refactor+perf+fix: G5 模块边界 SPI 化、C5 流式解析、C6/C7 查询优化、D13 队列持久化、A3/A4/A7 鉴权

模块边界（G5 / G7）
- 新增 task/spi/TaskModuleHeartbeatSpi + 13 个模块实现：TaskHeartbeatService 不再 import 任何
  业务模块（原先注入 12 个 CacheService 并用 switch 分发）；启动校验重复注册
- 新增 task/spi/BrandTaskHeartbeatSpi（品牌任务心跳/中断）、BrandTaskStaleRepairSpi（陈旧修复）、
  CollectDataItemCleanupSpi（历史清理）：跨模块 Mapper 操作收回业务模块
- G7：10 个被跨模块借用的 productrisk VO 迁至 common/model/vo
- 架构棘轮收紧：TASK_TO_BUSINESS_BASELINE 119 → 6（实测）
- 新增 TaskModuleHeartbeatSpiCoverageTest（moduleType 覆盖与拼写）

性能与容量（C5/C6/C7/D13）
- C5 流式解析：SkipPriceAsinService（含两行表头语义）、AppearancePatentExcelParser、
  BrandTaskService、DeleteBrandRunService、LocalFileStorageService.getExcelInfo 改 ExcelStreamReader；
  行数上限改为迭代中生效
- C6 去重总数据列表：关键字改前缀匹配（命中 uk_data_value）；V124 删除永不生效的 idx_country
- C7 撞款扫描：按店逐批取数（索引前缀），不再全表 GROUP BY + JOIN + 全量拉内存
- D13 待删对象本地日志 PendingDeleteJournal（启动回放 + 收敛重写），RustfsDeleteRetryService 与
  TransientPayloadDeleteOrchestrator 接入；异步删除失败对象写回日志
- V123 删除 biz_file_result 两个被复合索引覆盖的单列索引

安全（A3/A4/A7 + 守卫名单）
- A4 数字人版本写操作要求管理员；A7 视频密钥按登录身份（超管例外）
- A3 上传接口加危险扩展名黑名单（可配置）
- AdminApiGuardFilter 用户态名单补 /api/ziniao（controller 已 requireAdmin，此处为开关打开后的第二层）

容量（明细表保留期）
- 新增 ShopDataCrawlItemRetentionService：biz_shop_data_crawl_item 按保留期（默认 30 天）分批清理
  （该表此前无任何清理策略，是增长最快的表），job 锁 + 单轮批次上限

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskCandidateVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import com.fasterxml.jackson.annotation.JsonProperty;
 import io.swagger.v3.oas.annotations.media.Schema;

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskCountryPreferenceVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import com.fasterxml.jackson.annotation.JsonProperty;
 import io.swagger.v3.oas.annotations.media.Schema;

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskDashboardVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import com.fasterxml.jackson.annotation.JsonProperty;
 import io.swagger.v3.oas.annotations.media.Schema;

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskHistoryVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.Data;

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskMatchShopsVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.Data;

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskResultItemVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import com.fasterxml.jackson.annotation.JsonAlias;
 import com.fasterxml.jackson.annotation.JsonProperty;

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskShopQueueItemVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import com.fasterxml.jackson.annotation.JsonAlias;
 import com.fasterxml.jackson.annotation.JsonProperty;

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskTaskBatchVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.Data;

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskTaskDetailVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.Data;

backend-java/src/main/java/com/nanri/aiimage/common/model/vo/ProductRiskTaskItemVo.java

@@ -1,4 +1,4 @@
-package com.nanri.aiimage.modules.productrisk.model.vo;
+package com.nanri.aiimage.common.model.vo;
 
 import com.fasterxml.jackson.annotation.JsonProperty;
 import com.nanri.aiimage.modules.shopmatch.model.vo.ShopMatchTaskStageVo;

backend-java/src/main/java/com/nanri/aiimage/common/module/TaskModuleRegistry.java

@@ -0,0 +1,111 @@
+package com.nanri.aiimage.common.module;
+
+import java.util.LinkedHashMap;
+import java.util.List;
+import java.util.Map;
+import java.util.Set;
+import java.util.stream.Collectors;
+
+/**
+ * 任务模块清单的**单一来源**（2026-09 全维度审查 G6）。
+ *
+ * 背景：新增一个工具模块此前要同时改 5 处枚举（结果文件 Job 白名单、按天清理清单、
+ * 站内通知模块名、任务心跳分支、陈旧判死巡检线），其中 4 处没有任何自检，
+ * 漏改表现为"功能静默不生效"。现在把**数据驱动的三处**（结果文件 Job、按天清理、通知中文名）
+ * 统一从这里派生，并由 {@code TaskModuleCoverageTest} 守住覆盖面。
+ *
+ * 心跳与陈旧判死两处是代码分支驱动（switch / 逐模块委派），无法纯数据派生，
+ * 新增模块时仍需实现对应 Handler，测试会提示缺失。
+ */
+public final class TaskModuleRegistry {
+
+    /**
+     * 单个任务模块的声明。
+     *
+     * @param type            moduleType（与 biz_file_task.module_type 一致）
+     * @param label           站内通知/界面用中文名
+     * @param resultFileJob   是否产出结果文件（需要注册 ResultFileJobHandler）
+     * @param ageCleanup      是否参与按天清理（ModuleCleanupProperties）
+     * @param delegatedStaleCheck 是否以"委派"方式并入 stale-check 巡检线
+     *                        （DeleteBrandStaleTaskService.delegatedStaleChecks，必须逐模块登记动作）
+     * @param selfScheduledStaleCheck 该模块的陈旧判死自带更快的调度（publish 60s / collect-data 30s），
+     *                        刻意不并入 2 分钟一轮的巡检线——并入会显著拉长判死时延。
+     *                        这类模块的"不在委托名单"是设计差异，不是漏接；由覆盖面测试固定。
+     */
+    public record Module(String type, String label, boolean resultFileJob, boolean ageCleanup,
+                         boolean delegatedStaleCheck, boolean selfScheduledStaleCheck) {
+    }
+
+    private static final List<Module> MODULES = List.of(
+            new Module("PUBLISH", "上架", true, false, false, true),
+            new Module("DEDUPE", "数据去重", false, true, false, false),
+            new Module("SPLIT", "数据拆分", false, true, false, false),
+            new Module("CONVERT", "格式转换", false, true, false, false),
+            new Module("DELETE_BRAND", "删除ASIN", true, true, false, false),
+            new Module("PRODUCT_RISK_RESOLVE", "商品风险解决", true, true, false, false),
+            new Module("PRICE_TRACK", "跟价", true, true, false, false),
+            new Module("SHOP_MATCH", "定时匹配", true, true, false, false),
+            new Module("PATROL_DELETE", "巡店删除", true, true, false, false),
+            new Module("QUERY_ASIN", "查询ASIN", true, true, false, false),
+            new Module("WITHDRAW", "取款", true, true, false, false),
+            new Module("APPEARANCE_PATENT", "外观专利检测", true, true, true, false),
+            new Module("SIMILAR_ASIN", "货源查询", true, true, true, false),
+            new Module("COLLECT_DATA", "采集数据", true, true, false, true),
+            new Module("SHOP_DATA_CRAWL", "店铺数据抓取", true, false, true, false),
+            new Module("BRAND", "品牌检测", true, false, true, false)
+    );
+
+    private TaskModuleRegistry() {
+    }
+
+    public static List<Module> modules() {
+        return MODULES;
+    }
+
+    /** 全部 moduleType。 */
+    public static Set<String> moduleTypes() {
+        return MODULES.stream().map(Module::type).collect(Collectors.toUnmodifiableSet());
+    }
+
+    /** moduleType → 中文名（站内通知等展示用）。 */
+    public static Map<String, String> labels() {
+        Map<String, String> labels = new LinkedHashMap<>();
+        for (Module module : MODULES) {
+            labels.put(module.type(), module.label());
+        }
+        return Map.copyOf(labels);
+    }
+
+    /** 产出结果文件的模块（需要注册 Handler 的模块）。 */
+    public static Set<String> resultFileJobModuleTypes() {
+        return MODULES.stream().filter(Module::resultFileJob).map(Module::type)
+                .collect(Collectors.toUnmodifiableSet());
+    }
+
+    /** 参与按天清理的模块。 */
+    public static Set<String> ageCleanupModuleTypes() {
+        return MODULES.stream().filter(Module::ageCleanup).map(Module::type)
+                .collect(Collectors.toUnmodifiableSet());
+    }
+
+    /** 自带更快巡检节奏、刻意不并入集中巡检线的模块。 */
+    public static Set<String> selfScheduledStaleCheckModuleTypes() {
+        return MODULES.stream().filter(Module::selfScheduledStaleCheck).map(Module::type)
+                .collect(Collectors.toUnmodifiableSet());
+    }
+
+    /** 以"委派"方式并入 stale-check 巡检线的模块（DeleteBrandStaleTaskService 必须逐模块登记动作）。 */
+    public static Set<String> delegatedStaleCheckModuleTypes() {
+        return MODULES.stream().filter(Module::delegatedStaleCheck).map(Module::type)
+                .collect(Collectors.toUnmodifiableSet());
+    }
+
+    public static String labelOf(String moduleType) {
+        for (Module module : MODULES) {
+            if (module.type().equals(moduleType)) {
+                return module.label();
+            }
+        }
+        return moduleType;
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/common/service/NodeSharedStore.java

@@ -0,0 +1,169 @@
+package com.nanri.aiimage.common.service;
+
+import com.fasterxml.jackson.databind.ObjectMapper;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.data.redis.core.StringRedisTemplate;
+
+import java.time.Duration;
+import java.util.concurrent.ConcurrentHashMap;
+import java.util.concurrent.atomic.AtomicLong;
+
+/**
+ * 跨节点的小容量状态存储（2026-09 全维度审查 D2）。
+ *
+ * <p>用途：导入/导出任务的进度、归属用户、分组等状态。原先只存节点本地内存，
+ * 客户端轮询落到另一节点就报"任务不存在"（nginx 的 user_id 亲和只覆盖常态）。
+ * 现在本地 Map 作快路径、Redis 作跨节点真源，任何节点都能读到。
+ *
+ * <p>写节流：调用方可能**逐行**刷新进度（数十万行），逐次写 Redis 不可接受；
+ * 默认 500ms 内只写一次（终态用 {@link #putNow} 立即落库）。
+ *
+ * <p>语义取舍：读-改-写不保证原子（进度类状态可接受）；{@code redis}/{@code objectMapper}
+ * 为空时退化为纯本地（单测与未注入场景）。
+ */
+@Slf4j
+public final class NodeSharedStore<K, V> {
+
+    private static final long DEFAULT_WRITE_THROTTLE_MILLIS = 500L;
+
+    private final String keyPrefix;
+    private final Duration ttl;
+    private final Class<V> valueType;
+    private final StringRedisTemplate redis;
+    private final ObjectMapper objectMapper;
+    private final long writeThrottleMillis;
+
+    private final ConcurrentHashMap<K, V> local = new ConcurrentHashMap<>();
+    private final ConcurrentHashMap<K, AtomicLong> lastWriteAt = new ConcurrentHashMap<>();
+
+    public NodeSharedStore(String keyPrefix,
+                           Duration ttl,
+                           Class<V> valueType,
+                           StringRedisTemplate redis,
+                           ObjectMapper objectMapper) {
+        this(keyPrefix, ttl, valueType, redis, objectMapper, DEFAULT_WRITE_THROTTLE_MILLIS);
+    }
+
+    public NodeSharedStore(String keyPrefix,
+                           Duration ttl,
+                           Class<V> valueType,
+                           StringRedisTemplate redis,
+                           ObjectMapper objectMapper,
+                           long writeThrottleMillis) {
+        this.keyPrefix = keyPrefix;
+        this.ttl = ttl;
+        this.valueType = valueType;
+        this.redis = redis;
+        this.objectMapper = objectMapper;
+        this.writeThrottleMillis = Math.max(0L, writeThrottleMillis);
+    }
+
+    /** 本地优先；本地没有则读 Redis 并回填本地（跨节点可见）。 */
+    public V get(K key) {
+        if (key == null) {
+            return null;
+        }
+        V cached = local.get(key);
+        if (cached != null) {
+            return cached;
+        }
+        V remote = readRemote(key);
+        if (remote != null) {
+            local.put(key, remote);
+        }
+        return remote;
+    }
+
+    public boolean containsKey(K key) {
+        return get(key) != null;
+    }
+
+    /** 写入并（按节流）同步到 Redis。 */
+    public void put(K key, V value) {
+        if (key == null || value == null) {
+            return;
+        }
+        local.put(key, value);
+        if (!throttleAllowsWrite(key)) {
+            return;
+        }
+        writeRemote(key, value);
+    }
+
+    /** 立即写入 Redis（终态、归属等一次性状态用）。 */
+    public void putNow(K key, V value) {
+        if (key == null || value == null) {
+            return;
+        }
+        local.put(key, value);
+        writeRemote(key, value);
+    }
+
+    /** 删除（本地 + Redis），返回删除前的值（可能为 null）。 */
+    public V remove(K key) {
+        if (key == null) {
+            return null;
+        }
+        V previous = local.remove(key);
+        lastWriteAt.remove(key);
+        if (redis != null) {
+            try {
+                redis.delete(fullKey(key));
+            } catch (Exception ex) {
+                log.warn("[node-shared-store] 删除远端状态失败 key={} msg={}", fullKey(key), ex.getMessage());
+            }
+        }
+        return previous;
+    }
+
+    /** 本地快照（仅用于日志/统计；不含其它节点的写入）。 */
+    public int localSize() {
+        return local.size();
+    }
+
+    private boolean throttleAllowsWrite(K key) {
+        if (writeThrottleMillis <= 0L) {
+            return true;
+        }
+        long now = System.currentTimeMillis();
+        AtomicLong last = lastWriteAt.computeIfAbsent(key, ignored -> new AtomicLong(0L));
+        long previous = last.get();
+        if (now - previous < writeThrottleMillis) {
+            return false;
+        }
+        return last.compareAndSet(previous, now);
+    }
+
+    private V readRemote(K key) {
+        if (redis == null || objectMapper == null) {
+            return null;
+        }
+        try {
+            String json = redis.opsForValue().get(fullKey(key));
+            if (json == null || json.isBlank()) {
+                return null;
+            }
+            return objectMapper.readValue(json, valueType);
+        } catch (Exception ex) {
+            log.warn("[node-shared-store] 读取远端状态失败 key={} msg={}", fullKey(key), ex.getMessage());
+            return null;
+        }
+    }
+
+    private void writeRemote(K key, V value) {
+        if (redis == null || objectMapper == null) {
+            return;
+        }
+        try {
+            String json = objectMapper.writeValueAsString(value);
+            redis.opsForValue().set(fullKey(key), json, ttl);
+        } catch (Exception ex) {
+            // 写失败不影响本地进度（下次 put 会重试），只留线索
+            log.warn("[node-shared-store] 写入远端状态失败 key={} msg={}", fullKey(key), ex.getMessage());
+        }
+    }
+
+    private String fullKey(K key) {
+        return keyPrefix + ":" + key;
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/common/util/ExcelStreamReader.java

@@ -56,6 +56,13 @@ public final class ExcelStreamReader {
         default void onHeader(String sheetName, Integer sheetNo, Map<Integer, String> headerMap) throws Exception {
         }
 
+        /**
+         * 表头之后回调本次 sheet 的近似总行数（EasyExcel 基于 sheet 尺寸，可能为 null）。
+         * 流式解析无法在读完前得到精确行数，需要展示进度/做前置上限校验的调用方可用它近似。
+         */
+        default void onSheetTotal(String sheetName, Integer sheetNo, Integer approximateTotalRows) throws Exception {
+        }
+
         void onRow(String sheetName, Integer sheetNo, int rowIndex, Map<Integer, String> headerMap, Map<Integer, String> rowMap) throws Exception;
     }
 
@@ -77,6 +84,7 @@ public final class ExcelStreamReader {
             currentHeaderMap = normalizedHeadMap;
             try {
                 handler.onHeader(sheetName(context), sheetNo(context), currentHeaderMap);
+                handler.onSheetTotal(sheetName(context), sheetNo(context), approximateTotalRows(context));
             } catch (RuntimeException ex) {
                 throw ex;
             } catch (Exception ex) {
@@ -109,6 +117,15 @@ public final class ExcelStreamReader {
         public void doAfterAllAnalysed(AnalysisContext context) {
         }
 
+        private Integer approximateTotalRows(AnalysisContext context) {
+            try {
+                return context.readSheetHolder() == null ? null
+                        : context.readSheetHolder().getApproximateTotalRowNumber();
+            } catch (Exception ex) {
+                return null;
+            }
+        }
+
         private String sheetName(AnalysisContext context) {
             return context.readSheetHolder() == null ? "" : context.readSheetHolder().getSheetName();
         }

backend-java/src/main/java/com/nanri/aiimage/common/util/PendingDeleteJournal.java

@@ -0,0 +1,138 @@
+package com.nanri.aiimage.common.util;
+
+import lombok.extern.slf4j.Slf4j;
+
+import java.io.IOException;
+import java.nio.charset.StandardCharsets;
+import java.nio.file.Files;
+import java.nio.file.Path;
+import java.nio.file.StandardCopyOption;
+import java.nio.file.StandardOpenOption;
+import java.util.ArrayList;
+import java.util.Collection;
+import java.util.LinkedHashSet;
+import java.util.List;
+import java.util.Set;
+
+/**
+ * 待删对象队列的本地落盘日志：进程重启后仍能恢复「待删除对象」清单。
+ *
+ * <p>使用场景：对象存储删除失败后的补偿队列原本只在节点内存里，重启即丢，
+ * 对应对象会一直残留在桶里直到生命周期规则过期。这里用一行一个 objectKey 的
+ * 追加日志做最小持久化——入队追加、队列收敛后整体重写、启动时回放。
+ *
+ * <p>并发：所有方法内部同步，调用方无需额外加锁；文件损坏/读写异常只记日志不抛出，
+ * 保证补偿链路本身不会因为落盘失败而中断业务。
+ *
+ * <p>上限：{@code maxEntries} 用于防止日志在异常堆积时无界增长（超出后丢弃最早的记录，
+ * 与内存队列的容量准入语义一致——丢的是「待删对象」，最坏结果是对象残留）。
+ */
+@Slf4j
+public final class PendingDeleteJournal {
+
+    private final Path path;
+    private final int maxEntries;
+    private final Object lock = new Object();
+
+    public PendingDeleteJournal(Path path, int maxEntries) {
+        this.path = path;
+        this.maxEntries = Math.max(1, maxEntries);
+    }
+
+    public Path getPath() {
+        return path;
+    }
+
+    /** 追加一条待删对象（重复追加由回放时的 Set 语义去重）。 */
+    public void record(String objectKey) {
+        if (objectKey == null || objectKey.isBlank()) {
+            return;
+        }
+        synchronized (lock) {
+            try {
+                Path parent = path.getParent();
+                if (parent != null) {
+                    Files.createDirectories(parent);
+                }
+                Files.writeString(path, sanitize(objectKey) + System.lineSeparator(),
+                        StandardCharsets.UTF_8, StandardOpenOption.CREATE, StandardOpenOption.APPEND);
+            } catch (Exception ex) {
+                log.warn("[delete-journal] 追加待删对象失败 path={} msg={}", path, ex.getMessage());
+            }
+        }
+    }
+
+    /** 启动回放：返回日志中的待删对象（按首次出现顺序去重）。 */
+    public List<String> readAll() {
+        synchronized (lock) {
+            if (!Files.isRegularFile(path)) {
+                return List.of();
+            }
+            List<String> lines;
+            try {
+                lines = Files.readAllLines(path, StandardCharsets.UTF_8);
+            } catch (Exception ex) {
+                log.warn("[delete-journal] 读取待删对象日志失败 path={} msg={}", path, ex.getMessage());
+                return List.of();
+            }
+            Set<String> unique = new LinkedHashSet<>();
+            for (String line : lines) {
+                if (line == null) {
+                    continue;
+                }
+                String key = line.trim();
+                if (!key.isEmpty()) {
+                    unique.add(key);
+                }
+            }
+            return new ArrayList<>(unique);
+        }
+    }
+
+    /** 队列收敛（成功删除/引用仍在）后整体重写为剩余的待删对象；剩余为空则删除日志。 */
+    public void rewrite(Collection<String> remaining) {
+        synchronized (lock) {
+            Set<String> keep = new LinkedHashSet<>();
+            if (remaining != null) {
+                for (String key : remaining) {
+                    if (key != null && !key.isBlank()) {
+                        keep.add(sanitize(key));
+                        if (keep.size() >= maxEntries) {
+                            break;
+                        }
+                    }
+                }
+            }
+            try {
+                if (keep.isEmpty()) {
+                    Files.deleteIfExists(path);
+                    return;
+                }
+                Path parent = path.getParent();
+                if (parent != null) {
+                    Files.createDirectories(parent);
+                }
+                Path tmp = path.resolveSibling(path.getFileName() + ".tmp");
+                StringBuilder content = new StringBuilder();
+                for (String key : keep) {
+                    content.append(key).append(System.lineSeparator());
+                }
+                Files.writeString(tmp, content.toString(), StandardCharsets.UTF_8,
+                        StandardOpenOption.CREATE, StandardOpenOption.TRUNCATE_EXISTING);
+                try {
+                    Files.move(tmp, path, StandardCopyOption.REPLACE_EXISTING, StandardCopyOption.ATOMIC_MOVE);
+                } catch (IOException atomicUnsupported) {
+                    // 少数文件系统不支持 ATOMIC_MOVE，退化为普通替换
+                    Files.move(tmp, path, StandardCopyOption.REPLACE_EXISTING);
+                }
+            } catch (Exception ex) {
+                log.warn("[delete-journal] 重写待删对象日志失败 path={} msg={}", path, ex.getMessage());
+            }
+        }
+    }
+
+    /** 单行一条记录：去掉换行避免破坏行结构。 */
+    private static String sanitize(String objectKey) {
+        return objectKey.replace('\n', ' ').replace('\r', ' ').trim();
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/config/AdminApiGuardFilter.java

@@ -79,6 +79,9 @@ public class AdminApiGuardFilter extends OncePerRequestFilter {
             "/api/shop-data-crawl",
             "/api/withdraw",
             "/api/task-file-jobs",
+            // 2026-09 复核补：该前缀全部接口已在 controller 内 requireAdmin（含可换取员工店铺登录
+            // 令牌的 /shops/open），此处纳入名单是"开关打开后的第二层"；开关关闭时行为不变
+            "/api/ziniao",
             // /api/tasks/{taskId}/interrupted 仅凭 taskId 即可把 RUNNING 任务置为 FAILED，
             // 匿名遍历 taskId 就能批量打断线上任务
             "/api/tasks",

backend-java/src/main/java/com/nanri/aiimage/config/ModuleCleanupProperties.java

@@ -1,5 +1,6 @@
 package com.nanri.aiimage.config;
 
+import com.nanri.aiimage.common.module.TaskModuleRegistry;
 import lombok.Data;
 import org.springframework.boot.context.properties.ConfigurationProperties;
 
