: 'scene' | 'prop') {
+        if (blocked.value || !hasStyle.value || !batchValid.value || !candidateLimitValid.value || dirty.value) return
+        const id = projectId.value
+        const target = getIdentitySession(id)
+        const title = `批量生成${module === 'scene' ? '场景' : '道具'}身份母版`
+        target.receipt = null
+        await runOperation(id, title, async () => {
+            target.receipt = {
+                kind: 'stable-anchors',
+                title,
+                result: await subjectIdentityApi.generateStableAnchors(id, {
+                    module,
+                    limit: candidateLimit.value,
+                    concurrency: concurrency.value
+                })
+            }
+        })
+        await Promise.all([catalog.refresh(), detail.refresh()])
+    }
+
     /** 小批量只为 missing_anchor 的 Character 生成候选，不自动选择演员。 */
     async function generateCastingCandidates() {
         if (
@@ -309,6 +350,7 @@ export function useSubjectIdentity() {
         save,
         generate,
         generateProject,
+        generateStableAnchors,
         generateCastingCandidates,
         generateImage,
         generateCastingCandidate,
diff --git a/src/features/subject-images/SubjectImagesPage.vue b/src/features/subject-images/SubjectImagesPage.vue
index 1c9ad36..eb8e0b3 100644
--- a/src/features/subject-images/SubjectImagesPage.vue
+++ b/src/features/subject-images/SubjectImagesPage.vue
@@ -61,6 +61,7 @@ const {
     limit,
     promptConcurrency,
     promptForce,
+    batchModule,
     generatePrompt,
     generatePrompts,
     blocked,
@@ -345,6 +346,21 @@ const rules = { concurrency: integerRule('批量并发'), limit: integerRule('
                     v-model:force="promptForce"
                     @generate="generatePrompts"
                 />
+                <div class="mt-5 flex flex-wrap items-center gap-3">
+                    <span class="text-xs text-muted">批处理范围</span>
+                    <NSelect
+                        v-model:value="batchModule"
+                        class="w-40"
+                        aria-label="形态批处理范围"
+                        :options="[
+                            { label: String('全部类型'), value: 'all' },
+                            { label: String('仅人物'), value: 'character' },
+                            { label: String('仅场景'), value: 'scene' },
+                            { label: String('仅道具'), value: 'prop' }
+                        ]"
+                    />
+                    <span class="text-xs text-muted">同时应用于批量提示词与批量生图。</span>
+                </div>
                 <div class="panel mt-5 p-5">
                     <div class="flex flex-wrap items-center justify-between gap-4">
                         <p class="text-sm">
diff --git a/src/features/subject-images/api.ts b/src/features/subject-images/api.ts
index 5a9ced0..f66ef99 100644
--- a/src/features/subject-images/api.ts
+++ b/src/features/subject-images/api.ts
@@ -6,7 +6,8 @@ import type {
     FormPromptResult,
     ImageBatchResult,
     SubjectFormAsset,
-    SubjectImage
+    SubjectImage,
+    SubjectImageGenerationPreview
 } from './types'
 
 /** 对正式 SubjectForm 数据库 ID 编码，不接受按名称拼出的路径。 */
@@ -33,6 +34,9 @@ export const subjectImagesApi = {
     listForms: (projectId: string, signal?: AbortSignal) =>
         request<SubjectFormAsset[]>(`/projects/${encodeURIComponent(projectId)}/subject-forms`, { signal }),
     listImages: (formId: string, signal?: AbortSignal) => request<SubjectImage[]>(formPath(formId), { signal }),
+    /** 解析正式 Prompt、参考图顺序与后端生成配置，不创建图片任务。 */
+    preview: (formId: string) =>
+        request<SubjectImageGenerationPreview>(`${formPath(formId)}/preview`, { method: 'POST', body: {} }),
     generate: (formId: string, input: GenerateFormImageInput) =>
         request<SubjectImage | null>(formPath(formId), { method: 'POST', body: input, timeoutMs: 0 }),