@@ -19,5 +20,6 @@ public class ModuleCleanupProperties {
     private int batchSize = 500;
     // SHOP_DATA_CRAWL keeps one per-shop daily workbook in its task service
     // and must not be removed by the age-based sweep.
-    private List<String> moduleTypes = new ArrayList<>(List.of("DEDUPE", "SPLIT", "CONVERT", "DELETE_BRAND", "PRODUCT_RISK_RESOLVE", "PRICE_TRACK", "SHOP_MATCH", "PATROL_DELETE", "QUERY_ASIN", "WITHDRAW", "APPEARANCE_PATENT", "SIMILAR_ASIN", "COLLECT_DATA"));
+    /** 参与按天清理的模块：取自模块注册表（G6）。 */
+    private List<String> moduleTypes = new ArrayList<>(TaskModuleRegistry.ageCleanupModuleTypes());
 }

backend-java/src/main/java/com/nanri/aiimage/config/TaskFileJobConfig.java

@@ -33,6 +33,7 @@ import io.micrometer.core.instrument.MeterRegistry;
 import org.springframework.beans.factory.ObjectProvider;
 import org.springframework.beans.factory.annotation.Qualifier;
 import org.springframework.beans.factory.annotation.Value;
+import com.nanri.aiimage.common.module.TaskModuleRegistry;
 import org.springframework.context.annotation.Bean;
 import org.springframework.context.annotation.Configuration;
 import org.springframework.core.task.TaskExecutor;
@@ -52,11 +53,11 @@ import java.util.concurrent.Semaphore;
 public class TaskFileJobConfig {
 
     /** 结果文件 Job 支持的全部 moduleType（启动校验枚举源，见 ResultFileJobHandlerRegistry.validateCoverage） */
-    public static final Set<String> RESULT_FILE_JOB_MODULE_TYPES = Set.of(
-            "SHOP_MATCH", "PRICE_TRACK", "PRODUCT_RISK_RESOLVE",
-            "PUBLISH", "QUERY_ASIN", "SHOP_DATA_CRAWL", "WITHDRAW",
-            "PATROL_DELETE", "APPEARANCE_PATENT", "SIMILAR_ASIN",
-            "DELETE_BRAND", "BRAND", "COLLECT_DATA");
+    /**
+     * 结果文件 Job 支持的 moduleType：取自模块注册表（G6），
+     * 与 {@code ResultFileJobHandlerRegistry.validateCoverage} 的启动自检配合使用。
+     */
+    public static final Set<String> RESULT_FILE_JOB_MODULE_TYPES = TaskModuleRegistry.resultFileJobModuleTypes();
 
     @Bean("taskFileJobDispatchExecutor")
     public TaskExecutor taskFileJobDispatchExecutor(

backend-java/src/main/java/com/nanri/aiimage/modules/appearancepatent/service/AppearancePatentTaskHeartbeatSpi.java

@@ -0,0 +1,30 @@
+package com.nanri.aiimage.modules.appearancepatent.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 外观专利检测 的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>心跳逻辑（缓存刷新 / 任务缓存回写）从 task 模块收回本模块，task 侧只依赖 SPI 接口，
+ * 消除 task → 业务模块的编译期依赖。
+ */
+@Service
+@RequiredArgsConstructor
+public class AppearancePatentTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final AppearancePatentTaskCacheService cacheService;
+
+    @Override
+    public String moduleType() {
+        return "APPEARANCE_PATENT";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.touchTaskHeartbeat(taskId);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/appearancepatent/service/support/AppearancePatentExcelParser.java

@@ -1,17 +1,17 @@
 package com.nanri.aiimage.modules.appearancepatent.service.support;
 
 import com.nanri.aiimage.common.exception.BusinessException;
+import com.nanri.aiimage.common.util.ExcelStreamReader;
 import lombok.extern.slf4j.Slf4j;
-import org.apache.poi.ss.usermodel.DataFormatter;
-import org.apache.poi.ss.usermodel.Row;
-import org.apache.poi.ss.usermodel.Sheet;
-import org.apache.poi.ss.usermodel.Workbook;
-import org.apache.poi.ss.usermodel.WorkbookFactory;
 
+import java.io.BufferedInputStream;
 import java.io.File;
 import java.io.FileInputStream;
+import java.io.IOException;
 import java.io.InputStream;
+import java.io.PushbackInputStream;
 import java.util.ArrayList;
+import java.util.Arrays;
 import java.util.LinkedHashMap;
 import java.util.List;
 import java.util.Locale;
@@ -37,9 +37,8 @@ public class AppearancePatentExcelParser {
         if (input == null) {
             throw new IllegalArgumentException("input must not be null");
         }
-        try (FileInputStream fis = new FileInputStream(input);
-             Workbook workbook = WorkbookFactory.create(fis)) {
-            return parseWorkbook(workbook, maxRows);
+        try (InputStream inputStream = new FileInputStream(input)) {
+            return readStreaming(inputStream, maxRows);
         } catch (BusinessException ex) {
             throw ex;
         } catch (Exception ex) {
@@ -52,8 +51,8 @@ public class AppearancePatentExcelParser {
         if (input == null) {
             throw new IllegalArgumentException("input must not be null");
         }
-        try (Workbook workbook = WorkbookFactory.create(input)) {
-            return parseWorkbook(workbook, DEFAULT_MAX_ROWS);
+        try {
+            return readStreaming(input, DEFAULT_MAX_ROWS);
         } catch (BusinessException ex) {
             throw ex;
         } catch (Exception ex) {
@@ -62,83 +61,154 @@ public class AppearancePatentExcelParser {
         }
     }
 
-    private ParsedSheet parseWorkbook(Workbook workbook, int maxRows) {
-        int safeMaxRows = Math.max(1, maxRows);
-        DataFormatter formatter = new DataFormatter();
-        Sheet sheet = workbook.getSheetAt(0);
-        Row header = sheet.getRow(0);
-        if (header == null) {
+    /**
+     * 流式解析（EasyExcel SAX），替代 POI WorkbookFactory 全量 DOM 加载：
+     * 大表不再整表驻留堆内存，行数上限在迭代过程中即时生效（超限抛错）。
+     * 语义与原 POI 路径一致：cell 归一化、表头别名匹配、空行跳过、必填表头缺失抛错、无字段截断。
+     * 注意不关闭传入的 InputStream（由调用方负责）。
+     */
+    private ParsedSheet readStreaming(InputStream inputStream, int maxRows) throws Exception {
+        // EasyExcel 会把非 zip 文本当 CSV 解析成功；原 WorkbookFactory 只认 xlsx/xls，
+        // 这里先做文件魔数校验，保持「垃圾文件→解析 Excel 失败」的语义并拒绝 CSV 误解析。
+        PushbackInputStream pb = new PushbackInputStream(new BufferedInputStream(inputStream, 8192), 8);
+        requireExcelMagic(pb);
+        SheetContext ctx = new SheetContext(Math.max(1, maxRows));
+        ExcelStreamReader.readFirstSheet(pb, new ExcelStreamReader.SheetRowHandler() {
+            @Override
+            public void onHeader(String sheetName, Integer sheetNo, Map<Integer, String> headerMap) {
+                ctx.initHeader(headerMap);
+            }
+
+            @Override
+            public void onRow(String sheetName, Integer sheetNo, int rowIndex,
+                              Map<Integer, String> headerMap, Map<Integer, String> rowMap) {
+                ctx.consumeRow(rowIndex, rowMap);
+            }
+        });
+        if (ctx.headers == null) {
+            // 空表 / 表头行整行为空（EasyExcel 不上报表头回调）：与旧实现 header == null 一致
             throw new BusinessException("Excel 表头为空");
         }
-        Map<String, Integer> headerMap = buildHeaderMap(header, formatter);
-        List<String> headers = readHeaders(header, formatter);
-        int idCol = findRequiredHeader(headerMap, "id");
-        int asinCol = findRequiredHeader(headerMap, "asin");
-        int countryCol = findRequiredHeader(headerMap, "国家", "country");
-        int priceCol = findOptionalHeaderExact(headerMap, "价格", "price");
-        int skuCol = findOptionalHeaderExact(headerMap,
-                "sku", "seller sku", "seller_sku", "merchant sku", "merchant_sku", "商品sku", "商品 sku", "库存sku");
-        int urlCol = findOptionalHeaderExact(headerMap,
-                "url", "rul", "link", "image", "img", "pic", "picture",
-                "链接", "商品链接", "图片", "商品图片", "主图", "商品主图", "图片链接", "主图链接");
-        int titleCol = findOptionalHeaderExact(headerMap,
-                "标题", "title", "listing title", "product title", "商品标题", "商品名称", "产品名称");
-
-        List<AppearanceExcelRow> rows = new ArrayList<>();
-        for (int i = 1; i <= sheet.getLastRowNum(); i++) {
-            Row row = sheet.getRow(i);
-            if (row == null) {
-                continue;
-            }
-            String id = cell(row, idCol, formatter);
-            String asin = cell(row, asinCol, formatter).toUpperCase(Locale.ROOT);
-            String country = cell(row, countryCol, formatter);
-            if (id.isBlank() && asin.isBlank() && country.isBlank()) {
-                continue;
-            }
-            if (rows.size() >= safeMaxRows) {
-                throw new BusinessException("解析行数超过上限: " + safeMaxRows);
-            }
-            rows.add(new AppearanceExcelRow(
-                    i + 1,
-                    id,
-                    asin,
-                    country,
-                    priceCol >= 0 ? cell(row, priceCol, formatter) : "",
-                    skuCol >= 0 ? cell(row, skuCol, formatter) : "",
-                    urlCol >= 0 ? cell(row, urlCol, formatter) : "",
-                    titleCol >= 0 ? cell(row, titleCol, formatter) : "",
-                    readRowValues(row, headers, formatter)));
-        }
-        return new ParsedSheet(headers, rows);
+        return new ParsedSheet(ctx.headers, ctx.rows);
     }
 
-    private Map<String, Integer> buildHeaderMap(Row header, DataFormatter formatter) {
+    /** 校验文件头魔数：xlsx=PK(zip)、xls=OLE2(CFD)。判非抛「解析 Excel 失败」；用 unread 回退已读字节。 */
+    private void requireExcelMagic(PushbackInputStream in) throws IOException {
+        byte[] head = new byte[8];
+        int n = 0;
+        while (n < head.length) {
+            int r = in.read(head, n, head.length - n);
+            if (r < 0) {
+                break;
+            }
+            n += r;
+        }
+        if (n > 0) {
+            in.unread(head, 0, n);
+        }
+        boolean isZip = n >= 4 && head[0] == 'P' && head[1] == 'K' && (head[2] == 3 || head[2] == 5 || head[2] == 7);
+        boolean isOle2 = n >= 8
+                && (head[0] & 0xFF) == 0xD0 && (head[1] & 0xFF) == 0xCF
+                && (head[2] & 0xFF) == 0x11 && (head[3] & 0xFF) == 0xE0;
+        if (!isZip && !isOle2) {
+            log.warn("[appearance-patent] parse rejected non-excel magic head={}", Arrays.copyOf(head, Math.max(n, 0)));
+            throw new BusinessException("解析 Excel 失败");
+        }
+    }
+
+    /** 单表解析上下文：表头就绪后逐行累积结果行。 */
+    private final class SheetContext {
+
+        private final int maxRows;
+        private final List<AppearanceExcelRow> rows = new ArrayList<>();
+        private List<String> headers;
+        private int idCol;
+        private int asinCol;
+        private int countryCol;
+        private int priceCol;
+        private int skuCol;
+        private int urlCol;
+        private int titleCol;
+
+        SheetContext(int maxRows) {
+            this.maxRows = maxRows;
+        }
+
+        void initHeader(Map<Integer, String> rawHeaderMap) {
+            // EasyExcel 回调给出 列号 → 表头文本，与旧 Row 遍历等价（缺列一般为 null/空串）
+            int lastColumnCount = lastColumnCount(rawHeaderMap);
             Map<String, Integer> map = new LinkedHashMap<>();
-        for (int i = 0; i < header.getLastCellNum(); i++) {
-            String val = normalize(formatter.formatCellValue(header.getCell(i)));
+            List<String> headerNames = new ArrayList<>();
+            for (int i = 0; i < lastColumnCount; i++) {
+                String val = normalize(rawHeaderMap.getOrDefault(i, ""));
+                headerNames.add(val.isBlank() ? "列" + (i + 1) : val);
                 if (!val.isBlank()) {
                     map.putIfAbsent(val.toLowerCase(Locale.ROOT), i);
                 }
             }
-        return map;
+            this.headers = headerNames;
+            this.idCol = findRequiredHeader(map, "id");
+            this.asinCol = findRequiredHeader(map, "asin");
+            this.countryCol = findRequiredHeader(map, "国家", "country");
+            this.priceCol = findOptionalHeaderExact(map, "价格", "price");
+            this.skuCol = findOptionalHeaderExact(map,
+                    "sku", "seller sku", "seller_sku", "merchant sku", "merchant_sku", "商品sku", "商品 sku", "库存sku");
+            this.urlCol = findOptionalHeaderExact(map,
+                    "url", "rul", "link", "image", "img", "pic", "picture",
+                    "链接", "商品链接", "图片", "商品图片", "主图", "商品主图", "图片链接", "主图链接");
+            this.titleCol = findOptionalHeaderExact(map,
+                    "标题", "title", "listing title", "product title", "商品标题", "商品名称", "产品名称");
         }
 
-    private List<String> readHeaders(Row header, DataFormatter formatter) {
-        List<String> headers = new ArrayList<>();
-        for (int i = 0; i < header.getLastCellNum(); i++) {
-            String val = normalize(formatter.formatCellValue(header.getCell(i)));
-            headers.add(val.isBlank() ? "列" + (i + 1) : val);
+        void consumeRow(int rowIndex, Map<Integer, String> rowMap) {
+            if (headers == null) {
+                return;
             }
-        return headers;
+            // EasyExcel rowIndex 从 0 起（0 为表头），POI 原实现行号同样 0 起并 +1 展示
+            String id = streamCell(rowMap, idCol);
+            String asin = streamCell(rowMap, asinCol).toUpperCase(Locale.ROOT);
+            String country = streamCell(rowMap, countryCol);
+            if (id.isBlank() && asin.isBlank() && country.isBlank()) {
+                return;
+            }
+            if (rows.size() >= maxRows) {
+                throw new BusinessException("解析行数超过上限: " + maxRows);
             }
-
-    private Map<String, String> readRowValues(Row row, List<String> headers, DataFormatter formatter) {
             Map<String, String> values = new LinkedHashMap<>();
             for (int i = 0; i < headers.size(); i++) {
-            values.put(headers.get(i), cell(row, i, formatter));
+                values.put(headers.get(i), streamCell(rowMap, i));
+            }
+            rows.add(new AppearanceExcelRow(
+                    rowIndex + 1,
+                    id,
+                    asin,
+                    country,
+                    priceCol >= 0 ? streamCell(rowMap, priceCol) : "",
+                    skuCol >= 0 ? streamCell(rowMap, skuCol) : "",
+                    urlCol >= 0 ? streamCell(rowMap, urlCol) : "",
+                    titleCol >= 0 ? streamCell(rowMap, titleCol) : "",
+                    values));
+        }
+
+        private int lastColumnCount(Map<Integer, String> rowMap) {
+            if (rowMap == null || rowMap.isEmpty()) {
+                return 0;
+            }
+            int lastColumnCount = 0;
+            for (Integer columnIndex : rowMap.keySet()) {
+                if (columnIndex != null && columnIndex >= lastColumnCount) {
+                    lastColumnCount = columnIndex + 1;
+                }
+            }
+            return lastColumnCount;
+        }
+
+        private String streamCell(Map<Integer, String> rowMap, int col) {
+            if (col < 0 || rowMap == null) {
+                return "";
+            }
+            return normalize(rowMap.get(col));
         }
-        return values;
     }
 
     private int findRequiredHeader(Map<String, Integer> map, String... names) {
@@ -177,10 +247,6 @@ public class AppearancePatentExcelParser {
         return normalized.replaceAll("[\\s_\\-()（）\\[\\]{}:：/\\\\]+", "");
     }
 
-    private String cell(Row row, int col, DataFormatter formatter) {
-        return col < 0 ? "" : normalize(formatter.formatCellValue(row.getCell(col)));
-    }
-
     private String normalize(String val) {
         return val == null ? "" : val.replace(String.valueOf((char) 0xFEFF), "").replace((char) 0x3000, ' ').trim().replaceAll("\\s+", " ");
     }

backend-java/src/main/java/com/nanri/aiimage/modules/brand/service/BrandCrawlTaskHeartbeatSpi.java

@@ -0,0 +1,97 @@
+package com.nanri.aiimage.modules.brand.service;
+
+import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
+import com.baomidou.mybatisplus.core.conditions.update.LambdaUpdateWrapper;
+import com.nanri.aiimage.modules.brand.mapper.BrandCrawlTaskMapper;
+import com.nanri.aiimage.modules.brand.model.entity.BrandCrawlTaskEntity;
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.vo.TaskHeartbeatVo;
+import com.nanri.aiimage.modules.task.spi.BrandTaskHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.stereotype.Service;
+
+import java.time.LocalDateTime;
+
+/**
+ * 品牌检测任务（brand_crawl_tasks）的心跳/中断实现（2026-09 全维度审查 G5）。
+ *
+ * <p>品牌任务有独立的表与状态字面量（running/pending/cancelled），原先这段逻辑散在
+ * {@code TaskHeartbeatService} 里并直接依赖本模块的 Mapper 与进度缓存；现整体收回本模块。
+ */
+@Service
+@RequiredArgsConstructor
+@Slf4j
+public class BrandCrawlTaskHeartbeatSpi implements BrandTaskHeartbeatSpi {
+
+    private static final String MODULE_BRAND = "BRAND";
+    private static final String BRAND_STATUS_RUNNING = "running";
+
+    private final BrandCrawlTaskMapper brandCrawlTaskMapper;
+    private final BrandTaskProgressCacheService brandTaskProgressCacheService;
+
+    @Override
+    public TaskHeartbeatVo heartbeat(Long taskId, TaskHeartbeatRequest request) {
+        BrandCrawlTaskEntity task = selectTask(taskId);
+        if (task == null) {
+            return null;
+        }
+        String status = task.getStatus();
+        if (!BRAND_STATUS_RUNNING.equals(status)) {
+            log.warn("[task-heartbeat] brand task is not running taskId={} actualUserId={} status={}",
+                    task.getId(), task.getUserId(), status);
+            return TaskHeartbeatVo.notAlive(MODULE_BRAND, status, "task is not running");
+        }
+        int updated = brandCrawlTaskMapper.update(null, new LambdaUpdateWrapper<BrandCrawlTaskEntity>()
+                .eq(BrandCrawlTaskEntity::getId, task.getId())
+                .eq(BrandCrawlTaskEntity::getStatus, BRAND_STATUS_RUNNING)
+                .set(BrandCrawlTaskEntity::getUpdatedAt, LocalDateTime.now()));
+        if (updated <= 0) {
+            BrandCrawlTaskEntity latest = brandCrawlTaskMapper.selectById(task.getId());
+            log.warn("[task-heartbeat] brand task heartbeat update missed taskId={} actualUserId={} status={} latestStatus={}",
+                    task.getId(), task.getUserId(), status,
+                    latest == null ? null : latest.getStatus());
+            return TaskHeartbeatVo.notAlive(MODULE_BRAND, latest == null ? status : latest.getStatus(),
+                    "task is not running");
+        }
+        brandTaskProgressCacheService.touchHeartbeat(
+                task.getId(),
+                request == null ? null : request.getPhase(),
+                request == null ? null : request.getCurrent(),
+                request == null ? null : request.getTotal());
+        return TaskHeartbeatVo.alive(MODULE_BRAND, BRAND_STATUS_RUNNING);
+    }
+
+    @Override
+    public TaskHeartbeatVo markInterrupted(Long taskId, String reason) {
+        BrandCrawlTaskEntity task = selectTask(taskId);
+        if (task == null) {
+            return null;
+        }
+        String status = task.getStatus();
+        if ("running".equalsIgnoreCase(status) || "pending".equalsIgnoreCase(status)) {
+            // 与 file 分支统一改为条件更新：整行 updateById 会拿读取快照覆盖并发写入的字段
+            // （客户端重启上报与品牌任务自身状态流转同时发生时）
+            int updated = brandCrawlTaskMapper.update(null, new LambdaUpdateWrapper<BrandCrawlTaskEntity>()
+                    .eq(BrandCrawlTaskEntity::getId, task.getId())
+                    .in(BrandCrawlTaskEntity::getStatus, "running", "pending", "RUNNING", "PENDING")
+                    .set(BrandCrawlTaskEntity::getStatus, "cancelled")
+                    .set(BrandCrawlTaskEntity::getErrorMessage, reason)
+                    .set(BrandCrawlTaskEntity::getUpdatedAt, LocalDateTime.now()));
+            if (updated > 0) {
+                log.warn("[task-interrupted] brand task marked cancelled by client restart taskId={} reason={}",
+                        taskId, reason);
+                return TaskHeartbeatVo.notAlive(MODULE_BRAND, "cancelled", "marked cancelled");
+            }
+        }
+        log.info("[task-interrupted] brand task not in running/pending, skipped taskId={} status={}", taskId, status);
+        return TaskHeartbeatVo.notAlive(MODULE_BRAND, status, "task is not running");
+    }
+
+    private BrandCrawlTaskEntity selectTask(Long taskId) {
+        LambdaQueryWrapper<BrandCrawlTaskEntity> brandQuery = new LambdaQueryWrapper<BrandCrawlTaskEntity>()
+                .eq(BrandCrawlTaskEntity::getId, taskId)
+                .last("limit 1");
+        return brandCrawlTaskMapper.selectOne(brandQuery);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/brand/service/BrandTaskService.java

@@ -8,6 +8,7 @@ import com.fasterxml.jackson.core.type.TypeReference;
 import com.fasterxml.jackson.databind.ObjectMapper;
 import com.nanri.aiimage.common.exception.BusinessException;
 import com.nanri.aiimage.common.service.DistributedJobLockService;
+import com.nanri.aiimage.common.util.ExcelStreamReader;
 import com.nanri.aiimage.config.BrandProgressProperties;
 import com.nanri.aiimage.config.HttpClientPool;
 import com.nanri.aiimage.config.StorageProperties;
@@ -39,12 +40,7 @@ import com.nanri.aiimage.modules.task.service.TaskProgressSnapshotService;
 import jakarta.annotation.PreDestroy;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
-import org.apache.poi.ss.usermodel.Cell;
-import org.apache.poi.ss.usermodel.DataFormatter;
-import org.apache.poi.ss.usermodel.Row;
 import org.apache.poi.ss.usermodel.Sheet;
-import org.apache.poi.ss.usermodel.Workbook;
-import org.apache.poi.ss.usermodel.WorkbookFactory;
 import org.apache.poi.xssf.streaming.SXSSFWorkbook;
 import org.springframework.stereotype.Service;
 
@@ -882,54 +878,60 @@ public class BrandTaskService {
     }
 
     private ParsedBrandFile parseBrandFile(File inputFile) {
-        DataFormatter formatter = new DataFormatter();
-        try (FileInputStream fis = new FileInputStream(inputFile); Workbook workbook = WorkbookFactory.create(fis)) {
-            Sheet sheet = workbook.getSheetAt(0);
-            Row headerRow = sheet.getRow(0);
-            if (headerRow == null) {
-                throw new BusinessException("Excel 表头为空");
+        // 2026-09 全维度审查 C5：改逐行流式解析（ExcelStreamReader → EasyExcel SAX）。
+        // 此前 WorkbookFactory.create 把整表读成 DOM，品牌源文件（几十万行）会整表驻留堆内存。
+        // 解析结果（columns/rows/uniqueBrands/行号）与错误文案不变，行号沿用 sheet 绝对 0 基行号 +1。
+        if (!hasExcelMagic(inputFile)) {
+            log.warn("[brand] 源文件不是 Excel（疑似 CSV/文本），拒绝解析 file={}", inputFile.getName());
+            throw new BusinessException("读取 Excel 失败");
         }
-            List<String> columns = extractHeaders(headerRow, formatter);
-            if (columns.isEmpty()) {
-                throw new BusinessException("未读取到有效表头");
-            }
-            Map<String, Integer> headerIndexes = buildHeaderIndexes(headerRow, formatter, columns);
-            List<Map<String, Object>> rows = new ArrayList<>();
-            Set<String> uniqueBrands = new LinkedHashSet<>();
-            for (int rowNum = 1; rowNum <= sheet.getLastRowNum(); rowNum++) {
-                Row row = sheet.getRow(rowNum);
-                if (row == null) {
-                    continue;
-                }
-                Map<String, Object> rowData = new LinkedHashMap<>();
-                for (String column : columns) {
-                    Integer index = headerIndexes.get(column);
-                    String value = index == null ? "" : normalizeCellText(formatter.formatCellValue(row.getCell(index)));
-                    rowData.put(column, value);
-                }
-                String brand = normalizeCellText(Objects.toString(rowData.getOrDefault("品牌", ""), ""));
-                if (brand.isBlank()) {
-                    rowData.put("__rowIndex", rowNum + 1);
-                    rows.add(rowData);
-                    continue;
-                }
-                if (uniqueBrands.add(brand)) {
-                    rowData.put("__rowIndex", rowNum + 1);
-                    rows.add(rowData);
-                }
-            }
-            return new ParsedBrandFile(sheet.getSheetName(), columns, rows, new ArrayList<>(uniqueBrands));
+        BrandFileSheetContext context = new BrandFileSheetContext();
+        try {
+            ExcelStreamReader.readFirstSheet(inputFile, context);
+        } catch (BusinessException ex) {
+            throw ex;
         } catch (IOException ex) {
             throw new BusinessException("读取 Excel 失败");
         }
+        return context.finish();
     }
 
-    private List<String> extractHeaders(Row headerRow, DataFormatter formatter) {
+    /**
+     * 文件魔数校验：xlsx=PK(zip)、xls=OLE2(CFD)。
+     * EasyExcel 会把非 zip 文本当 CSV 静默解析成功，而旧 POI WorkbookFactory 只认 xlsx/xls（垃圾文件直接失败），
+     * 这里保持「非 Excel 源文件 → 读取失败」的语义（与 SimilarAsinExcelParser 同口径）。
+     */
+    private boolean hasExcelMagic(File file) {
+        byte[] head = new byte[8];
+        try (InputStream inputStream = new FileInputStream(file)) {
+            int n = inputStream.readNBytes(head, 0, head.length);
+            boolean isZip = n >= 4 && head[0] == 'P' && head[1] == 'K' && (head[2] == 3 || head[2] == 5 || head[2] == 7);
+            boolean isOle2 = n >= 8
+                    && (head[0] & 0xFF) == 0xD0 && (head[1] & 0xFF) == 0xCF
+                    && (head[2] & 0xFF) == 0x11 && (head[3] & 0xFF) == 0xE0;
+            return isZip || isOle2;
+        } catch (IOException ex) {
+            return false;
+        }
+    }
+
+    /** 品牌源文件流式解析上下文：表头解析 + 按品牌去重累积结果行（内存不随整表 DOM 放大）。 */
+    private final class BrandFileSheetContext implements ExcelStreamReader.SheetRowHandler {
+
+        private final List<Map<String, Object>> rows = new ArrayList<>();
+        private final Set<String> uniqueBrands = new LinkedHashSet<>();
+        private List<String> columns;
+        private Map<String, Integer> headerIndexes;
+        private String sheetName = "";
+
+        @Override
+        public void onHeader(String sheetName, Integer sheetNo, Map<Integer, String> headerMap) {
+            this.sheetName = sheetName == null ? "" : sheetName;
+            int lastColumnCount = lastColumnCount(headerMap);
             List<String> headers = new ArrayList<>();
             Set<String> seen = new LinkedHashSet<>();
-        for (int i = 0; i < headerRow.getLastCellNum(); i++) {
-            Cell cell = headerRow.getCell(i);
-            String value = normalizeHeaderValue(cell == null ? null : formatter.formatCellValue(cell));
+            for (int i = 0; i < lastColumnCount; i++) {
+                String value = normalizeHeaderValue(headerMap.get(i));
                 if (value.isBlank() || seen.contains(value)) {
                     continue;
                 }
@@ -939,24 +941,70 @@ public class BrandTaskService {
                     break;
                 }
             }
-        return headers;
+            if (headers.isEmpty()) {
+                throw new BusinessException("未读取到有效表头");
             }
-
-    private Map<String, Integer> buildHeaderIndexes(Row headerRow, DataFormatter formatter, List<String> columns) {
-        Map<String, Integer> headerIndexes = new LinkedHashMap<>();
-        Set<String> allowed = new LinkedHashSet<>(columns);
-        for (int i = 0; i < headerRow.getLastCellNum(); i++) {
-            Cell cell = headerRow.getCell(i);
-            String value = normalizeHeaderValue(cell == null ? null : formatter.formatCellValue(cell));
-            if (value.isBlank() || headerIndexes.containsKey(value) || !allowed.contains(value)) {
+            Map<String, Integer> indexes = new LinkedHashMap<>();
+            Set<String> allowed = new LinkedHashSet<>(headers);
+            for (int i = 0; i < lastColumnCount; i++) {
+                String value = normalizeHeaderValue(headerMap.get(i));
+                if (value.isBlank() || indexes.containsKey(value) || !allowed.contains(value)) {
                     continue;
                 }
-            headerIndexes.put(value, i);
+                indexes.put(value, i);
                 if ("缩略图地址8".equals(value)) {
                     break;
                 }
             }
-        return headerIndexes;
+            this.columns = headers;
+            this.headerIndexes = indexes;
+        }
+
+        @Override
+        public void onRow(String sheetName, Integer sheetNo, int rowIndex,
+                          Map<Integer, String> headerMap, Map<Integer, String> rowMap) {
+            if (columns == null) {
+                // 表头行整行为空时 EasyExcel 不上报表头回调：与旧实现 headerRow == null 一致
+                throw new BusinessException("Excel 表头为空");
+            }
+            Map<String, Object> rowData = new LinkedHashMap<>();
+            for (String column : columns) {
+                Integer index = headerIndexes.get(column);
+                String value = index == null ? "" : normalizeCellText(rowMap.get(index));
+                rowData.put(column, value);
+            }
+            String brand = normalizeCellText(Objects.toString(rowData.getOrDefault("品牌", ""), ""));
+            if (brand.isBlank()) {
+                rowData.put("__rowIndex", rowIndex + 1);
+                rows.add(rowData);
+                return;
+            }
+            if (uniqueBrands.add(brand)) {
+                rowData.put("__rowIndex", rowIndex + 1);
+                rows.add(rowData);
+            }
+        }
+
+        private ParsedBrandFile finish() {
+            if (columns == null) {
+                // 空表（无任何行）或表头行缺失：与旧实现 headerRow == null 一致
+                throw new BusinessException("Excel 表头为空");
+            }
+            return new ParsedBrandFile(sheetName, columns, rows, new ArrayList<>(uniqueBrands));
+        }
+
+        private int lastColumnCount(Map<Integer, String> headerMap) {
+            if (headerMap == null || headerMap.isEmpty()) {
+                return 0;
+            }
+            int lastColumnCount = 0;
+            for (Integer columnIndex : headerMap.keySet()) {
+                if (columnIndex != null && columnIndex >= lastColumnCount) {
+                    lastColumnCount = columnIndex + 1;
+                }
+            }
+            return lastColumnCount;
+        }
     }
 
     private String normalizeHeaderValue(String value) {

backend-java/src/main/java/com/nanri/aiimage/modules/brand/service/BrandTaskStaleRepairSpiImpl.java

@@ -0,0 +1,47 @@
+package com.nanri.aiimage.modules.brand.service;
+
+import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
+import com.baomidou.mybatisplus.core.conditions.update.LambdaUpdateWrapper;
+import com.nanri.aiimage.modules.brand.mapper.BrandCrawlTaskMapper;
+import com.nanri.aiimage.modules.brand.model.entity.BrandCrawlTaskEntity;
+import com.nanri.aiimage.modules.task.spi.BrandTaskStaleRepairSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+import java.time.LocalDateTime;
+import java.util.List;
+
+/**
+ * brand_crawl_tasks 的陈旧修复实现（2026-09 全维度审查 G5：逻辑从 task 模块收回本模块）。
+ */
+@Service
+@RequiredArgsConstructor
+public class BrandTaskStaleRepairSpiImpl implements BrandTaskStaleRepairSpi {
+
+    private static final String STATUS_PENDING = "pending";
+    private static final String STATUS_RUNNING = "running";
+    private static final String STATUS_FAILED = "failed";
+
+    private final BrandCrawlTaskMapper brandCrawlTaskMapper;
+
+    @Override
+    public List<Long> failStaleBrandTasks(LocalDateTime cutoff, int limit) {
+        List<BrandCrawlTaskEntity> stale = brandCrawlTaskMapper.selectList(
+                new LambdaQueryWrapper<BrandCrawlTaskEntity>()
+                        .select(BrandCrawlTaskEntity::getId)
+                        .in(BrandCrawlTaskEntity::getStatus, STATUS_PENDING, STATUS_RUNNING)
+                        .lt(BrandCrawlTaskEntity::getUpdatedAt, cutoff)
+                        .last("limit " + Math.max(1, limit)));
+        if (stale.isEmpty()) {
+            return List.of();
+        }
+        List<Long> ids = stale.stream().map(BrandCrawlTaskEntity::getId).toList();
+        brandCrawlTaskMapper.update(null, new LambdaUpdateWrapper<BrandCrawlTaskEntity>()
+                .in(BrandCrawlTaskEntity::getId, ids)
+                .in(BrandCrawlTaskEntity::getStatus, STATUS_PENDING, STATUS_RUNNING)
+                .lt(BrandCrawlTaskEntity::getUpdatedAt, cutoff)
+                .set(BrandCrawlTaskEntity::getStatus, STATUS_FAILED)
+                .set(BrandCrawlTaskEntity::getErrorMessage, "任务长期无心跳，已自动失败"));
+        return ids;
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/service/CollectDataItemCleanupSpiImpl.java

@@ -0,0 +1,29 @@
+package com.nanri.aiimage.modules.collectdata.service;
+
+import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
+import com.nanri.aiimage.modules.collectdata.mapper.CollectDataItemMapper;
+import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataItemEntity;
+import com.nanri.aiimage.modules.task.spi.CollectDataItemCleanupSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+import java.util.List;
+
+/**
+ * 采集明细行的历史清理实现（2026-09 全维度审查 G5：逻辑从 task 模块收回本模块）。
+ */
+@Service
+@RequiredArgsConstructor
+public class CollectDataItemCleanupSpiImpl implements CollectDataItemCleanupSpi {
+
+    private final CollectDataItemMapper collectDataItemMapper;
+
+    @Override
+    public int deleteItemsByTaskIds(List<Long> taskIds) {
+        if (taskIds == null || taskIds.isEmpty()) {
+            return 0;
+        }
+        return collectDataItemMapper.delete(new LambdaQueryWrapper<CollectDataItemEntity>()
+                .in(CollectDataItemEntity::getTaskId, taskIds));
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/service/CollectDataTaskHeartbeatSpi.java

@@ -0,0 +1,28 @@
+package com.nanri.aiimage.modules.collectdata.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 采集数据的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>进度由采集模块自己的导入进度表维护，心跳直接把客户端上报的进度写进去。
+ */
+@Service
+@RequiredArgsConstructor
+public class CollectDataTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final CollectDataService collectDataService;
+
+    @Override
+    public String moduleType() {
+        return "COLLECT_DATA";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        collectDataService.updateProgress(taskId, request);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/dedupe/service/DedupeTotalDataService.java

@@ -3,6 +3,7 @@ package com.nanri.aiimage.modules.dedupe.service;
 import cn.hutool.core.util.IdUtil;
 import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.nanri.aiimage.common.exception.BusinessException;
+import com.nanri.aiimage.common.util.ExcelStreamReader;
 import com.nanri.aiimage.modules.dedupe.mapper.DedupeTotalDataMapper;
 import com.nanri.aiimage.modules.dedupe.model.dto.DedupeTotalDataCreateRequest;
 import com.nanri.aiimage.modules.dedupe.model.dto.DedupeTotalDataUpdateRequest;
@@ -122,7 +123,10 @@ public class DedupeTotalDataService {
         String safeCountry = country == null ? "" : country.trim();
         AccessScope scope = resolveAccessScope(operatorId);
         LambdaQueryWrapper<DedupeTotalDataEntity> query = new LambdaQueryWrapper<DedupeTotalDataEntity>()
-                .like(!safeKeyword.isEmpty(), DedupeTotalDataEntity::getDataValue, safeKeyword)
+                // 2026-09 全维度审查 C6：关键字由"包含"改为"前缀"匹配——后台该输入框标注的是
+                // 「输入 ASIN」（相邻的用户名框才是"模糊搜索"），前缀匹配才能命中 data_value 索引；
+                // 此前的 %kw% 前置通配使 idx 永不生效，93 万行表每翻页两次全表扫
+                .likeRight(!safeKeyword.isEmpty(), DedupeTotalDataEntity::getDataValue, safeKeyword)
                 .like(!safeUsername.isEmpty(), DedupeTotalDataEntity::getUploaderUsername, safeUsername)
                 .apply(!safeCountry.isEmpty(),
                         "FIND_IN_SET({0}, IFNULL(country, '')) > 0",
@@ -667,101 +671,162 @@ public class DedupeTotalDataService {
             throw new BusinessException("仅支持 .xlsx 或 .xls 文件");
         }
 
-        DataFormatter formatter = new DataFormatter();
-        try (Workbook workbook = org.apache.poi.ss.usermodel.WorkbookFactory.create(inputStream)) {
-            Sheet sheet = workbook.getSheetAt(0);
-            Row headerRow = sheet.getRow(0);
-            if (headerRow == null) {
-                throw new BusinessException("Excel 表头为空");
-            }
+        // 2026-09 全维度审查 C5：改逐行流式解析（ExcelStreamReader → EasyExcel SAX）。
+        // 此前 WorkbookFactory.create 把整表读成 DOM（50 万行约 1~2GB 堆），且行数上限在
+        // 加载完之后才校验（等于没有防线）。现在内存只保留单批待插入值与文件内去重集合，
+        // 行数上限在迭代过程中即时生效。
+        ImportSession session = new ImportSession(progress, groupId, uploaderUserId, uploaderUsername);
+        try {
+            ExcelStreamReader.readFirstSheet(inputStream, new ExcelStreamReader.SheetRowHandler() {
+                private Integer countryColumnIndex;
+                private boolean headerReady;
 
-            Map<String, Integer> headerMap = new HashMap<>();
-            for (int i = 0; i < headerRow.getLastCellNum(); i++) {
-                Cell cell = headerRow.getCell(i);
-                String value = normalizeExcelText(cell == null ? null : formatter.formatCellValue(cell));
-                if (!value.isBlank() && !headerMap.containsKey(value)) {
-                    headerMap.put(value, i);
-                }
-            }
-
-            Integer asinIndex = headerMap.get("ASIN");
-            if (asinIndex == null) {
+                @Override
+                public void onHeader(String sheetName, Integer sheetNo, Map<Integer, String> headerMap) {
+                    session.headerIndex = buildHeaderIndex(headerMap);
+                    if (session.headerIndex.get("ASIN") == null) {
                         throw new BusinessException("缺少 ASIN 列");
                     }
-            Integer countryIndex = headerMap.get("国家");
+                    countryColumnIndex = session.headerIndex.get("国家");
+                    headerReady = true;
+                }
 
-            Set<String> seenInFile = new HashSet<>();
-            List<String> pendingValues = new ArrayList<>(IMPORT_BATCH_SIZE);
-            Map<String, String> pendingCountries = new HashMap<>();
-            int totalRows = Math.max(sheet.getLastRowNum(), 0);
-            if (maxImportRows > 0 && totalRows > maxImportRows) {
+                @Override
+                public void onSheetTotal(String sheetName, Integer sheetNo, Integer approximateTotalRows) {
+                    if (progress != null && approximateTotalRows != null && approximateTotalRows > 0) {
+                        progress.setTotalRows(approximateTotalRows);
+                    }
+                }
+
+                @Override
+                public void onRow(String sheetName, Integer sheetNo, int rowIndex,
+                                  Map<Integer, String> headerMap, Map<Integer, String> rowMap) {
+                    if (!headerReady) {
+                        // 无表头（空表）：与旧 DOM 实现一致，不处理任何数据行
+                        return;
+                    }
+                    session.acceptRow(rowMap, countryColumnIndex);
+                }
+            });
+        } catch (BusinessException e) {
+            throw e;
+        } catch (Exception e) {
+            log.error("[dedupe-total-data] 导入 Excel 失败 filename={} processedRows={}",
+                    filename, session.processedRows(), e);
+            throw new BusinessException("导入 Excel 失败", e);
+        }
+        return session.finish();
+    }
+
+    /** 表头名 → 列索引（首次出现优先、忽略空表头），与旧 DOM 实现口径一致。 */
+    private Map<String, Integer> buildHeaderIndex(Map<Integer, String> headerMap) {
+        Map<String, Integer> index = new HashMap<>();
+        if (headerMap == null) {
+            return index;
+        }
+        for (Map.Entry<Integer, String> entry : headerMap.entrySet()) {
+            if (entry.getKey() == null) {
+                continue;
+            }
+            String value = normalizeExcelText(entry.getValue());
+            if (!value.isBlank() && !index.containsKey(value)) {
+                index.put(value, entry.getKey());
+            }
+        }
+        return index;
+    }
+
+    /** 流式解析下按列索引取行内文本。 */
+    private String cellText(Map<Integer, String> rowMap, Integer columnIndex) {
+        if (rowMap == null || columnIndex == null) {
+            return null;
+        }
+        return rowMap.get(columnIndex);
+    }
+
+    /** 流式导入的累加器：单批待插入值 + 计数与进度（内存只随批大小增长，不随文件行数增长）。 */
+    private final class ImportSession {
+
+        private final DedupeTotalDataImportProgressVo progress;
+        private final Long groupId;
+        private final Long uploaderUserId;
+        private final String uploaderUsername;
+        private final Set<String> seenInFile = new HashSet<>();
+        private final List<String> pendingValues = new ArrayList<>(IMPORT_BATCH_SIZE);
+        private final Map<String, String> pendingCountries = new HashMap<>();
+        private Map<String, Integer> headerIndex = Map.of();
+        private int processedRows;
+        private int asinCount;
+        private int insertedCount;
+        private int skippedCount;
+
+        private ImportSession(DedupeTotalDataImportProgressVo progress, Long groupId,
+                              Long uploaderUserId, String uploaderUsername) {
+            this.progress = progress;
+            this.groupId = groupId;
+            this.uploaderUserId = uploaderUserId;
+            this.uploaderUsername = uploaderUsername;
+        }
+
+        private void acceptRow(Map<Integer, String> rowMap, Integer countryColumnIndex) {
+            processedRows++;
+            if (maxImportRows > 0 && processedRows > maxImportRows) {
+                // 流式下无法前置知道总行数：超限即中断，已插入批次保留（与旧实现的"截断"语义不同，
+                // 但旧实现在加载后才校验，实际是 OOM 之后才失败）
                 throw new BusinessException("导入行数超过上限: " + maxImportRows);
             }
-            int asinCount = 0;
-            int insertedCount = 0;
-            int skippedCount = 0;
-            if (progress != null) {
-                progress.setTotalRows(totalRows);
-                progress.setProcessedRows(0);
-                progress.setAsinCount(0);
-                progress.setInsertedCount(0);
-                progress.setSkippedCount(0);
-                progress.setErrorMessage(null);
-            }
-
-            for (int rowNum = 1; rowNum <= sheet.getLastRowNum(); rowNum++) {
-                Row row = sheet.getRow(rowNum);
-                if (row == null) {
-                    skippedCount++;
-                    updateImportProgress(progress, rowNum, asinCount, insertedCount, skippedCount);
-                    continue;
-                }
-                String asin = normalizeComparableValueOrBlank(formatter.formatCellValue(row.getCell(asinIndex)));
+            Integer asinIndex = headerIndex.get("ASIN");
+            String asin = asinIndex == null ? "" : normalizeComparableValueOrBlank(cellText(rowMap, asinIndex));
             if (asin.isBlank()) {
                 skippedCount++;
-                    updateImportProgress(progress, rowNum, asinCount, insertedCount, skippedCount);
-                    continue;
+                updateImportProgress(progress, processedRows, asinCount, insertedCount, skippedCount);
+                return;
             }
             asinCount++;
             if (!seenInFile.add(asin)) {
                 skippedCount++;
-                    updateImportProgress(progress, rowNum, asinCount, insertedCount, skippedCount);
-                    continue;
+                updateImportProgress(progress, processedRows, asinCount, insertedCount, skippedCount);
+                return;
             }
             pendingValues.add(asin);
-                if (countryIndex != null) {
-                    pendingCountries.put(asin, parseCountries(formatter.formatCellValue(row.getCell(countryIndex))));
+            if (countryColumnIndex != null) {
+                pendingCountries.put(asin, parseCountries(cellText(rowMap, countryColumnIndex)));
             }
             if (pendingValues.size() >= IMPORT_BATCH_SIZE) {
-                    ImportBatchResult batch = insertImportBatch(
-                            pendingValues, pendingCountries, groupId, uploaderUserId, uploaderUsername);
-                    insertedCount += batch.insertedCount();
-                    skippedCount += batch.skippedCount();
-                    pendingValues.clear();
-                    pendingCountries.clear();
+                flushBatch();
             }
-                updateImportProgress(progress, rowNum, asinCount, insertedCount, skippedCount);
+            updateImportProgress(progress, processedRows, asinCount, insertedCount, skippedCount);
         }
-            if (!pendingValues.isEmpty()) {
-                ImportBatchResult batch = insertImportBatch(
-                        pendingValues, pendingCountries, groupId, uploaderUserId, uploaderUsername);
-                insertedCount += batch.insertedCount();
-                skippedCount += batch.skippedCount();
-                pendingValues.clear();
-                pendingCountries.clear();
-            }
-            updateImportProgress(progress, totalRows, asinCount, insertedCount, skippedCount);
 