     generateProject: (projectId: string, input: GenerateProjectImagesInput) =>
diff --git a/src/features/subject-images/components/GenerateImageDialog.vue b/src/features/subject-images/components/GenerateImageDialog.vue
index e39eac9..b7c7902 100644
--- a/src/features/subject-images/components/GenerateImageDialog.vue
+++ b/src/features/subject-images/components/GenerateImageDialog.vue
@@ -2,11 +2,13 @@
 import AppForm from '../../../components/ui/AppForm.vue'
 import { NFormItem } from 'naive-ui'
 import { sizeRules } from '../../../lib/form-rules'
-import { NAlert, NButton, NCheckbox, NCollapse, NCollapseItem, NInput, NInputNumber } from 'naive-ui'
+import { NAlert, NButton, NCheckbox, NCollapse, NCollapseItem, NInput, NInputNumber, NTag } from 'naive-ui'
 import { computed, ref, watch } from 'vue'
 import { AppDialog } from '../../../components/ui'
+import { errorMessage } from '../../../lib/http'
+import { subjectImagesApi } from '../api'
 import { primaryImage, validImageSize } from '../model'
-import type { GenerateFormImageInput, SubjectFormAsset } from '../types'
+import type { GenerateFormImageInput, SubjectFormAsset, SubjectImageGenerationPreview } from '../types'
 
 /** 单形态生成确认：尺寸与自定义 Prompt 只用于本次请求。 */
 const props = defineProps<{ form: SubjectFormAsset | null; disabled: boolean }>()
@@ -17,6 +19,9 @@ const width = ref<number | ''>('')
 const height = ref<number | ''>('')
 const setPrimary = ref(false)
 const acknowledged = ref(false)
+const preview = ref<SubjectImageGenerationPreview | null>(null)
+const previewError = ref('')
+const previewPending = ref(false)
 const dimensionsValid = computed(() => validImageSize(width.value, height.value))
 
 /** 每次打开重新建立表单，避免把另一形态的 Prompt 或尺寸带入请求。 */
@@ -27,11 +32,31 @@ watch(
         width.value = ''
         height.value = ''
         acknowledged.value = false
+        preview.value = null
+        previewError.value = ''
+        previewPending.value = false
         setPrimary.value = props.form ? !primaryImage(props.form.images) : false
     },
     { immediate: true }
 )
 
+/** 预览由后端按当前正式配置编译，不提交图片生成任务。 */
+async function loadPreview() {
+    if (!props.form || previewPending.value) return
+    previewPending.value = true
+    previewError.value = ''
+    try {
+        const result = await subjectImagesApi.preview(props.form.id)
+        if (!result || result.subjectFormId !== props.form.id) throw new Error('预览结果与当前形态不匹配。')
+        preview.value = result
+    } catch (cause) {
+        preview.value = null
+        previewError.value = errorMessage(cause)
+    } finally {
+        previewPending.value = false
+    }
+}
+
 /** 完成前置确认后只发出一次明确的生图请求，不触发整条生产流程。 */
 function submit() {
     if (!props.form || props.disabled || !dimensionsValid.value || !acknowledged.value) return
@@ -68,6 +93,33 @@ const rules = { ...sizeRules(() => formModel.value) }
                 人物、场景、道具均会引用已锁定身份的当前母版，保持身份、空间骨架或物件结构；身份未锁定或没有母版时不会继承该图片。更换母版不会自动更新已有形态图。
                 身份母版与此处的“形态主参考图”是两种不同用途的图片。
             </p>
+            <div class="flex flex-wrap items-center gap-3">
+                <NButton :loading="previewPending" :disabled="!form" @click="loadPreview">预览实际生成输入</NButton>
+                <span class="text-xs text-muted">只读取后端编译结果，不调用图片模型。</span>
+            </div>
+            <NAlert v-if="previewError" type="error" :show-icon="false">{{ previewError }}</NAlert>
+            <div v-if="preview" class="panel p-4 text-xs leading-6">
+                <div class="flex flex-wrap items-center gap-2">
+                    <strong>{{ preview.provider }} / {{ preview.model }}</strong>
+                    <NTag size="small" :bordered="false">{{ preview.references.length }} 张参考图</NTag>
+                </div>
+                <ul v-if="preview.references.length" class="mt-3 space-y-1 text-muted">
+                    <li v-for="item in preview.references" :key="`${item.imageNumber}:${item.imageId}`">