+        private void flushBatch() {
+            if (pendingValues.isEmpty()) {
+                return;
+            }
+            ImportBatchResult batch = insertImportBatch(
+                    pendingValues, pendingCountries, groupId, uploaderUserId, uploaderUsername);
+            insertedCount += batch.insertedCount();
+            skippedCount += batch.skippedCount();
+            pendingValues.clear();
+            pendingCountries.clear();
+        }
+
+        private int processedRows() {
+            return processedRows;
+        }
+
+        private DedupeTotalDataImportVo finish() {
+            flushBatch();
+            updateImportProgress(progress, processedRows, asinCount, insertedCount, skippedCount);
+            if (progress != null && processedRows > progress.getTotalRows()) {
+                // 近似总行数缺失（或与实际物理行数不一致）时用实际处理行数兜底，避免展示成 x/0
+                progress.setTotalRows(processedRows);
+            }
             DedupeTotalDataImportVo vo = new DedupeTotalDataImportVo();
-            vo.setTotalRows(totalRows);
+            vo.setTotalRows(processedRows);
             vo.setAsinCount(asinCount);
             vo.setInsertedCount(insertedCount);
             vo.setSkippedCount(skippedCount);
             return vo;
-        } catch (BusinessException e) {
-            throw e;
-        } catch (Exception e) {
-            throw new BusinessException("导入 Excel 失败");
         }
     }
 
@@ -850,112 +915,120 @@ public class DedupeTotalDataService {
             throw new BusinessException("仅支持 .xlsx 或 .xls 文件");
         }
 
-        DataFormatter formatter = new DataFormatter();
-        try (Workbook workbook = org.apache.poi.ss.usermodel.WorkbookFactory.create(inputStream)) {
-            Sheet sheet = workbook.getSheetAt(0);
-            Row headerRow = sheet.getRow(0);
-            if (headerRow == null) {
-                throw new BusinessException("Excel 表头为空");
-            }
+        // 2026-09 全维度审查 C5：与导入同样改流式解析，避免 50 万行时 DOM 占 1~2GB 堆
+        DeleteSession session = new DeleteSession(progress, scope, groupId);
+        try {
+            ExcelStreamReader.readFirstSheet(inputStream, new ExcelStreamReader.SheetRowHandler() {
+                private boolean headerReady;
 
-            Map<String, Integer> headerMap = new HashMap<>();
-            for (int i = 0; i < headerRow.getLastCellNum(); i++) {
-                Cell cell = headerRow.getCell(i);
-                String value = normalizeExcelText(cell == null ? null : formatter.formatCellValue(cell));
-                if (!value.isBlank() && !headerMap.containsKey(value)) {
-                    headerMap.put(value, i);
-                }
-            }
-
-            Integer asinIndex = headerMap.get("ASIN");
-            if (asinIndex == null) {
+                @Override
+                public void onHeader(String sheetName, Integer sheetNo, Map<Integer, String> headerMap) {
+                    session.headerIndex = buildHeaderIndex(headerMap);
+                    if (session.headerIndex.get("ASIN") == null) {
                         throw new BusinessException("缺少 ASIN 列");
                     }
-
-            Set<String> seenInFile = new HashSet<>();
-            List<String> pendingDeletes = new ArrayList<>();
-            int totalRows = Math.max(sheet.getLastRowNum(), 0);
-            if (maxImportRows > 0 && totalRows > maxImportRows) {
-                throw new BusinessException("导入行数超过上限: " + maxImportRows);
-            }
-            int asinCount = 0;
-            int deletedCount = 0;
-            int skippedCount = 0;
-            if (progress != null) {
-                progress.setTotalRows(totalRows);
-                progress.setProcessedRows(0);
-                progress.setAsinCount(0);
-                progress.setInsertedCount(0);
-                progress.setSkippedCount(0);
-                progress.setErrorMessage(null);
+                    headerReady = true;
                 }
 
-            for (int rowNum = 1; rowNum <= sheet.getLastRowNum(); rowNum++) {
-                Row row = sheet.getRow(rowNum);
-                if (row == null) {
-                    skippedCount++;
-                    if (progress != null) {
-                        progress.setProcessedRows(rowNum);
-                        progress.setAsinCount(asinCount);
-                        progress.setInsertedCount(deletedCount);
-                        progress.setSkippedCount(skippedCount);
+                @Override
+                public void onSheetTotal(String sheetName, Integer sheetNo, Integer approximateTotalRows) {
+                    if (progress != null && approximateTotalRows != null && approximateTotalRows > 0) {
+                        progress.setTotalRows(approximateTotalRows);
                     }
-                    continue;
-                }
-                String asin = normalizeComparableValueOrBlank(formatter.formatCellValue(row.getCell(asinIndex)));
-                if (asin.isBlank()) {
-                    skippedCount++;
-                    if (progress != null) {
-                        progress.setProcessedRows(rowNum);
-                        progress.setAsinCount(asinCount);
-                        progress.setInsertedCount(deletedCount);
-                        progress.setSkippedCount(skippedCount);
-                    }
-                    continue;
-                }
-                asinCount++;
-                String dataValue = asin;
-                if (!seenInFile.add(dataValue)) {
-                    skippedCount++;
-                    if (progress != null) {
-                        progress.setProcessedRows(rowNum);
-                        progress.setAsinCount(asinCount);
-                        progress.setInsertedCount(deletedCount);
-                        progress.setSkippedCount(skippedCount);
-                    }
-                    continue;
                 }
 
-                // 批量删除：攒批 + IN 一次删，避免 50 万行 = 50 万个 REQUIRES_NEW 事务的 N+1
-                pendingDeletes.add(dataValue);
-                if (pendingDeletes.size() >= DELETE_BATCH_SIZE) {
-                    deletedCount += deleteBatchByDataValues(pendingDeletes, scope, groupId);
-                    pendingDeletes = new ArrayList<>();
+                @Override
+                public void onRow(String sheetName, Integer sheetNo, int rowIndex,
+                                  Map<Integer, String> headerMap, Map<Integer, String> rowMap) {
+                    if (!headerReady) {
+                        return;
                     }
-                if (progress != null) {
-                    progress.setProcessedRows(rowNum);
-                    progress.setAsinCount(asinCount);
-                    progress.setInsertedCount(deletedCount);
-                    progress.setSkippedCount(skippedCount);
+                    session.acceptRow(rowMap);
                 }
-            }
-            if (!pendingDeletes.isEmpty()) {
-                deletedCount += deleteBatchByDataValues(pendingDeletes, scope, groupId);
-            }
-
-            DedupeTotalDataImportVo vo = new DedupeTotalDataImportVo();
-            vo.setTotalRows(totalRows);
-            vo.setAsinCount(asinCount);
-            vo.setInsertedCount(deletedCount);
-            vo.setSkippedCount(skippedCount);
-            return vo;
+            });
         } catch (BusinessException e) {
             throw e;
         } catch (Exception e) {
             // 此前丢弃原始异常：排查时只能看到一句「删除 Excel 匹配数据失败」
-            log.error("[dedupe-total-data] 删除 Excel 匹配数据失败", e);
+            log.error("[dedupe-total-data] 删除 Excel 匹配数据失败 filename={} processedRows={}",
+                    filename, session.processedRows(), e);
             throw new BusinessException("删除 Excel 匹配数据失败", e);
         }
+        return session.finish();
+    }
+
+    /** 流式删除匹配的累加器：单批待删值 + 计数与进度。 */
+    private final class DeleteSession {
+
+        private final DedupeTotalDataImportProgressVo progress;
+        private final AccessScope scope;
+        private final Long groupId;
+        private final Set<String> seenInFile = new HashSet<>();
+        private final List<String> pendingDeletes = new ArrayList<>();
+        private Map<String, Integer> headerIndex = Map.of();
+        private int processedRows;
+        private int asinCount;
+        private int deletedCount;
+        private int skippedCount;
+
+        private DeleteSession(DedupeTotalDataImportProgressVo progress, AccessScope scope, Long groupId) {
+            this.progress = progress;
+            this.scope = scope;
+            this.groupId = groupId;
+        }
+
+        private void acceptRow(Map<Integer, String> rowMap) {
+            processedRows++;
+            if (maxImportRows > 0 && processedRows > maxImportRows) {
+                throw new BusinessException("导入行数超过上限: " + maxImportRows);
+            }
+            Integer asinIndex = headerIndex.get("ASIN");
+            String asin = asinIndex == null ? "" : normalizeComparableValueOrBlank(cellText(rowMap, asinIndex));
+            if (asin.isBlank() || !seenInFile.add(asin)) {
+                skippedCount++;
+                updateProgress();
+                return;
+            }
+            asinCount++;
+            // 批量删除：攒批 + IN 一次删，避免 50 万行 = 50 万个 REQUIRES_NEW 事务的 N+1
+            pendingDeletes.add(asin);
+            if (pendingDeletes.size() >= DELETE_BATCH_SIZE) {
+                deletedCount += deleteBatchByDataValues(pendingDeletes, scope, groupId);
+                pendingDeletes.clear();
+            }
+            updateProgress();
+        }
+
+        private void updateProgress() {
+            if (progress == null) {
+                return;
+            }
+            progress.setProcessedRows(processedRows);
+            progress.setAsinCount(asinCount);
+            progress.setInsertedCount(deletedCount);
+            progress.setSkippedCount(skippedCount);
+        }
+
+        private int processedRows() {
+            return processedRows;
+        }
+
+        private DedupeTotalDataImportVo finish() {
+            if (!pendingDeletes.isEmpty()) {
+                deletedCount += deleteBatchByDataValues(pendingDeletes, scope, groupId);
+                pendingDeletes.clear();
+            }
+            updateProgress();
+            if (progress != null && processedRows > progress.getTotalRows()) {
+                progress.setTotalRows(processedRows);
+            }
+            DedupeTotalDataImportVo vo = new DedupeTotalDataImportVo();
+            vo.setTotalRows(processedRows);
+            vo.setAsinCount(asinCount);
+            vo.setInsertedCount(deletedCount);
+            vo.setSkippedCount(skippedCount);
+            return vo;
+        }
     }
 
     @Transactional

backend-java/src/main/java/com/nanri/aiimage/modules/deletebrand/service/DeleteBrandRunService.java

@@ -7,6 +7,7 @@ import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.fasterxml.jackson.core.type.TypeReference;
 import com.fasterxml.jackson.databind.ObjectMapper;
 import com.nanri.aiimage.common.exception.BusinessException;
+import com.nanri.aiimage.common.util.ExcelStreamReader;
 import com.nanri.aiimage.config.DeleteBrandProgressProperties;
 import com.nanri.aiimage.config.TaskPressureProperties;
 import com.nanri.aiimage.modules.deletebrand.model.cache.DeleteBrandParsedFileCacheDto;
@@ -44,10 +45,8 @@ import com.nanri.aiimage.modules.ziniao.service.ZiniaoShopSwitchService;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
 import org.apache.poi.ss.usermodel.Cell;
-import org.apache.poi.ss.usermodel.DataFormatter;
 import org.apache.poi.ss.usermodel.Row;
 import org.apache.poi.ss.usermodel.Sheet;
-import org.apache.poi.ss.usermodel.Workbook;
 import org.apache.poi.xssf.streaming.SXSSFWorkbook;
 import org.springframework.stereotype.Service;
 import org.springframework.transaction.PlatformTransactionManager;
@@ -59,6 +58,7 @@ import org.springframework.transaction.support.TransactionTemplate;
 import java.io.File;
 import java.io.FileInputStream;
 import java.io.FileOutputStream;
+import java.io.IOException;
 import java.time.Duration;
 import java.time.LocalDateTime;
 import java.util.ArrayList;
@@ -515,39 +515,102 @@ public class DeleteBrandRunService {
     }
 
     private ParsedDeleteBrandFile parseDeleteBrandFile(File inputFile) {
-        DataFormatter formatter = new DataFormatter();
-        try (FileInputStream fis = new FileInputStream(inputFile); Workbook workbook = org.apache.poi.ss.usermodel.WorkbookFactory.create(fis)) {
-            Sheet sheet = workbook.getSheetAt(0);
-            Row titleRow = sheet.getRow(0);
-            Row headerRow = sheet.getRow(1);
-            if (titleRow == null || headerRow == null) {
-                throw new BusinessException("Excel 表头不完整");
+        // 2026-09 全维度审查 C5：改逐行流式解析（ExcelStreamReader → EasyExcel SAX）。
+        // 此前 WorkbookFactory.create 把整表读成 DOM，删除品牌源文件（几十万行）会整表驻留堆内存。
+        // 解析结果（国家分组/预览行/行号）与错误文案不变，行号沿用 sheet 绝对 0 基行号 +1。
+        if (!hasExcelMagic(inputFile)) {
+            log.warn("[delete-brand] 源文件不是 Excel（疑似 CSV/文本），拒绝解析 file={}", inputFile.getName());
+            throw new BusinessException("读取删除品牌 Excel 失败");
+        }
+        DeleteBrandSheetContext context = new DeleteBrandSheetContext();
+        try {
+            ExcelStreamReader.readFirstSheet(inputFile, context);
+        } catch (BusinessException ex) {
+            throw ex;
+        } catch (Exception ex) {
+            // 此前该 catch 既不记日志也不带 cause，线上只有一句「读取删除品牌 Excel 失败」
+            log.warn("[delete-brand] 读取删除品牌 Excel 失败 file={} err={}", inputFile.getName(), ex.getMessage(), ex);
+            throw new BusinessException("读取删除品牌 Excel 失败");
+        }
+        return context.finish();
     }
 
-            List<CountryColumnPair> pairs = resolveCountryPairs(titleRow, headerRow, formatter);
+    /**
+     * 文件魔数校验：xlsx=PK(zip)、xls=OLE2(CFD)。
+     * EasyExcel 会把非 zip 文本当 CSV 静默解析成功，而旧 POI WorkbookFactory 只认 xlsx/xls（垃圾文件直接失败），
+     * 这里保持「非 Excel 源文件 → 读取失败」的语义（与 SimilarAsinExcelParser 同口径）。
+     */
+    private boolean hasExcelMagic(File file) {
+        byte[] head = new byte[8];
+        try (FileInputStream inputStream = new FileInputStream(file)) {
+            int n = inputStream.readNBytes(head, 0, head.length);
+            boolean isZip = n >= 4 && head[0] == 'P' && head[1] == 'K' && (head[2] == 3 || head[2] == 5 || head[2] == 7);
+            boolean isOle2 = n >= 8
+                    && (head[0] & 0xFF) == 0xD0 && (head[1] & 0xFF) == 0xCF
+                    && (head[2] & 0xFF) == 0x11 && (head[3] & 0xFF) == 0xE0;
+            return isZip || isOle2;
+        } catch (IOException ex) {
+            return false;
+        }
+    }
+
+    /**
+     * 删除品牌源文件流式解析上下文：两行表头（国家标题行 + 字段行）解析列组 + 逐行按国家收集 ASIN。
+     * dataStartRowIndex 语义与旧实现一致（删除ASIN 表头记为 2，其余"状态"表头记为 1）。
+     */
+    private final class DeleteBrandSheetContext implements ExcelStreamReader.SheetRowHandler {
+
+        private final Map<String, Map<String, DeleteBrandCountryAsinVo>> grouped = new LinkedHashMap<>();
+        private final Map<String, String> displayCountryNames = new LinkedHashMap<>();
+        private Map<Integer, String> titleRowMap;
+        private boolean titleRowSeen;
+        private boolean headerRowSeen;
+        private List<CountryColumnPair> pairs;
+        private int firstDataRow = 2;
+
+        @Override
+        public void onHeader(String sheetName, Integer sheetNo, Map<Integer, String> headerMap) {
+            // 第 0 行 = 国家标题行
+            titleRowSeen = true;
+            titleRowMap = headerMap == null ? Map.of() : headerMap;
+        }
+
+        @Override
+        public void onRow(String sheetName, Integer sheetNo, int rowIndex,
+                          Map<Integer, String> headerMap, Map<Integer, String> rowMap) {
+            if (!headerRowSeen) {
+                if (rowIndex != 1) {
+                    // 第 1 行整行为空（EasyExcel 不上报）：与旧实现 headerRow == null 一致
+                    throw new BusinessException("Excel 表头不完整");
+                }
+                if (!titleRowSeen) {
+                    // 第 0 行整行为空（EasyExcel 不上报表头回调）：与旧实现 titleRow == null 一致
+                    throw new BusinessException("Excel 表头不完整");
+                }
+                headerRowSeen = true;
+                pairs = resolveCountryPairs(titleRowMap, rowMap);
                 if (pairs.isEmpty()) {
                     throw new BusinessException("未识别到删除品牌表头");
                 }
-
-            Map<String, Map<String, DeleteBrandCountryAsinVo>> grouped = new LinkedHashMap<>();
-            Map<String, String> displayCountryNames = new LinkedHashMap<>();
-
-            int firstDataRow = pairs.stream()
+                firstDataRow = pairs.stream()
                         .mapToInt(CountryColumnPair::dataStartRowIndex)
                         .min()
                         .orElse(2);
-            for (int rowNum = firstDataRow; rowNum <= sheet.getLastRowNum(); rowNum++) {
-                Row row = sheet.getRow(rowNum);
-                if (row == null) {
-                    continue;
             }
+            if (rowIndex < firstDataRow) {
+                return;
+            }
+            consumeDataRow(rowIndex, rowMap);
+        }
+
+        private void consumeDataRow(int rowNum, Map<Integer, String> rowMap) {
             for (CountryColumnPair pair : pairs) {
                 if (rowNum < pair.dataStartRowIndex()) {
                     continue;
                 }
                 String country = pair.country();
-                    String asin = normalizeCellText(formatter.formatCellValue(row.getCell(pair.asinColumnIndex())));
-                    String status = normalizeCellText(formatter.formatCellValue(row.getCell(pair.statusColumnIndex())));
+                String asin = normalizeCellText(cellText(rowMap, pair.asinColumnIndex()));
+                String status = normalizeCellText(cellText(rowMap, pair.statusColumnIndex()));
                 if (asin.isBlank() && status.isBlank()) {
                     continue;
                 }
@@ -571,6 +634,11 @@ public class DeleteBrandRunService {
             }
         }
 
+        private ParsedDeleteBrandFile finish() {
+            if (!headerRowSeen) {
+                // 空表 / 表头行缺失：与旧实现 titleRow == null || headerRow == null 一致
+                throw new BusinessException("Excel 表头不完整");
+            }
             List<DeleteBrandCountryGroupVo> countries = new ArrayList<>();
             List<DeleteBrandPreviewRowVo> previewRows = new ArrayList<>();
             int totalRows = 0;
@@ -598,10 +666,6 @@ public class DeleteBrandRunService {
             }
 
             return new ParsedDeleteBrandFile(totalRows, countries, previewRows);
-        } catch (BusinessException ex) {
-            throw ex;
-        } catch (Exception ex) {
-            throw new BusinessException("读取删除品牌 Excel 失败");
         }
     }
 
@@ -612,13 +676,13 @@ public class DeleteBrandRunService {
         return value.replace(" ", "").trim().toUpperCase(Locale.ROOT);
     }
 
-    private List<CountryColumnPair> resolveCountryPairs(Row titleRow, Row headerRow, DataFormatter formatter) {
+    private List<CountryColumnPair> resolveCountryPairs(Map<Integer, String> titleRowMap, Map<Integer, String> headerRowMap) {
         List<CountryColumnPair> pairs = new ArrayList<>();
-        int lastCellNum = Math.max(titleRow.getLastCellNum(), headerRow.getLastCellNum());
+        int lastCellNum = Math.max(lastColumnCount(titleRowMap), lastColumnCount(headerRowMap));
         for (int i = 0; i < lastCellNum; i++) {
-            String title = normalizeCellText(formatter.formatCellValue(titleRow.getCell(i)));
-            String firstHeader = normalizeCellText(formatter.formatCellValue(headerRow.getCell(i)));
-            String secondHeader = normalizeCellText(formatter.formatCellValue(headerRow.getCell(i + 1)));
+            String title = normalizeCellText(cellText(titleRowMap, i));
+            String firstHeader = normalizeCellText(cellText(headerRowMap, i));
+            String secondHeader = normalizeCellText(cellText(headerRowMap, i + 1));
             if (title.isBlank()) {
                 continue;
             }
@@ -633,6 +697,27 @@ public class DeleteBrandRunService {
         return pairs;
     }
 
+    /** 流式解析下按列索引取行内文本（与旧 formatter.formatCellValue(row.getCell(index)) 口径一致）。 */
+    private String cellText(Map<Integer, String> rowMap, Integer columnIndex) {
+        if (rowMap == null || columnIndex == null || columnIndex < 0) {
+            return "";
+        }
+        return rowMap.get(columnIndex);
+    }
+
+    private int lastColumnCount(Map<Integer, String> rowMap) {
+        if (rowMap == null || rowMap.isEmpty()) {
+            return 0;
+        }
+        int lastColumnCount = 0;
+        for (Integer columnIndex : rowMap.keySet()) {
+            if (columnIndex != null && columnIndex >= lastColumnCount) {
+                lastColumnCount = columnIndex + 1;
+            }
+        }
+        return lastColumnCount;
+    }
+
     private String resolveShopName(DeleteBrandSourceFileDto sourceFile) {
         String name = sourceFile.getOriginalFilename();
         if (name == null || name.isBlank()) {

backend-java/src/main/java/com/nanri/aiimage/modules/deletebrand/service/DeleteBrandStaleTaskService.java

@@ -38,6 +38,7 @@ import java.time.Duration;
 import java.time.Instant;
 import java.time.LocalDateTime;
 import java.util.ArrayList;
+import java.util.LinkedHashMap;
 import java.util.List;
 import java.util.Map;
 import java.util.stream.Collectors;
@@ -103,13 +104,9 @@ public class DeleteBrandStaleTaskService {
             ShopMatchStaleCheckStats patrolDeleteStats = failStalePatrolDeleteTasks();
             ShopMatchStaleCheckStats queryAsinStats = failStaleQueryAsinTasks();
             ShopMatchStaleCheckStats withdrawStats = failStaleWithdrawTasks();
-            runModuleStaleCheck("brand", brandTaskService::failStaleRunningTasks);
-            runModuleStaleCheck("appearance-patent", appearancePatentTaskService::finalizeStaleTasks);
-            runModuleStaleCheck("similar-asin", similarAsinTaskService::finalizeStaleTasks);
-            // 商品管理采集并入本巡检线（2026-09）：此前它自带 @Scheduled 且按 owner 过滤，
-            // 双实例下 owner 宕机即无人判死（P1-8）；并入后由本方法的 job 锁保证单实例扫描，
-            // 它自己改为全局判死 + 任务锁 + status CAS
-            runModuleStaleCheck("shop-data-crawl", shopDataCrawlTaskService::finalizeStaleTasks);
+            for (Map.Entry<String, Runnable> delegated : delegatedStaleChecks().entrySet()) {
+                runModuleStaleCheck(delegated.getKey(), delegated.getValue());
+            }
             // 周期每 2 分钟一轮，各模块 summary 合并为单行，避免定期刷屏
             // 注意：每段占位符数量必须与实参一致——此前每段 5 个占位符只传 4 个参数，
             // 导致 withdraw 之后的取值整体错位、末尾 elapsedMs/thread 打成字面量
@@ -125,6 +122,23 @@ public class DeleteBrandStaleTaskService {
         }
     }
 
+    /**
+     * 委派式陈旧判死：moduleType → 处理动作。
+     * {@code TaskModuleRegistry} 中 delegatedStaleCheck=true 的模块都必须在这里登记，
+     * 由 TaskModuleCoverageTest 守住一致性（新增模块漏登记会红测试）。
+     */
+    Map<String, Runnable> delegatedStaleChecks() {
+        Map<String, Runnable> handlers = new LinkedHashMap<>();
+        handlers.put("BRAND", brandTaskService::failStaleRunningTasks);
+        handlers.put("APPEARANCE_PATENT", appearancePatentTaskService::finalizeStaleTasks);
+        handlers.put("SIMILAR_ASIN", similarAsinTaskService::finalizeStaleTasks);
+        // 商品管理采集并入本巡检线（2026-09）：此前它自带 @Scheduled 且按 owner 过滤，
+        // 双实例下 owner 宕机即无人判死（P1-8）；并入后由本方法的 job 锁保证单实例扫描，
+        // 它自己改为全局判死 + 任务锁 + status CAS
+        handlers.put("SHOP_DATA_CRAWL", shopDataCrawlTaskService::finalizeStaleTasks);
+        return handlers;
+    }
+
     private void runModuleStaleCheck(String moduleName, Runnable action) {
         long startedAt = System.currentTimeMillis();
         try {

backend-java/src/main/java/com/nanri/aiimage/modules/deletebrand/service/DeleteBrandTaskHeartbeatSpi.java

@@ -0,0 +1,57 @@
+package com.nanri.aiimage.modules.deletebrand.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+import java.util.LinkedHashMap;
+import java.util.Map;
+
+/**
+ * 删除 ASIN 的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>本模块的心跳不写独立的"心跳键"，而是整段进度写进任务缓存（saveProgress，forceWrite=true），
+ * 并携带客户端上报的 phase/current/total。
+ */
+@Service
+@RequiredArgsConstructor
+public class DeleteBrandTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final DeleteBrandTaskCacheService cacheService;
+
+    @Override
+    public String moduleType() {
+        return "DELETE_BRAND";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.saveProgress(taskId, buildHeartbeatProgress(request), true);
+    }
+
+    @Override
+    public void saveTaskCache(FileTaskEntity task) {
+        cacheService.saveTaskCache(task);
+    }
+
+    private Map<String, String> buildHeartbeatProgress(TaskHeartbeatRequest request) {
+        String now = String.valueOf(System.currentTimeMillis());
+        Map<String, String> values = new LinkedHashMap<>();
+        values.put("last_heartbeat_at", now);
+        values.put("updated_at", now);
+        if (request != null) {
+            putIfPresent(values, "phase", request.getPhase());
+            putIfPresent(values, "current", request.getCurrent());
+            putIfPresent(values, "total", request.getTotal());
+        }
+        return values;
+    }
+
+    private void putIfPresent(Map<String, String> values, String key, Object value) {
+        if (value != null) {
+            values.put(key, String.valueOf(value));
+        }
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/digitalhuman/controller/DigitalHumanVersionController.java

@@ -1,12 +1,14 @@
 package com.nanri.aiimage.modules.digitalhuman.controller;
 
 import com.nanri.aiimage.common.api.ApiResponse;
+import com.nanri.aiimage.modules.admin.support.AdminAuthSupport;
 import com.nanri.aiimage.modules.digitalhuman.model.dto.UploadVersionRequest;
 import com.nanri.aiimage.modules.digitalhuman.model.vo.DigitalHumanVersionVo;
 import com.nanri.aiimage.modules.digitalhuman.service.DigitalHumanVersionService;
 import io.swagger.v3.oas.annotations.Operation;
 import io.swagger.v3.oas.annotations.Parameter;
 import io.swagger.v3.oas.annotations.tags.Tag;
+import jakarta.servlet.http.HttpServletRequest;
 import jakarta.validation.Valid;
 import lombok.RequiredArgsConstructor;
 import org.springframework.http.MediaType;
@@ -27,10 +29,22 @@ import java.util.List;
 public class DigitalHumanVersionController {
 
     private final DigitalHumanVersionService versionService;
+    private final AdminAuthSupport adminAuthSupport;
+
+    /**
+     * 写操作（上传/发布/设为最新/删除）必须管理员身份：2026-09 全维度审查 A4——
+     * 此前整套接口匿名可达，任何人可发布/删除客户端版本。
+     * 读接口（列表/latest/详情/download-url）保持匿名：桌面客户端自身要查版本与下载地址。
+     */
+    private void requireAdmin(HttpServletRequest request) {
+        adminAuthSupport.requireAdminOrInternal(request);
+    }
 
     @PostMapping(value = "/upload", consumes = MediaType.MULTIPART_FORM_DATA_VALUE)
     @Operation(summary = "上传新版本", description = "上传数字人程序新版本到 OSS，并记录版本信息")
-    public ApiResponse<DigitalHumanVersionVo> uploadVersion(@Valid @ModelAttribute UploadVersionRequest request) {
+    public ApiResponse<DigitalHumanVersionVo> uploadVersion(@Valid @ModelAttribute UploadVersionRequest request,
+                                                           HttpServletRequest httpRequest) {
+        requireAdmin(httpRequest);
         DigitalHumanVersionVo vo = versionService.uploadVersion(
                 request.getVersion(),
                 request.getFile(),
@@ -63,21 +77,27 @@ public class DigitalHumanVersionController {
     @PostMapping("/{version}/release")
     @Operation(summary = "发布版本", description = "将草稿状态的版本发布，状态变更为 RELEASED")
     public ApiResponse<DigitalHumanVersionVo> releaseVersion(
-            @Parameter(description = "版本号", example = "1.0.0") @PathVariable String version) {
+            @Parameter(description = "版本号", example = "1.0.0") @PathVariable String version,
+            HttpServletRequest request) {
+        requireAdmin(request);
         return ApiResponse.success(versionService.releaseVersion(version));
     }
 
     @PostMapping("/{version}/set-latest")
     @Operation(summary = "设为最新版本", description = "将指定已发布版本设置为最新版本，清除其他版本的最新标记")
     public ApiResponse<DigitalHumanVersionVo> setLatest(
-            @Parameter(description = "版本号", example = "1.0.0") @PathVariable String version) {
+            @Parameter(description = "版本号", example = "1.0.0") @PathVariable String version,
+            HttpServletRequest request) {
+        requireAdmin(request);
         return ApiResponse.success(versionService.setLatest(version));
     }
 
     @DeleteMapping("/{version}")
     @Operation(summary = "删除版本", description = "删除指定版本，同时删除 OSS 上的文件。最新版本不允许删除")
     public ApiResponse<Void> deleteVersion(
-            @Parameter(description = "版本号", example = "1.0.0") @PathVariable String version) {
+            @Parameter(description = "版本号", example = "1.0.0") @PathVariable String version,
+            HttpServletRequest request) {
+        requireAdmin(request);
         versionService.deleteVersion(version);
         return ApiResponse.success(null);
     }

backend-java/src/main/java/com/nanri/aiimage/modules/file/service/LocalFileStorageService.java

@@ -3,34 +3,46 @@ package com.nanri.aiimage.modules.file.service;
 import cn.hutool.core.io.FileUtil;
 import cn.hutool.core.util.IdUtil;
 import com.nanri.aiimage.common.exception.BusinessException;
+import com.nanri.aiimage.common.util.ExcelStreamReader;
 import com.nanri.aiimage.config.StorageProperties;
 import com.nanri.aiimage.modules.file.model.vo.ExcelInfoVo;
 import com.nanri.aiimage.modules.file.model.vo.UploadFileVo;
 import lombok.RequiredArgsConstructor;
-import org.apache.poi.ss.usermodel.Cell;
-import org.apache.poi.ss.usermodel.DataFormatter;
-import org.apache.poi.ss.usermodel.Row;
-import org.apache.poi.ss.usermodel.Sheet;
-import org.apache.poi.ss.usermodel.Workbook;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.beans.factory.annotation.Value;
 import org.springframework.stereotype.Service;
 import org.springframework.web.multipart.MultipartFile;
 
 import java.io.File;
-import java.io.FileInputStream;
 import java.io.IOException;
 import java.util.ArrayList;
-import java.util.HashMap;
 import java.util.LinkedHashMap;
+import java.util.LinkedHashSet;
 import java.util.List;
 import java.util.Map;
+import java.util.Set;
 
 @Service
 @RequiredArgsConstructor
+@Slf4j
 public class LocalFileStorageService {
 
     /** fileKey → 文件名 索引容量：超限淘汰最旧条目，避免索引自身无界增长。 */
     static final int SOURCE_FILE_INDEX_CAPACITY = 1024;
 
+    /**
+     * 禁止上传的可执行/脚本类扩展名（2026-09 全维度审查 A3）。
+     * 上传接口是桌面端文件中转口（临时目录落盘，2GB/次），此前对类型无任何限制；
+     * 这里挡掉"落盘后一旦被其它组件误执行/误引用即可造成危害"的扩展名。
+     * 业务文件（xlsx/xls/csv/txt/json/zip/图片/视频）不受影响；
+     * 如需调整由运维在 application.yml 覆盖 aiimage.files.upload-blocked-extensions。
+     */
+    private static final String DEFAULT_BLOCKED_EXTENSIONS =
+            "exe,bat,cmd,com,scr,msi,dll,sh,ps1,vbs,vbe,wsf,hta,jsp,jspx,php,jar,so,dylib";
+
+    @Value("${aiimage.files.upload-blocked-extensions:" + DEFAULT_BLOCKED_EXTENSIONS + "}")
+    private String blockedExtensions = DEFAULT_BLOCKED_EXTENSIONS;
+
     private final StorageProperties storageProperties;
 
     /** 临时目录标准路径：供调用方做"文件必须落在上传临时目录内"的穿越校验。 */
@@ -51,6 +63,7 @@ public class LocalFileStorageService {
         FileUtil.mkdir(tempDir);
         String fileKey = IdUtil.fastSimpleUUID();
         String extName = FileUtil.extName(file.getOriginalFilename());
+        rejectBlockedExtension(file.getOriginalFilename(), extName);
         String filename = fileKey + (extName.isEmpty() ? "" : "." + extName);
         File target = FileUtil.file(tempDir, filename);
         file.transferTo(target);
@@ -65,49 +78,125 @@ public class LocalFileStorageService {
         return vo;
     }
 
+    /** 扩展名黑名单校验：命中即拒绝落盘（大小写不敏感）。 */
+    private void rejectBlockedExtension(String originalFilename, String extName) {
+        if (extName == null || extName.isBlank()) {
+            return;
+        }
+        String normalized = extName.trim().toLowerCase(java.util.Locale.ROOT);
+        for (String blocked : blockedExtensions.split(",")) {
+            if (!blocked.isBlank() && blocked.trim().toLowerCase(java.util.Locale.ROOT).equals(normalized)) {
+                log.warn("[file-upload] 拒绝上传被禁扩展名 originalFilename={} ext={}", originalFilename, normalized);
+                throw new BusinessException("不支持上传该类型文件：." + normalized);
+            }
+        }
+    }
+
     public ExcelInfoVo getExcelInfo(String fileKey) throws IOException {
         File inputFile = findLocalSourceFile(fileKey);
         if (inputFile == null || !inputFile.exists()) {
             throw new BusinessException("源文件不存在");
         }
 
-        DataFormatter formatter = new DataFormatter();
-        try (FileInputStream fis = new FileInputStream(inputFile); Workbook workbook = org.apache.poi.ss.usermodel.WorkbookFactory.create(fis)) {
-            Sheet sheet = workbook.getSheetAt(0);
-            Row headerRow = sheet.getRow(0);
-            if (headerRow == null) {
-                throw new BusinessException("Excel 表头为空");
+        // 2026-09 全维度审查 C5：改流式解析（ExcelStreamReader → EasyExcel SAX）。
+        // 此前 WorkbookFactory.create 把整表读成 DOM，只为取表头与行数却让整表驻留堆内存。
+        // 返回值结构与字段含义不变：headers 同旧口径（去重、空表头跳过、命中「缩略图地址8」截断），
+        // totalRows 仍是"数据行数"（旧实现取 lastRowNum，等于物理末行行号-1）。
+        try {
+            ExcelInfoStreamingReader reader = new ExcelInfoStreamingReader();
+            ExcelStreamReader.readFirstSheet(inputFile, reader);
+            reader.assertHeaderSeen();
+            ExcelInfoVo vo = new ExcelInfoVo();
+            vo.setHeaders(reader.headers());
+            vo.setTotalRows(reader.totalRows());
+            return vo;
+        } catch (BusinessException e) {
+            throw e;
+        } catch (Exception e) {
+            log.warn("[file-excel-info] 读取 Excel 信息失败 fileKey={} err={}", fileKey, e.getMessage(), e);
+            throw new BusinessException("读取 Excel 信息失败");
+        }
     }
 
-            Map<String, Integer> headerMap = new HashMap<>();
-            List<String> headers = new ArrayList<>();
-            for (int i = 0; i < headerRow.getLastCellNum(); i++) {
-                Cell cell = headerRow.getCell(i);
-                String value = normalizeCellText(cell == null ? null : formatter.formatCellValue(cell));
+    /**
+     * 表头与行数的流式读取：内存不随文件行数增长。
+     * 表头行整行为空（EasyExcel 不上报表头回调）时视为「Excel 表头为空」——旧实现仅在 getRow(0)==null
+     * 时报同一错误，此处口径略严（不存在的表头与空表头同样处理）。
+     */
+    private final class ExcelInfoStreamingReader implements ExcelStreamReader.SheetRowHandler {
+
+        private final List<String> headers = new ArrayList<>();
+        private final Set<String> seenHeaders = new LinkedHashSet<>();
+        private Integer approximateTotalRows;
+        private boolean headerSeen;
+        private boolean headerTruncated;
+        private int countedRows;
+
+        @Override
+        public void onHeader(String sheetName, Integer sheetNo, Map<Integer, String> headerMap) {
+            headerSeen = true;
+            int lastColumnCount = lastHeaderColumnCount(headerMap);
+            for (int i = 0; i < lastColumnCount && !headerTruncated; i++) {
+                String value = normalizeCellText(headerMap.get(i));
                 if (value.isBlank()) {
                     continue;
                 }
                 value = value.split("idASIN国家状态价格变体数量", 2)[0].trim().isEmpty()
                         ? value
                         : value.split("idASIN国家状态价格变体数量", 2)[0].trim();
-                if (headerMap.containsKey(value)) {
+                if (!seenHeaders.add(value)) {
                     continue;
                 }
-                headerMap.put(value, i);
                 headers.add(value);
                 if ("缩略图地址8".equals(value)) {
-                    break;
+                    headerTruncated = true;
+                }
             }
         }
 
-            ExcelInfoVo vo = new ExcelInfoVo();
-            vo.setHeaders(headers);
-            vo.setTotalRows(Math.max(sheet.getLastRowNum(), 0));
-            return vo;
-        } catch (BusinessException e) {
-            throw e;
-        } catch (Exception e) {
-            throw new BusinessException("读取 Excel 信息失败");
+        @Override
+        public void onSheetTotal(String sheetName, Integer sheetNo, Integer approximateTotalRows) {
+            this.approximateTotalRows = approximateTotalRows;
+        }
+
+        @Override
+        public void onRow(String sheetName, Integer sheetNo, int rowIndex,
+                          Map<Integer, String> headerMap, Map<Integer, String> rowMap) {
+            countedRows++;
+        }
+
+        private void assertHeaderSeen() {
+            if (!headerSeen) {
+                throw new BusinessException("Excel 表头为空");
+            }
+        }
+
+        private List<String> headers() {
+            return headers;
+        }
+
+        /**
+         * onSheetTotal 的近似值 = 含表头的物理末行行号（POI lastRowNum+1 口径），换算回旧的
+         * "数据行数"口径；近似值缺失（dimension 标签缺失）或偏小时用遍历计数兜底。
+         */
+        private int totalRows() {
+            int approximateDataRows = approximateTotalRows == null || approximateTotalRows <= 0
+                    ? 0
+                    : approximateTotalRows - 1;
+            return Math.max(approximateDataRows, countedRows);
+        }
+
+        private int lastHeaderColumnCount(Map<Integer, String> headerMap) {
+            if (headerMap == null || headerMap.isEmpty()) {
+                return 0;
+            }
+            int lastColumnCount = 0;
+            for (Integer columnIndex : headerMap.keySet()) {
+                if (columnIndex != null && columnIndex >= lastColumnCount) {
+                    lastColumnCount = columnIndex + 1;
+                }
+            }
+            return lastColumnCount;
         }
     }
 

backend-java/src/main/java/com/nanri/aiimage/modules/file/service/object/RustfsDeleteRetryService.java

@@ -1,11 +1,15 @@
 package com.nanri.aiimage.modules.file.service.object;
 
+import com.nanri.aiimage.common.util.PendingDeleteJournal;
 import com.nanri.aiimage.config.TransientStorageProperties;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
+import org.springframework.boot.context.event.ApplicationReadyEvent;
+import org.springframework.context.event.EventListener;
 import org.springframework.scheduling.annotation.Scheduled;
 import org.springframework.stereotype.Service;
 
+import java.nio.file.Path;
 import java.time.LocalDateTime;
 import java.util.ArrayList;
 import java.util.HashSet;
@@ -26,6 +30,48 @@ public class RustfsDeleteRetryService {
     private final ConcurrentLinkedQueue<RetryItem> queue = new ConcurrentLinkedQueue<>();
     private final ConcurrentHashMap<String, RetryItem> pending = new ConcurrentHashMap<>();
     private final Object admissionLock = new Object();
+    /**
+     * 待删对象的本地落盘日志：内存队列重启即丢，对应对象会一直残留在桶里直到生命周期过期。
+     * 入队时追加、每轮重试后按 pending 收敛重写、启动时回放。
+     */
+    private volatile PendingDeleteJournal journal;
+
+    /**
+     * 启动回放：把上次进程残留的待删对象重新入队（容量准入、去重语义与在线入队一致），
+     * 随后由 {@link #retryPendingDeletes()} 的周期任务继续重试。
+     */
+    @EventListener(ApplicationReadyEvent.class)
+    public void replayJournalOnStartup() {
+        if (!properties.isDeleteRetryEnabled()) {
+            return;
+        }
+        List<String> objectKeys = journal().readAll();
+        if (objectKeys.isEmpty()) {
+            return;
+        }
+        int replayed = 0;
+        for (String objectKey : objectKeys) {
+            enqueue(objectKey, null);
+            replayed++;
+        }
+        log.info("[rustfs] 启动回放待删对象 journalSize={} replayed={} pending={}",
+                objectKeys.size(), replayed, pending.size());
+    }
+
+    private PendingDeleteJournal journal() {
+        PendingDeleteJournal local = journal;
+        if (local != null) {
+            return local;
+        }
+        synchronized (admissionLock) {
+            if (journal == null) {
+                Path path = Path.of(System.getProperty("java.io.tmpdir"),
+                        "aiimage-delete-journal", "rustfs-delete-retry.log");
+                journal = new PendingDeleteJournal(path, Math.max(1, properties.getDeleteRetryQueueCapacity()));
+            }
+            return journal;
+        }
+    }
 
     public void enqueue(String objectKey, Throwable cause) {
         if (!properties.isDeleteRetryEnabled() || objectKey == null || objectKey.isBlank()) {
@@ -97,6 +143,8 @@ public class RustfsDeleteRetryService {
         }
         retryNextInvocation.forEach(this::requeue);
         if (processed > 0) {
+            // 收敛落盘：成功删除的对象从日志移除，仍失败/待重试的保留
+            journal().rewrite(pending.keySet());
             log.info("[rustfs] delete retry batch completed processed={} success={} failed={} pending={}",
                     processed, success, failed, pending.size());
         }
@@ -124,6 +172,8 @@ public class RustfsDeleteRetryService {
             }
             item = new RetryItem(objectKey);
             pending.put(objectKey, item);
+            // 只在"新准入"时记账：重写以 pending 为准，重复失败无需重复追加
+            journal().record(objectKey);
             return item;
         }
     }

backend-java/src/main/java/com/nanri/aiimage/modules/imagevideo/controller/ImageVideoController.java

@@ -1,6 +1,9 @@
 package com.nanri.aiimage.modules.imagevideo.controller;
 
 import com.nanri.aiimage.common.api.ApiResponse;
+import com.nanri.aiimage.common.exception.BusinessException;
+import com.nanri.aiimage.modules.permission.model.entity.AdminUserEntity;
+import com.nanri.aiimage.modules.admin.support.AdminAuthSupport;
 import com.nanri.aiimage.modules.imagevideo.model.dto.DouyinCopyRequest;
 import com.nanri.aiimage.modules.imagevideo.model.dto.ImageVideoSecretSaveRequest;
 import com.nanri.aiimage.modules.imagevideo.model.dto.ImageVideoVoiceCloneRequest;
@@ -17,6 +20,7 @@ import com.nanri.aiimage.modules.imagevideo.service.ImageVideoAsyncTaskService;
 import com.nanri.aiimage.modules.imagevideo.service.ImageVideoSecretService;
 import io.swagger.v3.oas.annotations.Operation;
 import io.swagger.v3.oas.annotations.tags.Tag;
+import jakarta.servlet.http.HttpServletRequest;
 import jakarta.validation.Valid;
 import lombok.RequiredArgsConstructor;
 import org.springframework.web.bind.annotation.GetMapping;
@@ -36,21 +40,42 @@ import org.springframework.web.multipart.MultipartFile;
 public class ImageVideoController {
 
     private final ImageVideoCozeService imageVideoCozeService;
+    private final AdminAuthSupport adminAuthSupport;
     private final ImageVideoSecretService imageVideoSecretService;
     private final ImageVideoAsyncTaskService imageVideoAsyncTaskService;
 
     @GetMapping("/secrets")
     @Operation(summary = "查询视频密钥配置状态")
-    public ApiResponse<ImageVideoSecretStatusVo> secretStatus(@RequestParam("user_id") Long userId) {
-        return ApiResponse.success(imageVideoSecretService.status(userId));
+    public ApiResponse<ImageVideoSecretStatusVo> secretStatus(@RequestParam("user_id") Long userId,
+                                                             HttpServletRequest request) {
+        return ApiResponse.success(imageVideoSecretService.status(resolveSelfUserId(request, userId)));
     }
 