+                        图 {{ item.imageNumber }} · {{ item.role }} · <code>{{ item.imageId }}</code>
+                    </li>
+                </ul>
+                <p v-else class="mt-3 text-muted">本次没有图片参考。</p>
+                <NCollapse class="mt-3" :default-expanded-names="['prompt']">
+                    <NCollapseItem name="prompt" title="最终 Provider Prompt">
+                        <p class="mt-3 whitespace-pre-wrap">{{ preview.providerPrompt }}</p>
+                    </NCollapseItem>
+                    <NCollapseItem name="config" title="生成配置">
+                        <pre class="mt-3 whitespace-pre-wrap">{{
+                            JSON.stringify(preview.generationConfig, null, 2)
+                        }}</pre>
+                    </NCollapseItem>
+                </NCollapse>
+            </div>
             <NFormItem
                 class="block"
                 path="prompt"
diff --git a/src/features/subject-images/types.ts b/src/features/subject-images/types.ts
index 55d6652..232c252 100644
--- a/src/features/subject-images/types.ts
+++ b/src/features/subject-images/types.ts
@@ -56,6 +56,8 @@ export interface GenerateFormImageInput {
 export interface GenerateProjectImagesInput {
     concurrency: number
     force: boolean
+    /** 可按主体模块缩小批处理范围；省略时处理全部模块。 */
+    module?: SubjectModule
     /** 仅限制本次真正生图的数量，留空表示不限制。 */
     limit?: number
 }
@@ -83,6 +85,24 @@ export type FormPromptResult = Omit<SubjectFormAsset, 'subject' | 'images'> & {
 export interface GenerateFormPromptsInput {
     concurrency: number
     force: boolean
+    module?: SubjectModule
+}
+
+export type SubjectModule = 'character' | 'scene' | 'prop'
+
+/** 形态生图前的无费用实际输入预览。 */
+export interface SubjectImageGenerationPreview {
+    subjectFormId: string
+    provider: string
+    model: string
+    generationConfig: Record<string, unknown>
+    references: Array<{
+        imageNumber: number
+        role: string
+        imageId: string
+        imageUrl: string
+    }>
+    providerPrompt: string
 }
 
 /** 长请求的回执按项目隔离，在会话内切换页面后仍可查看。 */
diff --git a/src/features/subject-images/useSubjectImages.ts b/src/features/subject-images/useSubjectImages.ts
index 3d9cdd5..e14b97a 100644
--- a/src/features/subject-images/useSubjectImages.ts
+++ b/src/features/subject-images/useSubjectImages.ts
@@ -1,12 +1,12 @@
 import { computed, ref } from 'vue'
-import { useQuery } from '../../composables/useQuery'
+import { usePolling } from '../../composables/usePolling'
 import { ApiError } from '../../lib/http'
 import { useProjectContext } from '../projects/context'
 import { getOperation, runOperation } from '../workflows/operations'
 import { workflowCheckpoints } from '../workflows/selectors'
 import { subjectImagesApi } from './api'
 import { getImageSession, hasRunningImages } from './model'
-import type { GenerateFormImageInput } from './types'
+import type { GenerateFormImageInput, SubjectModule } from './types'
 
 /** 形态图库统一读取数据库记录并管理单图／批量长请求，不调用生产流程。 */
 export function useSubjectImages() {
@@ -17,28 +17,33 @@ export function useSubjectImages() {
     const limit = ref<number | ''>('')
     const promptConcurrency = ref(3)
     const promptForce = ref(false)
-    const query = useQuery(id, async (projectId, signal) => {
-        if (!projectId) return []
-        try {
-            const forms = await subjectImagesApi.listForms(projectId, signal)
-            if (
-                forms.some(
-                    form =>
-                        form.subject.projectId !== projectId ||
-                        form.images.some(image => image.subjectFormId !== form.id)
+    const batchModule = ref<'all' | SubjectModule>('all')
+    const query = usePolling(
+        id,
+        async (projectId, signal) => {
+            if (!projectId) return []
+            try {
+                const forms = await subjectImagesApi.listForms(projectId, signal)