     @PutMapping("/secrets")
     @Operation(summary = "保存视频密钥配置")
-    public ApiResponse<ImageVideoSecretStatusVo> saveSecrets(@Valid @RequestBody ImageVideoSecretSaveRequest request) {
+    public ApiResponse<ImageVideoSecretStatusVo> saveSecrets(@Valid @RequestBody ImageVideoSecretSaveRequest request,
+                                                             HttpServletRequest httpRequest) {
+        resolveSelfUserId(httpRequest, request.getUserId());
         return ApiResponse.success(imageVideoSecretService.save(request));
     }
 
+    /**
+     * 2026-09 全维度审查 A7：视频密钥接口此前匿名可达，且 userId 取自请求体/查询串
+     * ——任何人可读写他人密钥配置。现要求登录身份（或可信内部代理），
+     * 且只允许操作自己（超管例外，保留后台代查能力）。
+     */
+    private Long resolveSelfUserId(HttpServletRequest request, Long requestedUserId) {
+        AdminUserEntity me = adminAuthSupport.requireUserOrInternal(request);
+        if (requestedUserId == null || requestedUserId <= 0) {
+            throw new BusinessException("用户 id 不合法");
+        }
+        boolean superAdmin = "super_admin".equals(adminAuthSupport.currentRole(me));
+        if (!superAdmin && !requestedUserId.equals(me.getId())) {
+            throw new BusinessException(403, "只能操作自己的视频密钥配置");
+        }
+        return requestedUserId;
+    }
+
     @PostMapping("/douyin-copy")
     @Operation(summary = "抖音文案识别和仿写")
     public ApiResponse<ImageVideoAsyncTaskVo> douyinCopy(@Valid @RequestBody DouyinCopyRequest request) {

backend-java/src/main/java/com/nanri/aiimage/modules/notification/service/NotificationScanScheduler.java

@@ -3,6 +3,7 @@ package com.nanri.aiimage.modules.notification.service;
 import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.nanri.aiimage.config.HttpClientPool;
 import com.nanri.aiimage.config.NotificationProperties;
+import com.nanri.aiimage.common.module.TaskModuleRegistry;
 import com.nanri.aiimage.common.service.DistributedJobLockService;
 import com.nanri.aiimage.modules.brand.mapper.BrandCrawlTaskMapper;
 import com.nanri.aiimage.modules.brand.model.entity.BrandCrawlTaskEntity;
@@ -54,24 +55,11 @@ public class NotificationScanScheduler {
 
     private static final DateTimeFormatter HOUR_FORMAT = DateTimeFormatter.ofPattern("yyyyMMddHH");
 
-    /** 任务模块类型 → 中文名（与 TaskHeartbeatService 的模块常量对齐）。 */
-    private static final Map<String, String> MODULE_LABELS = Map.ofEntries(
-            Map.entry("APPEARANCE_PATENT", "外观专利检测"),
-            Map.entry("SIMILAR_ASIN", "货源查询"),
-            Map.entry("PATROL_DELETE", "巡店删除"),
-            Map.entry("PRICE_TRACK", "跟价"),
-            Map.entry("PRODUCT_RISK_RESOLVE", "商品风险解决"),
-            Map.entry("QUERY_ASIN", "查询ASIN"),
-            Map.entry("WITHDRAW", "取款"),
-            Map.entry("SHOP_DATA_CRAWL", "店铺数据抓取"),
-            Map.entry("SHOP_MATCH", "定时匹配"),
-            Map.entry("BRAND", "品牌检测"),
-            Map.entry("COLLECT_DATA", "采集数据"),
-            Map.entry("DELETE_BRAND", "删除ASIN"),
-            Map.entry("SPLIT", "数据拆分"),
-            Map.entry("CONVERT", "格式转换"),
-            Map.entry("PUBLISH", "上架"),
-            Map.entry("DEDUPE", "数据去重"));
+    /**
+     * 任务模块类型 → 中文名：直接取模块注册表（G6），新增模块只需在
+     * {@code TaskModuleRegistry} 追加一行，不再需要同步这份表。
+     */
+    private static final Map<String, String> MODULE_LABELS = TaskModuleRegistry.labels();
 
     private final FileTaskMapper fileTaskMapper;
     private final BrandCrawlTaskMapper brandCrawlTaskMapper;

backend-java/src/main/java/com/nanri/aiimage/modules/patroldelete/controller/PatrolDeleteController.java

@@ -13,9 +13,9 @@ import com.nanri.aiimage.modules.patroldelete.service.PatrolDeleteResolveService
 import com.nanri.aiimage.modules.patroldelete.service.PatrolDeleteTaskService;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCandidateAddRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
 import io.swagger.v3.oas.annotations.Operation;
 import io.swagger.v3.oas.annotations.Parameter;
 import io.swagger.v3.oas.annotations.enums.ParameterIn;

backend-java/src/main/java/com/nanri/aiimage/modules/patroldelete/service/PatrolDeleteResolveService.java

@@ -10,9 +10,9 @@ import com.nanri.aiimage.modules.patroldelete.model.entity.PatrolDeleteShopCandi
 import com.nanri.aiimage.modules.patroldelete.model.vo.PatrolDeleteConditionVo;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCandidateAddRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskShopQueueItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskShopQueueItemVo;
 import com.nanri.aiimage.modules.ziniao.model.vo.ZiniaoShopMatchResultVo;
 import com.nanri.aiimage.modules.ziniao.service.ZiniaoShopIndexService;
 import com.nanri.aiimage.modules.ziniao.service.ZiniaoShopSwitchService;

backend-java/src/main/java/com/nanri/aiimage/modules/patroldelete/service/PatrolDeleteTaskHeartbeatSpi.java

@@ -0,0 +1,35 @@
+package com.nanri.aiimage.modules.patroldelete.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 巡店删除 的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>心跳逻辑（缓存刷新 / 任务缓存回写）从 task 模块收回本模块，task 侧只依赖 SPI 接口，
+ * 消除 task → 业务模块的编译期依赖。
+ */
+@Service
+@RequiredArgsConstructor
+public class PatrolDeleteTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final PatrolDeleteTaskCacheService cacheService;
+
+    @Override
+    public String moduleType() {
+        return "PATROL_DELETE";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.touchTaskHeartbeat(taskId);
+    }
+
+    @Override
+    public void saveTaskCache(FileTaskEntity task) {
+        cacheService.saveTaskCache(task);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/patroldelete/service/PatrolDeleteTaskService.java

@@ -21,7 +21,7 @@ import com.nanri.aiimage.modules.patroldelete.model.vo.PatrolDeleteTaskBatchVo;
 import com.nanri.aiimage.modules.patroldelete.model.vo.PatrolDeleteCreateTaskVo;
 import com.nanri.aiimage.modules.patroldelete.model.vo.PatrolDeleteHistoryVo;
 import com.nanri.aiimage.modules.patroldelete.model.vo.PatrolDeleteResultItemVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
 import com.nanri.aiimage.modules.task.mapper.FileResultMapper;
 import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
 import com.nanri.aiimage.modules.task.model.entity.FileResultEntity;

backend-java/src/main/java/com/nanri/aiimage/modules/pricetrack/service/PriceTrackTaskHeartbeatSpi.java

@@ -0,0 +1,35 @@
+package com.nanri.aiimage.modules.pricetrack.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 跟价 的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>心跳逻辑（缓存刷新 / 任务缓存回写）从 task 模块收回本模块，task 侧只依赖 SPI 接口，
+ * 消除 task → 业务模块的编译期依赖。
+ */
+@Service
+@RequiredArgsConstructor
+public class PriceTrackTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final PriceTrackTaskCacheService cacheService;
+
+    @Override
+    public String moduleType() {
+        return "PRICE_TRACK";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.touchTaskHeartbeat(taskId);
+    }
+
+    @Override
+    public void saveTaskCache(FileTaskEntity task) {
+        cacheService.saveTaskCache(task);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/productrisk/controller/ProductRiskResolveController.java

@@ -8,14 +8,14 @@ import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCreateTaskRequ
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskSubmitResultRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskTaskBatchRequest;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCountryPreferenceVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCountryPreferenceVo;
 import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCreateTaskVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskHistoryVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskHistoryVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
 import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskPendingDeleteVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskTaskBatchVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskTaskBatchVo;
 import com.nanri.aiimage.modules.productrisk.service.ProductRiskResolveService;
 import com.nanri.aiimage.modules.productrisk.service.ProductRiskTaskService;
 import io.swagger.v3.oas.annotations.Operation;

backend-java/src/main/java/com/nanri/aiimage/modules/productrisk/model/dto/ProductRiskCreateTaskRequest.java

@@ -1,7 +1,7 @@
 package com.nanri.aiimage.modules.productrisk.model.dto;
 
 import com.fasterxml.jackson.annotation.JsonProperty;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskShopQueueItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskShopQueueItemVo;
 import io.swagger.v3.oas.annotations.media.Schema;
 import jakarta.validation.Valid;
 import jakarta.validation.constraints.NotEmpty;

backend-java/src/main/java/com/nanri/aiimage/modules/productrisk/model/vo/ProductRiskCreateTaskVo.java

@@ -5,6 +5,7 @@ import lombok.Data;
 
 import java.util.ArrayList;
 import java.util.List;
+import com.nanri.aiimage.common.model.vo.ProductRiskResultItemVo;
 
 @Data
 @Schema(description = "创建任务响应：含任务 id 与各店初始快照（含 resultId），用于推队列与前端展示")

backend-java/src/main/java/com/nanri/aiimage/modules/productrisk/service/ProductRiskResolveService.java

@@ -12,10 +12,10 @@ import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequ
 import com.nanri.aiimage.modules.productrisk.model.entity.ProductRiskCountryPrefEntity;
 import com.nanri.aiimage.modules.productrisk.model.entity.ProductRiskShopCandidateEntity;
 import com.nanri.aiimage.modules.productrisk.model.enums.ProductRiskCountryCode;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCountryPreferenceVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskShopQueueItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCountryPreferenceVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskShopQueueItemVo;
 import com.nanri.aiimage.modules.ziniao.model.vo.ZiniaoShopMatchResultVo;
 import com.nanri.aiimage.modules.ziniao.service.ZiniaoShopIndexService;
 import com.nanri.aiimage.modules.ziniao.service.ZiniaoShopSwitchService;

backend-java/src/main/java/com/nanri/aiimage/modules/productrisk/service/ProductRiskTaskHeartbeatSpi.java

@@ -0,0 +1,35 @@
+package com.nanri.aiimage.modules.productrisk.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 商品风险解决 的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>心跳逻辑（缓存刷新 / 任务缓存回写）从 task 模块收回本模块，task 侧只依赖 SPI 接口，
+ * 消除 task → 业务模块的编译期依赖。
+ */
+@Service
+@RequiredArgsConstructor
+public class ProductRiskTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final ProductRiskTaskCacheService cacheService;
+
+    @Override
+    public String moduleType() {
+        return "PRODUCT_RISK_RESOLVE";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.touchTaskHeartbeat(taskId);
+    }
+
+    @Override
+    public void saveTaskCache(FileTaskEntity task) {
+        cacheService.saveTaskCache(task);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/productrisk/service/ProductRiskTaskService.java

@@ -13,14 +13,14 @@ import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskShopPayloadDto
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskSubmitResultRequest;
 import com.nanri.aiimage.modules.productrisk.model.entity.ProductRiskShopCandidateEntity;
 import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCreateTaskVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskHistoryVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskResultItemVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskTaskBatchVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskTaskDetailVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskHistoryVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskResultItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskTaskBatchVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskTaskDetailVo;
 import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskPendingDeleteVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskTaskItemVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskShopQueueItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskTaskItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskShopQueueItemVo;
 import com.nanri.aiimage.modules.task.mapper.FileResultMapper;
 import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
 import com.nanri.aiimage.modules.task.model.entity.FileResultEntity;

backend-java/src/main/java/com/nanri/aiimage/modules/publish/service/PublishTaskHeartbeatSpi.java

@@ -0,0 +1,35 @@
+package com.nanri.aiimage.modules.publish.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 上架模块的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>心跳前做任务归属校验（非本实例拥有的任务拒绝续命），心跳刷新走模块自己的进度服务；
+ * 本模块没有独立的任务缓存回写（历史行为保持一致）。
+ */
+@Service
+@RequiredArgsConstructor
+public class PublishTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final PublishTaskService publishTaskService;
+
+    @Override
+    public String moduleType() {
+        return PublishTaskService.MODULE_TYPE;
+    }
+
+    @Override
+    public void ensureOwnership(FileTaskEntity task, String operation) {
+        publishTaskService.ensureTaskOwnedByCurrentInstance(task, operation);
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        publishTaskService.touchHeartbeat(taskId, request);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/queryasin/controller/QueryAsinTaskController.java

@@ -3,9 +3,9 @@ package com.nanri.aiimage.modules.queryasin.controller;
 import com.nanri.aiimage.common.api.ApiResponse;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCandidateAddRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
 import com.nanri.aiimage.modules.queryasin.model.dto.QueryAsinCreateTaskRequest;
 import com.nanri.aiimage.modules.queryasin.model.dto.QueryAsinSubmitResultRequest;
 import com.nanri.aiimage.modules.queryasin.model.dto.QueryAsinTaskBatchRequest;

backend-java/src/main/java/com/nanri/aiimage/modules/queryasin/service/QueryAsinResolveService.java

@@ -4,9 +4,9 @@ import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.nanri.aiimage.common.exception.BusinessException;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCandidateAddRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskShopQueueItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskShopQueueItemVo;
 import com.nanri.aiimage.modules.queryasin.mapper.QueryAsinShopCandidateMapper;
 import com.nanri.aiimage.modules.queryasin.model.dto.QueryAsinCountryAsinsDto;
 import com.nanri.aiimage.modules.queryasin.model.entity.QueryAsinShopCandidateEntity;

backend-java/src/main/java/com/nanri/aiimage/modules/queryasin/service/QueryAsinTaskHeartbeatSpi.java

@@ -0,0 +1,35 @@
+package com.nanri.aiimage.modules.queryasin.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 查询ASIN 的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>心跳逻辑（缓存刷新 / 任务缓存回写）从 task 模块收回本模块，task 侧只依赖 SPI 接口，
+ * 消除 task → 业务模块的编译期依赖。
+ */
+@Service
+@RequiredArgsConstructor
+public class QueryAsinTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final QueryAsinTaskCacheService cacheService;
+
+    @Override
+    public String moduleType() {
+        return "QUERY_ASIN";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.touchTaskHeartbeat(taskId);
+    }
+
+    @Override
+    public void saveTaskCache(FileTaskEntity task) {
+        cacheService.saveTaskCache(task);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/queryasin/service/QueryAsinTaskService.java

@@ -21,7 +21,7 @@ import com.nanri.aiimage.modules.queryasin.model.vo.QueryAsinTaskBatchVo;
 import com.nanri.aiimage.modules.queryasin.model.vo.QueryAsinCreateTaskVo;
 import com.nanri.aiimage.modules.queryasin.model.vo.QueryAsinHistoryVo;
 import com.nanri.aiimage.modules.queryasin.model.vo.QueryAsinResultItemVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
 import com.nanri.aiimage.modules.task.mapper.FileResultMapper;
 import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
 import com.nanri.aiimage.modules.task.model.entity.FileResultEntity;

backend-java/src/main/java/com/nanri/aiimage/modules/shopdatacrawl/controller/ShopDataCrawlTaskController.java

@@ -4,10 +4,10 @@ import com.nanri.aiimage.common.api.ApiResponse;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCandidateAddRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCountryPreferenceSaveRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCountryPreferenceVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCountryPreferenceVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
 import com.nanri.aiimage.modules.shopdatacrawl.model.dto.ShopDataCrawlCreateTaskRequest;
 import com.nanri.aiimage.modules.shopdatacrawl.model.dto.ShopDataCrawlSubmitResultRequest;
 import com.nanri.aiimage.modules.shopdatacrawl.model.dto.ShopDataCrawlTaskBatchRequest;

backend-java/src/main/java/com/nanri/aiimage/modules/shopdatacrawl/mapper/ShopDataCrawlItemMapper.java

@@ -19,6 +19,20 @@ public interface ShopDataCrawlItemMapper extends BaseMapper<ShopDataCrawlItemEnt
             """)
     int deleteBatch(@Param("shopName") String shopName, @Param("businessDate") java.time.LocalDate businessDate);
 
+    /**
+     * 按批次日期分批清理历史明细（保留期用）。
+     *
+     * <p>该表是增长最快的表（店 × 日 × 国 × SKU）且此前**没有任何清理策略**（2026-09 全维度审查）。
+     * 撞款扫描只读每店最新批次，历史批次仅用于追溯，过期即可删；分批删避免长事务与大范围行锁，
+     * 走 idx_shop_data_crawl_item_biz_date 范围扫描。
+     */
+    @Delete("""
+            DELETE FROM biz_shop_data_crawl_item
+            WHERE business_date < #{cutoff}
+            LIMIT #{batchSize}
+            """)
+    int deleteOlderThanBatch(@Param("cutoff") java.time.LocalDate cutoff, @Param("batchSize") int batchSize);
+
     /** 删除归属某累计档的全部明细行（管理端真删店铺数据记录时清理，避免留下悬空引用）。 */
     @Delete("""
             DELETE FROM biz_shop_data_crawl_item

backend-java/src/main/java/com/nanri/aiimage/modules/shopdatacrawl/service/ShopDataCrawlItemRetentionService.java

@@ -0,0 +1,76 @@
+package com.nanri.aiimage.modules.shopdatacrawl.service;
+
+import com.nanri.aiimage.common.service.DistributedJobLockService;
+import com.nanri.aiimage.modules.shopdatacrawl.mapper.ShopDataCrawlItemMapper;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.beans.factory.annotation.Value;
+import org.springframework.scheduling.annotation.Scheduled;
+import org.springframework.stereotype.Service;
+
+import java.time.Duration;
+import java.time.LocalDate;
+
+/**
+ * 采集明细表（biz_shop_data_crawl_item）的保留期清理（2026-09 全维度审查）。
+ *
+ * <p>该表按「店 × 日 × 国 × SKU」增长，是全库增长最快的表，此前只有"按店按日替换当天批次"
+ * 与"删累计档时连带删除"两条删除路径，历史批次永不清理。撞款扫描只读每店**最新**批次，
+ * 因此保留期内（默认 30 天）之外的批次可安全删除。
+ *
+ * <p>双节点用 job 锁保证单实例执行；分批删除并限制单轮批次数，避免一次跑太久占住锁。
+ */
+@Service
+@RequiredArgsConstructor
+@Slf4j
+public class ShopDataCrawlItemRetentionService {
+
+    /** 单轮最多删除的批次数（每批 batchSize 行），剩余留给下一轮。 */
+    static final int MAX_BATCHES_PER_RUN = 20;
+
+    private final ShopDataCrawlItemMapper itemMapper;
+    private final DistributedJobLockService distributedJobLockService;
+
+    @Value("${aiimage.shop-data-crawl.item-retention-days:30}")
+    private int retentionDays = 30;
+
+    @Value("${aiimage.shop-data-crawl.item-retention-batch-size:5000}")
+    private int retentionBatchSize = 5000;
+
+    @Scheduled(cron = "${aiimage.shop-data-crawl.item-retention-cron:0 30 3 * * *}")
+    public void purgeExpiredItems() {
+        int days = Math.max(1, retentionDays);
+        int batchSize = Math.max(100, retentionBatchSize);
+        LocalDate cutoff = LocalDate.now().minusDays(days);
+
+        DistributedJobLockService.LockHandle lockHandle =
+                distributedJobLockService.tryLock("shop-data-crawl:item-retention", Duration.ofMinutes(15));
+        if (lockHandle == null) {
+            log.info("[shop-data-crawl] item retention skipped because another instance holds the lock");
+            return;
+        }
+        try (lockHandle) {
+            int totalDeleted = 0;
+            int batches = 0;
+            while (batches < MAX_BATCHES_PER_RUN) {
+                int deleted = itemMapper.deleteOlderThanBatch(cutoff, batchSize);
+                batches++;
+                totalDeleted += deleted;
+                if (deleted < batchSize) {
+                    break;
+                }
+            }
+            if (totalDeleted > 0) {
+                log.info("[shop-data-crawl] item retention deleted={} cutoff={} retentionDays={} batches={}",
+                        totalDeleted, cutoff, days, batches);
+            }
+        } catch (Exception ex) {
+            log.warn("[shop-data-crawl] item retention failed cutoff={} msg={}", cutoff, ex.getMessage(), ex);
+        }
+    }
+
+    /** 供测试读取当前生效的保留天数。 */
+    int effectiveRetentionDays() {
+        return Math.max(1, retentionDays);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/shopdatacrawl/service/ShopDataCrawlResolveService.java

@@ -12,10 +12,10 @@ import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequ
 import com.nanri.aiimage.modules.shopdatacrawl.model.entity.ShopDataCrawlCountryPrefEntity;
 import com.nanri.aiimage.modules.shopdatacrawl.model.entity.ShopDataCrawlShopCandidateEntity;
 import com.nanri.aiimage.modules.productrisk.model.enums.ProductRiskCountryCode;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCountryPreferenceVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskShopQueueItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCountryPreferenceVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskShopQueueItemVo;
 import com.nanri.aiimage.modules.shopkey.service.ShopManageService;
 import com.nanri.aiimage.modules.ziniao.model.vo.ZiniaoShopMatchResultVo;
 import com.nanri.aiimage.modules.ziniao.service.ZiniaoShopIndexService;

backend-java/src/main/java/com/nanri/aiimage/modules/shopdatacrawl/service/ShopDataCrawlTaskHeartbeatSpi.java

@@ -0,0 +1,41 @@
+package com.nanri.aiimage.modules.shopdatacrawl.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 店铺数据采集的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>除常规缓存刷新外，本模块在心跳前要做**任务归属校验**：非本实例拥有的任务拒绝心跳，
+ * 避免另一实例误把在途任务的 updated_at 续命。
+ */
+@Service
+@RequiredArgsConstructor
+public class ShopDataCrawlTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final ShopDataCrawlTaskCacheService cacheService;
+    private final ShopDataCrawlTaskService shopDataCrawlTaskService;
+
+    @Override
+    public String moduleType() {
+        return "SHOP_DATA_CRAWL";
+    }
+
+    @Override
+    public void ensureOwnership(FileTaskEntity task, String operation) {
+        shopDataCrawlTaskService.ensureTaskOwnedByCurrentInstance(task, operation);
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.touchTaskHeartbeat(taskId);
+    }
+
+    @Override
+    public void saveTaskCache(FileTaskEntity task) {
+        cacheService.saveTaskCache(task);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/shopdatacrawl/service/ShopDataCrawlTaskService.java

@@ -27,7 +27,7 @@ import com.nanri.aiimage.modules.shopdatacrawl.model.vo.ShopDataCrawlTaskBatchVo
 import com.nanri.aiimage.modules.shopdatacrawl.model.vo.ShopDataCrawlCreateTaskVo;
 import com.nanri.aiimage.modules.shopdatacrawl.model.vo.ShopDataCrawlHistoryVo;
 import com.nanri.aiimage.modules.shopdatacrawl.model.vo.ShopDataCrawlResultItemVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
 import com.nanri.aiimage.modules.task.mapper.FileResultMapper;
 import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
 import com.nanri.aiimage.modules.task.mapper.TaskChunkMapper;

backend-java/src/main/java/com/nanri/aiimage/modules/shopduplicatecheck/mapper/ShopDuplicateCheckItemMapper.java

@@ -16,7 +16,18 @@ import java.util.List;
 @Mapper
 public interface ShopDuplicateCheckItemMapper {
 
-    /** 每店最新 business_date 批次的全部明细行（撞款统计直接数据源）。 */
+    /**
+     * 全量店铺名（去重，走 idx_shop_data_crawl_item_shop_date 的索引扫描，不读行数据）。
+     * 撞款扫描按店逐批取数用，避免一次性把全部明细拉进堆。
+     */
+    @Select("SELECT DISTINCT shop_name FROM biz_shop_data_crawl_item")
+    List<String> selectDistinctShopNames();
+
+    /**
+     * 单店最新 business_date 批次的明细行。
+     * 内层 MAX(business_date) 走 (shop_name, business_date) 索引（松散索引扫描），
+     * 外层按索引前缀取该批次，只读目标店的行，不再全表 GROUP BY + JOIN。
+     */
     @Select("""
             SELECT i.shop_name AS shopName,
                    i.country AS country,
@@ -25,14 +36,12 @@ public interface ShopDuplicateCheckItemMapper {
                    i.price AS price,
                    i.brand AS brand
             FROM biz_shop_data_crawl_item i
-            JOIN (
-                SELECT shop_name, MAX(business_date) AS max_date
-                FROM biz_shop_data_crawl_item
-                GROUP BY shop_name
-            ) latest
-              ON latest.shop_name = i.shop_name AND latest.max_date = i.business_date
+            WHERE i.shop_name = #{shopName}
+              AND i.business_date = (
+                  SELECT MAX(business_date) FROM biz_shop_data_crawl_item WHERE shop_name = #{shopName}
+              )
             """)
-    List<ShopLatestItemRowDto> selectLatestShopItems();
+    List<ShopLatestItemRowDto> selectLatestItemsByShop(@Param("shopName") String shopName);
 
     /** 幂等判定：该店该日是否已有明细批次。 */
     @Select("""

backend-java/src/main/java/com/nanri/aiimage/modules/shopduplicatecheck/service/ShopDataDuplicateCheckScanService.java

@@ -232,23 +232,39 @@ public class ShopDataDuplicateCheckScanService {
      * 重复检查基数 = 已采集落库的店（超管应见 8 家）。
      */
     private List<ShopParsed> collectShopParsed(DistributedJobLockService.LockHandle lock) {
-        List<ShopLatestItemRowDto> rows = itemMapper.selectLatestShopItems();
-        if (rows.isEmpty()) {
+        // 按店逐批取数（2026-09 全维度审查 C7）：此前一条 full-table GROUP BY + JOIN 把全部明细
+        // 一次性拉进堆，而明细表随采集天数线性增长（每日每店一批）。改为先走索引扫描拿店名，
+        // 再逐店按 (shop_name, business_date) 索引取「该店最新批次」，
+        // 堆内同时只驻留单店行 + 已构建结构。
+        List<String> rawShopNames = itemMapper.selectDistinctShopNames();
+        if (rawShopNames.isEmpty()) {
             log.info("[shop-duplicate-check] 明细表暂无采集数据，本次扫描不出库");
             return List.of();
         }
         // 店铺名去重（分组标签来自店铺管理 biz_shop_manage，与页面口径一致）。
         Set<String> shopNames = new LinkedHashSet<>();
-        for (ShopLatestItemRowDto row : rows) {
-            if (row != null) {
-                shopNames.add(normalizeShopName(row.shopName()));
+        for (String rawShopName : rawShopNames) {
+            if (rawShopName != null && !rawShopName.isBlank()) {
+                shopNames.add(normalizeShopName(rawShopName));
             }
         }
         Map<String, String> groupLabels = new HashMap<>();
         for (ShopGroupLabelDto dto : sourceMapper.selectGroupLabels(new ArrayList<>(shopNames))) {
             groupLabels.put(shopKey(dto.getShopName()), dto.getGroupName() == null ? "" : dto.getGroupName());
         }
-        List<ShopParsed> parsedShops = buildParsedShops(rows, groupLabels);
+        List<ShopParsed> parsedShops = new ArrayList<>();
+        for (String rawShopName : rawShopNames) {
+            if (rawShopName == null || rawShopName.isBlank()) {
+                continue;
+            }
+            List<ShopLatestItemRowDto> rows = itemMapper.selectLatestItemsByShop(rawShopName);
+            if (rows.isEmpty()) {
+                // 该店历史批次存在但最新批次无行：跳过（不影响其它店）
+                log.info("[shop-duplicate-check] 店铺最新批次无明细行，跳过 shopName={}", rawShopName);
+                continue;
+            }
+            parsedShops.addAll(buildParsedShops(rows, groupLabels));
+        }
         log.info("[shop-duplicate-check] 明细数据源扫描：{} 家店铺，{} 行明细", parsedShops.size(), rawRowsCount(parsedShops));
         return parsedShops;
     }

backend-java/src/main/java/com/nanri/aiimage/modules/shopkey/service/SkipPriceAsinService.java

@@ -5,6 +5,7 @@ import com.baomidou.mybatisplus.core.toolkit.support.SFunction;
 import com.baomidou.mybatisplus.extension.plugins.pagination.Page;
 import cn.hutool.core.util.IdUtil;
 import com.nanri.aiimage.common.exception.BusinessException;
+import com.nanri.aiimage.common.util.ExcelStreamReader;
 import com.nanri.aiimage.modules.shopkey.mapper.SkipPriceAsinMapper;
 import com.nanri.aiimage.modules.shopkey.mapper.ShopManageMapper;
 import com.nanri.aiimage.modules.shopkey.model.dto.SkipPriceAsinCreateRequest;
@@ -19,12 +20,8 @@ import com.nanri.aiimage.modules.shopkey.model.vo.SkipPriceAsinPageVo;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
 import org.springframework.beans.factory.annotation.Value;
-import org.apache.poi.ss.usermodel.Cell;
-import org.apache.poi.ss.usermodel.DataFormatter;
 import org.apache.poi.ss.usermodel.Row;
 import org.apache.poi.ss.usermodel.Sheet;
-import org.apache.poi.ss.usermodel.Workbook;
-import org.apache.poi.ss.usermodel.WorkbookFactory;
 import org.apache.poi.xssf.streaming.SXSSFWorkbook;
 import org.springframework.scheduling.annotation.Scheduled;
 import org.springframework.stereotype.Service;
@@ -33,7 +30,6 @@ import org.springframework.web.multipart.MultipartFile;
 
 import java.io.ByteArrayOutputStream;
 import java.io.File;
-import java.io.FileInputStream;
 import java.nio.file.Files;
 import java.math.BigDecimal;
 import java.math.RoundingMode;
@@ -485,45 +481,136 @@ public class SkipPriceAsinService {
     private void processImportFile(File tempFile, String filename, Long groupId, String shopName,
                                    boolean deleteMode, QueryAsinImportProgressVo progress) throws Exception {
         validateExcelFilename(filename);
-        try (FileInputStream inputStream = new FileInputStream(tempFile);
-             Workbook workbook = WorkbookFactory.create(inputStream)) {
-            Sheet sheet = workbook.getNumberOfSheets() == 0 ? null : workbook.getSheetAt(0);
-            if (sheet == null) {
-                throw new BusinessException("Excel 为空");
+        // 2026-09 全维度审查 C5：改逐行流式解析（ExcelStreamReader → EasyExcel SAX）。
+        // 此前 WorkbookFactory.create 把整表读成 DOM（50 万行约 1~2GB 堆），且行数上限在整表加载
+        // 完之后才校验（等于没有防线）。现在内存只保留表头映射与计数（不随文件行数增长），
+        // 行数上限在迭代过程中即时生效，进度总行数用 onSheetTotal 的近似值。
+        ImportSession session = new ImportSession(progress, groupId, shopName, deleteMode);
+        ExcelStreamReader.readFirstSheet(tempFile, new ExcelStreamReader.SheetRowHandler() {
+            @Override
+            public void onHeader(String sheetName, Integer sheetNo, Map<Integer, String> headerMap) {
+                session.acceptHeader(headerMap);
             }
-            HeaderMapping mapping = resolveHeaderMapping(sheet);
+
+            @Override
+            public void onSheetTotal(String sheetName, Integer sheetNo, Integer approximateTotalRows) {
+                session.acceptApproximateTotalRows(approximateTotalRows);
+            }
+
+            @Override
+            public void onRow(String sheetName, Integer sheetNo, int rowIndex,
+                              Map<Integer, String> headerMap, Map<Integer, String> rowMap) {
+                session.acceptRow(rowIndex, rowMap);
+            }
+        });
+        session.finish();
+    }
+
+    /**
+     * 流式导入的累加器：表头映射 + 计数与进度，内存只与单行/单批写入相关，不随文件行数增长。
+     * 行号沿用 EasyExcel 的 sheet 绝对 0 基行号（与旧 POI Row.getRowNum() 口径一致）。
+     */
+    private final class ImportSession {
+
+        private final QueryAsinImportProgressVo progress;
+        private final Long groupId;
+        private final String shopName;
+        private final boolean deleteMode;
+        private final ImportCounters counters = new ImportCounters();
+        private Map<Integer, String> firstHeaderMap = Map.of();
+        private HeaderMapping mapping;
+        private Integer approximateTotalRows;
+        private boolean headerSeen;
+        private int processedRows;
+
+        private ImportSession(QueryAsinImportProgressVo progress, Long groupId, String shopName, boolean deleteMode) {
+            this.progress = progress;
+            this.groupId = groupId;
+            this.shopName = shopName;
+            this.deleteMode = deleteMode;
+        }
+
+        private void acceptHeader(Map<Integer, String> headerMap) {
+            this.firstHeaderMap = headerMap == null ? Map.of() : headerMap;
+            this.headerSeen = true;
+        }
+
+        private void acceptApproximateTotalRows(Integer approximateTotalRows) {
+            this.approximateTotalRows = approximateTotalRows;
+        }
+
+        private void acceptRow(int rowIndex, Map<Integer, String> rowMap) {
+            if (!headerSeen) {
+                // 表头行整行为空时 EasyExcel 不上报表头回调：与旧实现 sheet.getRow(0) == null 一致
+                throw new BusinessException("Excel 表头为空");
+            }
+            if (mapping == null) {
+                // 两行表头格式下，第 1 行（0 基）就是第二行表头，由首个 onRow 回调提供；
+                // 第 1 行为空行时 EasyExcel 不上报，等价于旧实现 secondHeader 为空行。
+                Map<Integer, String> secondHeaderMap = rowIndex == 1 ? rowMap : null;
+                long resolveStartedAt = System.nanoTime();
+                mapping = resolveHeaderMapping(firstHeaderMap, secondHeaderMap);
                 if (mapping.asinColumns().isEmpty()) {
                     throw new BusinessException("Excel 至少需要包含一个国家的 ASIN 列");
                 }
-
-            int firstDataRow = mapping.firstDataRowIndex();
-            int lastRow = Math.max(sheet.getLastRowNum(), firstDataRow - 1);
-            int totalRows = Math.max(0, lastRow - firstDataRow + 1);
-            if (maxImportRows > 0 && totalRows > maxImportRows) {
+                updateTotalRows();
+                log.info("[skip-price-asin-import] 流式解析表头 firstDataRowIndex={} asinColumns={} priceColumns={} "
+                                + "approxTotalRows={} totalRows={} resolveMs={} deleteMode={}",
+                        mapping.firstDataRowIndex(), mapping.asinColumns(), mapping.priceColumns(),
+                        approximateTotalRows, progress.getTotalRows(), elapsedMs(resolveStartedAt), deleteMode);
+                if (mapping.firstDataRowIndex() > rowIndex) {
+                    // 该行是第二行表头（未进入数据区）：不作为数据行处理
+                    return;
+                }
+            }
+            processedRows++;
+            if (maxImportRows > 0 && processedRows > maxImportRows) {
+                // 流式下无法前置知道总行数：超限即中断（旧实现在整表加载后才校验，实际是 OOM 之后才失败）
                 throw new BusinessException("导入行数超过上限: " + maxImportRows);
             }
-            progress.setTotalRows(totalRows);
-
-            DataFormatter formatter = new DataFormatter();
-            ImportCounters counters = new ImportCounters();
-            for (int rowIndex = firstDataRow; rowIndex <= lastRow; rowIndex++) {
-                Row row = sheet.getRow(rowIndex);
-                if (row == null || isBlankRow(row, formatter)) {
+            if (isBlankRow(rowMap)) {
                 counters.skippedCount++;
-                    updateImportProgress(progress, counters, rowIndex - firstDataRow + 1);
-                    continue;
+                updateImportProgress(progress, counters, processedRows);
+                return;
             }
             if (deleteMode) {
-                    consumeDeleteImportRow(row, formatter, mapping, groupId, shopName, counters);
+                consumeDeleteImportRow(rowMap, mapping, groupId, shopName, counters);
             } else {
-                    consumeAddImportRow(row, formatter, mapping, groupId, shopName, counters);
+                consumeAddImportRow(rowMap, mapping, groupId, shopName, counters);
             }
-                updateImportProgress(progress, counters, rowIndex - firstDataRow + 1);
+            updateImportProgress(progress, counters, processedRows);
         }
+
+        /** 近似总行数换算为数据行数（近似值 = 含表头的物理末行行号，减去数据起始行号）。 */
+        private void updateTotalRows() {
+            if (progress == null || approximateTotalRows == null || approximateTotalRows <= 0) {
+                return;
+            }
+            progress.setTotalRows(Math.max(0, approximateTotalRows - mapping.firstDataRowIndex()));
+        }
+
+        private void finish() {
+            if (!headerSeen) {
+                // 空表（无任何行）或表头行缺失：与旧实现 sheet.getRow(0) == null 一致
+                throw new BusinessException("Excel 表头为空");
+            }
+            Integer totalRows = progress == null ? null : progress.getTotalRows();
+            if (progress != null && (totalRows == null || processedRows > totalRows)) {
+                // 近似总行数缺失（或与实际物理行数不一致）时用实际处理行数兜底，避免展示成 x/0
+                progress.setTotalRows(processedRows);
+            }
+            log.info("[skip-price-asin-import] 流式导入完成 processedRows={} totalRows={} asinCount={} "
+                            + "insertedCount={} deletedCount={} skippedCount={} deleteMode={}",
+                    processedRows, progress == null ? null : progress.getTotalRows(), counters.asinCount,
+                    counters.insertedCount, counters.deletedCount, counters.skippedCount, deleteMode);
         }
     }
 
-    private void consumeAddImportRow(Row row, DataFormatter formatter, HeaderMapping mapping,
+    private long elapsedMs(long startedAtNanos) {
+        return (System.nanoTime() - startedAtNanos) / 1_000_000;
+    }
+
+    private void consumeAddImportRow(Map<Integer, String> rowMap, HeaderMapping mapping,
                                      Long groupId, String shopName, ImportCounters counters) {
         boolean acceptedAny = false;
         boolean sawAnyAsin = false;
@@ -532,7 +619,7 @@ public class SkipPriceAsinService {
         entity.setShopName(shopName);
         for (Map.Entry<String, Integer> entry : mapping.asinColumns().entrySet()) {
             String country = entry.getKey();
-            String asin = normalizeBlank(cellText(row, entry.getValue(), formatter)).toUpperCase(Locale.ROOT);
+            String asin = normalizeBlank(cellText(rowMap, entry.getValue())).toUpperCase(Locale.ROOT);
             if (asin.isEmpty()) {
                 continue;
             }
@@ -543,7 +630,7 @@ public class SkipPriceAsinService {
                 continue;
             }
             BigDecimal minimumPrice = parseOptionalMinimumPrice(
-                    cellText(row, mapping.priceColumns().get(country), formatter));
+                    cellText(rowMap, mapping.priceColumns().get(country)));
             if (existsCountryAsin(groupId, shopName, country, asin)) {
                 counters.skippedCount++;
                 continue;
@@ -561,13 +648,13 @@ public class SkipPriceAsinService {
         skipPriceAsinMapper.insert(entity);
     }
 
-    private void consumeDeleteImportRow(Row row, DataFormatter formatter, HeaderMapping mapping,
+    private void consumeDeleteImportRow(Map<Integer, String> rowMap, HeaderMapping mapping,
                                         Long groupId, String shopName, ImportCounters counters) {
         boolean acceptedAny = false;
         boolean sawAnyAsin = false;
         for (Map.Entry<String, Integer> entry : mapping.asinColumns().entrySet()) {
             String country = entry.getKey();
-            String asin = normalizeBlank(cellText(row, entry.getValue(), formatter)).toUpperCase(Locale.ROOT);
+            String asin = normalizeBlank(cellText(rowMap, entry.getValue())).toUpperCase(Locale.ROOT);
             if (asin.isEmpty()) {
                 continue;
             }
@@ -646,27 +733,43 @@ public class SkipPriceAsinService {
         return skipPriceAsinMapper.selectOne(query.last("LIMIT 1"));
     }
 
-    private HeaderMapping resolveHeaderMapping(Sheet sheet) {
-        DataFormatter formatter = new DataFormatter();
-        Row firstHeader = sheet.getRow(0);
-        Row secondHeader = sheet.getRow(1);
-        if (firstHeader == null) {
+    /**
+     * 由流式表头（第 0 行 headerMap）与第二行表头（两行表头格式时的首个 onRow 回调，可为 null）解析列映射。
+     * 保持与旧 DOM 实现完全相同的列别名匹配与 firstDataRowIndex 语义：
+     * 两行表头（国家行 + 字段行）匹配到 ASIN 列时 firstDataRowIndex = 2；
+     * 否则退回单行表头（国家与字段同列，如「德国ASIN」）匹配，firstDataRowIndex = 1。
+     */
+    private HeaderMapping resolveHeaderMapping(Map<Integer, String> firstHeaderMap,
+                                               Map<Integer, String> secondHeaderMap) {
+        if (firstHeaderMap == null || firstHeaderMap.isEmpty()) {
             throw new BusinessException("Excel 表头为空");
         }
 
-        int maxColumn = Math.max(lastCellNum(firstHeader), lastCellNum(secondHeader));
+        int maxColumn = Math.max(lastColumnCount(firstHeaderMap), lastColumnCount(secondHeaderMap));
         Map<String, Integer> asinColumns = new LinkedHashMap<>();
         Map<String, Integer> priceColumns = new LinkedHashMap<>();
+        // 兼容合并单元格表头：EasyExcel SAX 不回填被合并覆盖的单元格文本（旧 DOM 的 cellTextWithMerged
+        // 会读到主单元格），这里用「字段列未显式标注国家时继承左侧最近国家列」还原同样的列映射结果。
+        String carriedCountry = "";
+        List<Integer> carriedColumns = new ArrayList<>();
         for (int columnIndex = 0; columnIndex < maxColumn; columnIndex++) {
-            String countryText = cellTextWithMerged(sheet, firstHeader, columnIndex, formatter);
-            String fieldText = cellText(secondHeader, columnIndex, formatter);
+            String countryText = cellText(firstHeaderMap, columnIndex);
+            String fieldText = cellText(secondHeaderMap, columnIndex);
             String country = normalizeCountryAlias(countryText);
             String fieldKey = normalizeHeaderKey(fieldText);
             if (country.isEmpty()) {
                 country = normalizeCountryAlias(fieldText);
                 fieldKey = normalizeHeaderKey(countryText);
             }
-            if (country.isEmpty() || !SUPPORTED_COUNTRIES.contains(country)) {
+            if (country.isEmpty() && !carriedCountry.isEmpty()) {
+                String inheritedKey = normalizeHeaderKey(fieldText);
+                if (isAsinHeader(inheritedKey) || isPriceHeader(inheritedKey)) {
+                    country = carriedCountry;
+                    fieldKey = inheritedKey;
+                    carriedColumns.add(columnIndex);
+                }
+            }
+            if (country.isEmpty()) {
                 continue;
             }
             if (isAsinHeader(fieldKey)) {
@@ -674,10 +777,11 @@ public class SkipPriceAsinService {
             } else if (isPriceHeader(fieldKey)) {
                 priceColumns.putIfAbsent(country, columnIndex);
             }
+            carriedCountry = country;
         }
         if (asinColumns.isEmpty()) {
             for (int columnIndex = 0; columnIndex < maxColumn; columnIndex++) {
-                String headerText = cellText(firstHeader, columnIndex, formatter);
+                String headerText = cellText(firstHeaderMap, columnIndex);
                 String key = normalizeHeaderKey(headerText);
                 String country = normalizeCountryAlias(headerText);
                 if (country.isEmpty()) {
@@ -687,13 +791,31 @@ public class SkipPriceAsinService {
                     asinColumns.putIfAbsent(country, columnIndex);
                 }
             }
+            if (!carriedColumns.isEmpty()) {
+                log.info("[skip-price-asin-import] 表头按合并单元格口径继承国家列 carriedColumns={} 但未匹配到 ASIN 列",
+                        carriedColumns);
+            }
             return new HeaderMapping(asinColumns, priceColumns, 1);
         }
+        if (!carriedColumns.isEmpty()) {
+            log.info("[skip-price-asin-import] 检测到疑似合并单元格表头，被覆盖列按左侧国家列归属 carriedColumns={}",
+                    carriedColumns);
+        }
         return new HeaderMapping(asinColumns, priceColumns, 2);
     }
 
-    private int lastCellNum(Row row) {
-        return row == null || row.getLastCellNum() < 0 ? 0 : row.getLastCellNum();
+    /** 流式表头行最大列数（与旧 Row.getLastCellNum() 口径一致：无单元格时为 0）。 */
+    private int lastColumnCount(Map<Integer, String> rowMap) {
+        if (rowMap == null || rowMap.isEmpty()) {
+            return 0;
+        }
+        int lastColumnCount = 0;
+        for (Integer columnIndex : rowMap.keySet()) {
+            if (columnIndex != null && columnIndex >= lastColumnCount) {
+                lastColumnCount = columnIndex + 1;
+            }
+        }
+        return lastColumnCount;
     }
 
     private boolean isAsinHeader(String key) {
@@ -705,27 +827,12 @@ public class SkipPriceAsinService {
                 || key.contains("minprice") || key.equals("price");
     }
 