+                if (
+                    forms.some(
+                        form =>
+                            form.subject.projectId !== projectId ||
+                            form.images.some(image => image.subjectFormId !== form.id)
+                    )
                 )
-            )
-                throw new Error('形态图库返回了不匹配的项目或图片，请刷新后重试。')
-            return forms
-        } catch (error) {
-            if (error instanceof ApiError && error.status === 404)
-                throw new Error(
-                    '后端尚未提供形态图库查询，请更新后端 dev 并重启服务（GET /projects/:id/subject-forms）。',
-                    { cause: error }
-                )
-            throw error
-        }
-    })
+                    throw new Error('形态图库返回了不匹配的项目或图片，请刷新后重试。')
+                return forms
+            } catch (error) {
+                if (error instanceof ApiError && error.status === 404)
+                    throw new Error(
+                        '后端尚未提供形态图库查询，请更新后端 dev 并重启服务（GET /projects/:id/subject-forms）。',
+                        { cause: error }
+                    )
+                throw error
+            }
+        },
+        false
+    )
     const forms = computed(() => query.data.value ?? [])
     const operation = computed(() => getOperation(id.value))
     const session = computed(() => getImageSession(id.value))
@@ -87,7 +92,11 @@ export function useSubjectImages() {
     async function generatePrompts() {
         if (blocked.value || !promptValid.value || !forms.value.length) return
         const projectId = id.value
-        const input = { concurrency: promptConcurrency.value, force: promptForce.value }
+        const input = {
+            concurrency: promptConcurrency.value,
+            force: promptForce.value,
+            ...(batchModule.value === 'all' ? {} : { module: batchModule.value })
+        }
         const target = getImageSession(projectId)
         target.promptReceipt = null
         await runOperation(projectId, '批量生成形态正式提示词', async () => {
@@ -120,7 +129,8 @@ export function useSubjectImages() {
         const input = {
             concurrency: concurrency.value,
             force: force.value,
-            ...(limit.value === '' ? {} : { limit: limit.value })
+            ...(limit.value === '' ? {} : { limit: limit.value }),
+            ...(batchModule.value === 'all' ? {} : { module: batchModule.value })
         }
         const target = getImageSession(projectId)
         target.receipt = null
@@ -145,6 +155,7 @@ export function useSubjectImages() {
         limit,
         promptConcurrency,
         promptForce,
+        batchModule,
         promptValid,
         concurrencyValid,
         generatePrompt,
diff --git a/tests/features/production/production.test.ts b/tests/features/production/production.test.ts
index 4f7c5f7..1fd8dc3 100644
--- a/tests/features/production/production.test.ts
+++ b/tests/features/production/production.test.ts
@@ -140,4 +140,59 @@ describe('镜头生产数据契约', () => {
         await flushPromises()
         expect(wrapper.emitted('generate')?.at(-1)).toEqual([{ setPrimary: false }])
     })
+
+    it('首帧生成前可预览后端最终选图规划和 Prompt', async () => {
+        const fetcher = vi.fn<typeof fetch>().mockResolvedValue(
+            new Response(
+                JSON.stringify({
+                    data: {
+                        shotId: 'shot-1',
+                        episodeNo: 1,
+                        beatNo: 1,
+                        shotNo: 1,
+                        provider: 'seedream',
+                        model: 'image-model',
+                        generationConfig: {},
+                        references: [],
+                        referencePlan: {
+                            provider: 'seedream',
+                            maxReferenceImages: 4,
+                            totalCandidates: 2,
+                            strategy: 'subject-coverage',
+                            selected: [
+                                {
+                                    subjectRef: '@CH0001',
+                                    module: 'character',
+                                    role: 'identity-anchor',
+                                    imageId: 'anchor-1'
+                                }
+                            ],
+                            omitted: [
+                                {
+                                    subjectRef: '@SC0001',
+                                    module: 'scene',
+                                    role: 'form-primary',
+                                    imageId: 'scene-1'
+                                }
+                            ]
+                        },
+                        providerPrompt: '最终首帧 Prompt'
+                    }
+                })
+            )
+        )
+        vi.stubGlobal('fetch', fetcher)
+        wrapper = mount(KeyframeDialog, {
+            attachTo: document.body,
+            props: { open: true, shotId: 'shot-1', shotTitle: '开场', keyframes: [], disabled: false }
+        })
+        await flushPromises()
+        button('预览实际生成输入').click()
+        await flushPromises()
+        expect(fetcher.mock.calls[0]?.[0]).toBe('/api/storyboard-shots/shot-1/keyframe/preview')
+        expect(document.body.textContent).toContain('已选 1 / 2 张')
+        expect(document.body.textContent).toContain('省略 1 张')
+        expect(document.body.textContent).toContain('最终首帧 Prompt')
+        expect(wrapper.emitted('generate')).toBeUndefined()
+    })
 })
diff --git a/tests/features/subject-images/images.test.ts b/tests/features/subject-images/images.test.ts
index 8407ad7..413e988 100644
--- a/tests/features/subject-images/images.test.ts
+++ b/tests/features/subject-images/images.test.ts
@@ -70,6 +70,41 @@ describe('形态图片选择与操作', () => {
         expect(document.body.textContent).toContain('母版')
         expect(document.body.textContent).not.toContain('道具形态生图暂不自动引用')
     })
+    it('按需预览正式形态的实际参考图和 Provider Prompt，不触发生图事件', async () => {
+        const fetcher = vi.fn<typeof fetch>().mockResolvedValue(
+            new Response(
+                JSON.stringify({
+                    data: {
+                        subjectFormId: 'form-db-1',
+                        provider: 'seedream',
+                        model: 'image-model',
+                        generationConfig: { width: 2048 },
+                        references: [
+                            {
+                                imageNumber: 1,
+                                role: 'identity-anchor',
+                                imageId: 'anchor-1',
+                                imageUrl: '/storage/anchor.png'
+                            }
+                        ],
+                        providerPrompt: '最终形态 Prompt'
+                    }
+                })
+            )
+        )
+        vi.stubGlobal('fetch', fetcher)
+        wrapper = mount(GenerateImageDialog, {
+            attachTo: document.body,
+            props: { open: true, form: formFixture(), disabled: false }
+        })
+        await flushPromises()
+        button('预览实际生成输入').click()
+        await flushPromises()
+        expect(fetcher.mock.calls[0]?.[0]).toBe('/api/subject-forms/form-db-1/images/preview')
+        expect(document.body.textContent).toContain('seedream / image-model')
+        expect(document.body.textContent).toContain('最终形态 Prompt')
+        expect(wrapper.emitted('generate')).toBeUndefined()
+    })
     it('已有主图优先于新候选图和失败记录，无主图才回退到成功候选图', () => {
         const form = formFixture()
         form.images.unshift(
diff --git a/tests/features/subject-images/prompts.test.ts b/tests/features/subject-images/prompts.test.ts
index 1c60c9d..510efdc 100644
--- a/tests/features/subject-images/prompts.test.ts
+++ b/tests/features/subject-images/prompts.test.ts
@@ -112,6 +112,14 @@ describe('形态正式提示词与批量配置', () => {
         expect(service.session.value.promptReceipt?.result.failures[0]?.error).toBe('模型拒绝')
         expect(service.session.value.receipt).toBeNull()
     })
+    it('批量提示词和图片可按主体模块缩小范围', async () => {
+        const { service, posts } = await setup()
+        service.batchModule.value = 'scene'
+        await service.generatePrompts()
+        await service.generateProject()
+        expect(JSON.parse(String(posts()[0]?.[1]?.body))).toMatchObject({ module: 'scene' })
+        expect(JSON.parse(String(posts()[1]?.[1]?.body))).toMatchObject({ module: 'scene' })
+    })
     it('图片数量上限可选，非法上限和并发阻止提交', async () => {
         const { service, posts } = await setup()
         for (const limit of [0, -1, 1.5]) {
diff --git a/tests/lib/http.test.ts b/tests/lib/http.test.ts
index c13327c..e059c4e 100644
--- a/tests/lib/http.test.ts
+++ b/tests/lib/http.test.ts
@@ -4,6 +4,8 @@ import { projectsApi } from '@/features/projects/api'
 import { breakdownApi } from '@/features/breakdown/api'
 import { storyboardApi } from '@/features/storyboard/api'
 import { subjectImagesApi } from '@/features/subject-images/api'
+import { subjectIdentityApi } from '@/features/subject-identity/api'
+import { productionApi } from '@/features/production/api'
 
 afterEach(() => {
     vi.unstubAllGlobals()
@@ -43,6 +45,29 @@ describe('后端 API 契约', () => {
         })
         expect(fetcher.mock.calls[4]![1]!.body).toBeUndefined()
     })
+    it('新增预览、模块批处理和稳定母版接口使用正式数据库 ID', async () => {
+        const fetcher = vi.fn<typeof fetch>().mockImplementation(async () => new Response('{"data":{}}'))
+        vi.stubGlobal('fetch', fetcher)
+        await subjectImagesApi.preview('form/id')
+        await subjectImagesApi.generatePrompts('project/id', { concurrency: 2, force: false, module: 'scene' })
+        await subjectImagesApi.generateProject('project/id', { concurrency: 1, force: false, module: 'prop' })
+        await subjectIdentityApi.generateProject('project/id', { concurrency: 2, force: false, module: 'scene' })
+        await subjectIdentityApi.generateStableAnchors('project/id', { module: 'prop', limit: 3, concurrency: 2 })
+        await productionApi.previewKeyframe('shot/id')
+        expect(fetcher.mock.calls.map(([url, init]) => [url, init?.method])).toEqual([
+            ['/api/subject-forms/form%2Fid/images/preview', 'POST'],
+            ['/api/projects/project%2Fid/subject-forms/generation-prompts', 'POST'],
+            ['/api/projects/project%2Fid/subject-images/generate', 'POST'],
+            ['/api/projects/project%2Fid/subject-identities/generate', 'POST'],
+            ['/api/projects/project%2Fid/subject-identities/anchors/generate', 'POST'],
+            ['/api/storyboard-shots/shot%2Fid/keyframe/preview', 'POST']
+        ])
+        expect(JSON.parse(fetcher.mock.calls[4]![1]!.body as string)).toEqual({
+            module: 'prop',
+            limit: 3,
+            concurrency: 2
+        })
+    })
     it('分镜只调用正式项目接口，并保持 force 和零次修复参数', async () => {
         const fetcher = vi.fn<typeof fetch>().mockImplementation(async () => new Response('{"data":{}}'))
         vi.stubGlobal('fetch', fetcher)
@@ -110,12 +135,11 @@ describe('后端 API 契约', () => {
             )
         vi.stubGlobal('fetch', fetcher)
         expect(await projectsApi.list()).toEqual([{ id: 'p1' }])
-        expect(await projectsApi.create({ title: '测试剧本', topic: '故事', style: '悬疑', episodeCount: 3 })).toEqual({
+        expect(await projectsApi.create({ topic: '故事', style: '悬疑', episodeCount: 3 })).toEqual({
             projectId: 'p2',
             status: 'generating'
         })
         expect(JSON.parse(fetcher.mock.calls[1]![1].body as string)).toEqual({
-            title: '测试剧本',
             topic: '故事',
             style: '悬疑',
             episodeCount: 3