-    private String cellTextWithMerged(Sheet sheet, Row row, int columnIndex, DataFormatter formatter) {
-        if (row == null) {
+    /** 流式解析下按列索引取行内文本（归一化口径与旧 cellText(Row, ...) 一致）。 */
+    private String cellText(Map<Integer, String> rowMap, Integer columnIndex) {
+        if (rowMap == null || columnIndex == null || columnIndex < 0) {
             return "";
         }
-        int rowIndex = row.getRowNum();
-        for (int i = 0; i < sheet.getNumMergedRegions(); i++) {
-            var region = sheet.getMergedRegion(i);
-            if (region.isInRange(rowIndex, columnIndex)) {
-                Row firstRow = sheet.getRow(region.getFirstRow());
-                return cellText(firstRow, region.getFirstColumn(), formatter);
-            }
-        }
-        return cellText(row, columnIndex, formatter);
-    }
-
-    private String cellText(Row row, Integer columnIndex, DataFormatter formatter) {
-        if (row == null || columnIndex == null || columnIndex < 0) {
-            return "";
-        }
-        Cell cell = row.getCell(columnIndex);
-        return cell == null ? "" : normalizeExcelText(formatter.formatCellValue(cell));
+        return normalizeExcelText(rowMap.get(columnIndex));
     }
 
     private String normalizeExcelText(String value) {
@@ -798,12 +905,13 @@ public class SkipPriceAsinService {
         return left.compareTo(right) == 0;
     }
 
-    private boolean isBlankRow(Row row, DataFormatter formatter) {
-        if (row == null) {
+    /** 该行所有单元格文本是否全空白（归一化口径与旧 isBlankRow(Row, DataFormatter) 一致）。 */
+    private boolean isBlankRow(Map<Integer, String> rowMap) {
+        if (rowMap == null || rowMap.isEmpty()) {
             return true;
         }
-        for (int i = 0; i < lastCellNum(row); i++) {
-            if (!cellText(row, i, formatter).isEmpty()) {
+        for (String value : rowMap.values()) {
+            if (!normalizeExcelText(value).isEmpty()) {
                 return false;
             }
         }

backend-java/src/main/java/com/nanri/aiimage/modules/shopmatch/controller/ShopMatchController.java

@@ -5,12 +5,12 @@ import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCandidateAddRe
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCountryPreferenceSaveRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskTaskBatchRequest;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCountryPreferenceVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskHistoryVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskTaskBatchVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCountryPreferenceVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskHistoryVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskTaskBatchVo;
 import com.nanri.aiimage.modules.shopmatch.model.dto.ShopMatchCreateTaskRequest;
 import com.nanri.aiimage.modules.shopmatch.model.dto.ShopMatchStageCompleteRequest;
 import com.nanri.aiimage.modules.shopmatch.model.dto.ShopMatchSubmitResultRequest;

backend-java/src/main/java/com/nanri/aiimage/modules/shopmatch/service/ShopMatchResolveService.java

@@ -8,10 +8,10 @@ import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCandidateAddRe
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCountryPreferenceSaveRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
 import com.nanri.aiimage.modules.productrisk.model.enums.ProductRiskCountryCode;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCountryPreferenceVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskShopQueueItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCountryPreferenceVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskShopQueueItemVo;
 import com.nanri.aiimage.modules.shopmatch.mapper.ShopMatchCountryPrefMapper;
 import com.nanri.aiimage.modules.shopmatch.mapper.ShopMatchShopCandidateMapper;
 import com.nanri.aiimage.modules.shopmatch.model.entity.ShopMatchCountryPrefEntity;

backend-java/src/main/java/com/nanri/aiimage/modules/shopmatch/service/ShopMatchTaskHeartbeatSpi.java

@@ -0,0 +1,35 @@
+package com.nanri.aiimage.modules.shopmatch.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 定时匹配 的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>心跳逻辑（缓存刷新 / 任务缓存回写）从 task 模块收回本模块，task 侧只依赖 SPI 接口，
+ * 消除 task → 业务模块的编译期依赖。
+ */
+@Service
+@RequiredArgsConstructor
+public class ShopMatchTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final ShopMatchTaskCacheService cacheService;
+
+    @Override
+    public String moduleType() {
+        return "SHOP_MATCH";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.touchTaskHeartbeat(taskId);
+    }
+
+    @Override
+    public void saveTaskCache(FileTaskEntity task) {
+        cacheService.saveTaskCache(task);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/shopmatch/service/ShopMatchTaskService.java

@@ -9,13 +9,13 @@ import com.nanri.aiimage.common.exception.BusinessException;
 import com.nanri.aiimage.config.TaskPressureProperties;
 import com.nanri.aiimage.modules.file.service.oss.OssStorageService;
 import com.nanri.aiimage.modules.productrisk.model.enums.ProductRiskCountryCode;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskHistoryVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskResultItemVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskShopQueueItemVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskTaskBatchVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskTaskDetailVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskTaskItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskHistoryVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskResultItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskShopQueueItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskTaskBatchVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskTaskDetailVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskTaskItemVo;
 import com.nanri.aiimage.modules.pricetrack.model.vo.SkipPriceAsinPageVo;
 import com.nanri.aiimage.modules.queryasin.model.dto.QueryAsinCountryAsinsDto;
 import com.nanri.aiimage.modules.shopmatch.mapper.ShopMatchShopCandidateMapper;

backend-java/src/main/java/com/nanri/aiimage/modules/similarasin/service/SimilarAsinTaskHeartbeatSpi.java

@@ -0,0 +1,53 @@
+package com.nanri.aiimage.modules.similarasin.service;
+
+import com.nanri.aiimage.config.SimilarAsinProperties;
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.service.TaskFileJobService;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+import java.time.Duration;
+
+/**
+ * 货源查询（相似 ASIN）的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>本模块的心跳有两处特殊：
+ * 1. 客户端心跳间隔远小于 stale 判定阈值，因此 biz_file_task.updated_at 采用**检查点节流**
+ *    （按 similarAsinProperties 配置的间隔刷新），避免每次心跳都写库；
+ * 2. 心跳时顺带刷新结果文件组装作业的租约（避免作业被误判卡死）。
+ */
+@Service
+@RequiredArgsConstructor
+public class SimilarAsinTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final SimilarAsinTaskCacheService cacheService;
+    private final TaskFileJobService taskFileJobService;
+    private final SimilarAsinProperties similarAsinProperties;
+
+    @Override
+    public String moduleType() {
+        return "SIMILAR_ASIN";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.touchTaskHeartbeat(taskId);
+    }
+
+    @Override
+    public void onHeartbeat(FileTaskEntity task) {
+        taskFileJobService.touchRunningAssembleJobsIfStale(
+                task.getId(), moduleType(), checkpointIntervalMillis());
+    }
+
+    @Override
+    public Duration checkpointInterval() {
+        return Duration.ofMillis(checkpointIntervalMillis());
+    }
+
+    private long checkpointIntervalMillis() {
+        return Math.max(1_000L, similarAsinProperties.getDbTaskTouchIntervalMillis());
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/task/service/ModuleHistoryCleanupService.java

@@ -4,8 +4,7 @@ import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.baomidou.mybatisplus.core.conditions.update.LambdaUpdateWrapper;
 import com.nanri.aiimage.common.service.DistributedJobLockService;
 import com.nanri.aiimage.config.ModuleCleanupProperties;
-import com.nanri.aiimage.modules.collectdata.mapper.CollectDataItemMapper;
-import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataItemEntity;
+import com.nanri.aiimage.modules.task.spi.CollectDataItemCleanupSpi;
 import com.nanri.aiimage.modules.task.mapper.FileResultMapper;
 import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
 import com.nanri.aiimage.modules.task.mapper.TaskChunkMapper;
@@ -55,7 +54,7 @@ public class ModuleHistoryCleanupService {
     private final TaskResultPayloadMapper taskResultPayloadMapper;
     private final TaskScopeStateMapper taskScopeStateMapper;
     private final TaskChunkMapper taskChunkMapper;
-    private final CollectDataItemMapper collectDataItemMapper;
+    private final CollectDataItemCleanupSpi collectDataItemCleanupSpi;
     private final DistributedJobLockService distributedJobLockService;
     private final TransientPayloadDeleteOrchestrator transientPayloadDeleteOrchestrator;
     private final TransactionTemplate transactionTemplate;
@@ -76,7 +75,7 @@ public class ModuleHistoryCleanupService {
                                        TaskResultPayloadMapper taskResultPayloadMapper,
                                        TaskScopeStateMapper taskScopeStateMapper,
                                        TaskChunkMapper taskChunkMapper,
-                                       CollectDataItemMapper collectDataItemMapper,
+                                       CollectDataItemCleanupSpi collectDataItemCleanupSpi,
                                        DistributedJobLockService distributedJobLockService,
                                        TransientPayloadDeleteOrchestrator transientPayloadDeleteOrchestrator,
                                        PlatformTransactionManager platformTransactionManager) {
@@ -89,7 +88,7 @@ public class ModuleHistoryCleanupService {
         this.taskResultPayloadMapper = taskResultPayloadMapper;
         this.taskScopeStateMapper = taskScopeStateMapper;
         this.taskChunkMapper = taskChunkMapper;
-        this.collectDataItemMapper = collectDataItemMapper;
+        this.collectDataItemCleanupSpi = collectDataItemCleanupSpi;
         this.distributedJobLockService = distributedJobLockService;
         this.transientPayloadDeleteOrchestrator = transientPayloadDeleteOrchestrator;
         this.transactionTemplate = new TransactionTemplate(platformTransactionManager);
@@ -218,8 +217,8 @@ public class ModuleHistoryCleanupService {
                 .in(TaskChunkEntity::getTaskId, cleanupTaskIds));
 
         if (!collectDataTaskIds.isEmpty()) {
-            collectDataItemMapper.delete(new LambdaQueryWrapper<CollectDataItemEntity>()
-                    .in(CollectDataItemEntity::getTaskId, collectDataTaskIds));
+            // 采集明细属于 collectdata 模块（G5：消除 task → 业务依赖，走 SPI）
+            collectDataItemCleanupSpi.deleteItemsByTaskIds(collectDataTaskIds);
         }
 
         fileResultMapper.delete(new LambdaQueryWrapper<FileResultEntity>()

backend-java/src/main/java/com/nanri/aiimage/modules/task/service/StaleTaskRepairService.java

@@ -3,8 +3,7 @@ package com.nanri.aiimage.modules.task.service;
 import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.baomidou.mybatisplus.core.conditions.update.LambdaUpdateWrapper;
 import com.nanri.aiimage.common.service.DistributedJobLockService;
-import com.nanri.aiimage.modules.brand.mapper.BrandCrawlTaskMapper;
-import com.nanri.aiimage.modules.brand.model.entity.BrandCrawlTaskEntity;
+import com.nanri.aiimage.modules.task.spi.BrandTaskStaleRepairSpi;
 import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
 import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
 import lombok.RequiredArgsConstructor;
@@ -32,16 +31,15 @@ public class StaleTaskRepairService {
     private static final String STATUS_RUNNING = "RUNNING";
     private static final String STATUS_FAILED = "FAILED";
 
-    private static final String STATUS_BRAND_PENDING = "pending";
-    private static final String STATUS_BRAND_RUNNING = "running";
-    private static final String STATUS_BRAND_FAILED = "failed";
-    private static final String STATUS_BRAND_CANCELLED = "cancelled";
 
     /** 中间态（无人接单）或 RUNNING（心跳残留）超过该时长未更新即判死。真活任务心跳会不断刷新 updatedAt，不会被误杀。 */
     private static final long STALE_IDLE_MINUTES = 120;
 
+    /** 单轮品牌陈旧任务处理上限（原实现内联 limit 500）。 */
+    private static final int MAX_BRAND_STALE_BATCH = 500;
+
     private final FileTaskMapper fileTaskMapper;
-    private final BrandCrawlTaskMapper brandCrawlTaskMapper;
+    private final BrandTaskStaleRepairSpi brandTaskStaleRepairSpi;
     private final DistributedJobLockService distributedJobLockService;
 
     /**
@@ -117,26 +115,15 @@ public class StaleTaskRepairService {
         }
     }
 
-    /** brand_crawl_tasks 的 pending/running 陈旧 → failed。 */
+    /**
+     * brand_crawl_tasks 的 pending/running 陈旧 → failed。
+     * 品牌表与状态字面量属于 brand 模块，实现放在 BrandTaskStaleRepairSpiImpl（G5：消除 task → 业务依赖）。
+     */
     private void repairBrandStale() {
         LocalDateTime cutoff = LocalDateTime.now().minusMinutes(STALE_IDLE_MINUTES);
-        List<BrandCrawlTaskEntity> stale = brandCrawlTaskMapper.selectList(new LambdaQueryWrapper<BrandCrawlTaskEntity>()
-                .select(BrandCrawlTaskEntity::getId)
-                .in(BrandCrawlTaskEntity::getStatus, STATUS_BRAND_PENDING, STATUS_BRAND_RUNNING)
-                .lt(BrandCrawlTaskEntity::getUpdatedAt, cutoff)
-                .last("limit 500"));
-        if (stale.isEmpty()) {
-            return;
-        }
-        int updated = brandCrawlTaskMapper.update(null, new LambdaUpdateWrapper<BrandCrawlTaskEntity>()
-                .in(BrandCrawlTaskEntity::getId, stale.stream().map(BrandCrawlTaskEntity::getId).toList())
-                .in(BrandCrawlTaskEntity::getStatus, STATUS_BRAND_PENDING, STATUS_BRAND_RUNNING)
-                .lt(BrandCrawlTaskEntity::getUpdatedAt, cutoff)
-                .set(BrandCrawlTaskEntity::getStatus, STATUS_BRAND_FAILED)
-                .set(BrandCrawlTaskEntity::getErrorMessage, "任务长期无心跳，已自动失败"));
-        if (updated > 0) {
-            log.warn("[stale-task-repair] brand 陈旧任务已标失败 count={} ids={}",
-                    updated, stale.stream().map(BrandCrawlTaskEntity::getId).toList());
+        List<Long> failedIds = brandTaskStaleRepairSpi.failStaleBrandTasks(cutoff, MAX_BRAND_STALE_BATCH);
+        if (!failedIds.isEmpty()) {
+            log.warn("[stale-task-repair] brand 陈旧任务已标失败 count={} ids={}", failedIds.size(), failedIds);
         }
     }
 }

backend-java/src/main/java/com/nanri/aiimage/modules/task/service/TaskHeartbeatService.java

@@ -2,76 +2,66 @@ package com.nanri.aiimage.modules.task.service;
 
 import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.baomidou.mybatisplus.core.conditions.update.LambdaUpdateWrapper;
-import com.nanri.aiimage.config.SimilarAsinProperties;
-import com.nanri.aiimage.modules.appearancepatent.service.AppearancePatentTaskCacheService;
-import com.nanri.aiimage.modules.brand.mapper.BrandCrawlTaskMapper;
-import com.nanri.aiimage.modules.brand.model.entity.BrandCrawlTaskEntity;
-import com.nanri.aiimage.modules.brand.service.BrandTaskProgressCacheService;
-import com.nanri.aiimage.modules.collectdata.service.CollectDataService;
-import com.nanri.aiimage.modules.deletebrand.service.DeleteBrandTaskCacheService;
-import com.nanri.aiimage.modules.patroldelete.service.PatrolDeleteTaskCacheService;
-import com.nanri.aiimage.modules.pricetrack.service.PriceTrackTaskCacheService;
-import com.nanri.aiimage.modules.productrisk.service.ProductRiskTaskCacheService;
-import com.nanri.aiimage.modules.publish.service.PublishTaskService;
-import com.nanri.aiimage.modules.queryasin.service.QueryAsinTaskCacheService;
-import com.nanri.aiimage.modules.shopdatacrawl.service.ShopDataCrawlTaskCacheService;
-import com.nanri.aiimage.modules.shopdatacrawl.service.ShopDataCrawlTaskService;
-import com.nanri.aiimage.modules.shopmatch.service.ShopMatchTaskCacheService;
-import com.nanri.aiimage.modules.similarasin.service.SimilarAsinTaskCacheService;
 import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
 import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
 import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
 import com.nanri.aiimage.modules.task.model.vo.TaskHeartbeatVo;
-import com.nanri.aiimage.modules.withdraw.service.WithdrawTaskCacheService;
-import lombok.RequiredArgsConstructor;
+import com.nanri.aiimage.modules.task.spi.BrandTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
 import lombok.extern.slf4j.Slf4j;
 import org.springframework.stereotype.Service;
 
 import java.time.Duration;
 import java.time.LocalDateTime;
 import java.util.LinkedHashMap;
+import java.util.List;
 import java.util.Map;
 
+/**
+ * 任务心跳入口。
+ *
+ * <p>2026-09 全维度审查 G5：本类此前直接注入 12 个业务模块的 CacheService/Service 并用 switch 分发，
+ * 与"业务模块实现 task 的 Handler SPI"形成 task ↔ 业务的循环依赖。现改为依赖
+ * {@link TaskModuleHeartbeatSpi}（各业务模块自己实现）与 {@link BrandTaskHeartbeatSpi}（品牌任务，
+ * 独立于 biz_file_task），task 侧不再 import 任何业务模块。
+ */
 @Service
-@RequiredArgsConstructor
 @Slf4j
 public class TaskHeartbeatService {
 
     private static final String STATUS_RUNNING = "RUNNING";
-    private static final String BRAND_STATUS_RUNNING = "running";
-
-    private static final String MODULE_DELETE_BRAND = "DELETE_BRAND";
-    private static final String MODULE_PRODUCT_RISK = "PRODUCT_RISK_RESOLVE";
-    private static final String MODULE_PRICE_TRACK = "PRICE_TRACK";
-    private static final String MODULE_SHOP_MATCH = "SHOP_MATCH";
-    private static final String MODULE_PATROL_DELETE = "PATROL_DELETE";
-    private static final String MODULE_QUERY_ASIN = "QUERY_ASIN";
-    private static final String MODULE_SHOP_DATA_CRAWL = "SHOP_DATA_CRAWL";
-    private static final String MODULE_WITHDRAW = "WITHDRAW";
-    private static final String MODULE_APPEARANCE_PATENT = "APPEARANCE_PATENT";
-    private static final String MODULE_SIMILAR_ASIN = "SIMILAR_ASIN";
-    private static final String MODULE_COLLECT_DATA = "COLLECT_DATA";
     private static final String MODULE_BRAND = "BRAND";
+    private static final String MODULE_SHOP_DATA_CRAWL = "SHOP_DATA_CRAWL";
 
     private final FileTaskMapper fileTaskMapper;
     private final TaskProgressSnapshotService taskProgressSnapshotService;
-    private final BrandCrawlTaskMapper brandCrawlTaskMapper;
-    private final ProductRiskTaskCacheService productRiskTaskCacheService;
-    private final PublishTaskService publishTaskService;
-    private final PriceTrackTaskCacheService priceTrackTaskCacheService;
-    private final ShopMatchTaskCacheService shopMatchTaskCacheService;
-    private final PatrolDeleteTaskCacheService patrolDeleteTaskCacheService;
-    private final QueryAsinTaskCacheService queryAsinTaskCacheService;
-    private final ShopDataCrawlTaskCacheService shopDataCrawlTaskCacheService;
-    private final ShopDataCrawlTaskService shopDataCrawlTaskService;
-    private final WithdrawTaskCacheService withdrawTaskCacheService;
-    private final AppearancePatentTaskCacheService appearancePatentTaskCacheService;
-    private final SimilarAsinTaskCacheService similarAsinTaskCacheService;
-    private final SimilarAsinProperties similarAsinProperties;
-    private final TaskFileJobService taskFileJobService;
-    private final DeleteBrandTaskCacheService deleteBrandTaskCacheService;
-    private final BrandTaskProgressCacheService brandTaskProgressCacheService;
-    private final CollectDataService collectDataService;
+    private final BrandTaskHeartbeatSpi brandTaskHeartbeatSpi;
+    /** moduleType → 模块心跳实现（启动时建索引并校验重复注册） */
+    private final Map<String, TaskModuleHeartbeatSpi> moduleHandlers;
+
+    public TaskHeartbeatService(FileTaskMapper fileTaskMapper,
+                                TaskProgressSnapshotService taskProgressSnapshotService,
+                                BrandTaskHeartbeatSpi brandTaskHeartbeatSpi,
+                                List<TaskModuleHeartbeatSpi> moduleHeartbeatHandlers) {
+        this.fileTaskMapper = fileTaskMapper;
+        this.taskProgressSnapshotService = taskProgressSnapshotService;
+        this.brandTaskHeartbeatSpi = brandTaskHeartbeatSpi;
+        Map<String, TaskModuleHeartbeatSpi> handlers = new LinkedHashMap<>();
+        for (TaskModuleHeartbeatSpi handler : moduleHeartbeatHandlers == null ? List.<TaskModuleHeartbeatSpi>of() : moduleHeartbeatHandlers) {
+            String moduleType = handler.moduleType();
+            if (moduleType == null || moduleType.isBlank()) {
+                throw new IllegalStateException("TaskModuleHeartbeatSpi 未声明 moduleType: "
+                        + handler.getClass().getName());
+            }
+            TaskModuleHeartbeatSpi exists = handlers.put(moduleType, handler);
+            if (exists != null) {
+                throw new IllegalStateException("moduleType=" + moduleType + " 注册了多个心跳实现: "
+                        + exists.getClass().getName() + " / " + handler.getClass().getName());
+            }
+        }
+        this.moduleHandlers = Map.copyOf(handlers);
+        log.info("[task-heartbeat-spi] 模块心跳实现注册完成 count={} modules={}", handlers.size(), handlers.keySet());
+    }
 
     public TaskHeartbeatVo heartbeat(Long taskId, TaskHeartbeatRequest request) {
         if (taskId == null || taskId <= 0) {
@@ -82,15 +72,12 @@ public class TaskHeartbeatService {
         String requestedModuleType = request == null || request.getModuleType() == null
                 ? null
                 : request.getModuleType().trim();
-        FileTaskEntity fileTask = MODULE_BRAND.equalsIgnoreCase(requestedModuleType)
-                ? null
-                : selectFileTask(taskId);
-        BrandCrawlTaskEntity brandTask = MODULE_SHOP_DATA_CRAWL.equalsIgnoreCase(requestedModuleType)
-                ? null
-                : selectBrandTask(taskId);
+        // 品牌任务在独立表里：客户端声明是 BRAND 时不必再查 biz_file_task（历史行为保持一致）
+        boolean brandOnly = MODULE_BRAND.equalsIgnoreCase(requestedModuleType);
+        boolean skipBrandLookup = MODULE_SHOP_DATA_CRAWL.equalsIgnoreCase(requestedModuleType);
 
-        TaskHeartbeatVo fileResult = touchFileTaskIfRunning(fileTask, request);
-        TaskHeartbeatVo brandResult = touchBrandTaskIfRunning(brandTask, request);
+        TaskHeartbeatVo fileResult = brandOnly ? null : touchFileTaskIfRunning(selectFileTask(taskId), request);
+        TaskHeartbeatVo brandResult = skipBrandLookup ? null : brandTaskHeartbeatSpi.heartbeat(taskId, request);
 
         if (fileResult != null && fileResult.isAlive() && brandResult != null && brandResult.isAlive()) {
             return TaskHeartbeatVo.alive(fileResult.getModuleType() + "," + MODULE_BRAND, STATUS_RUNNING);
@@ -158,58 +145,34 @@ public class TaskHeartbeatService {
             syncTerminalState(fileTask, status, fileTask.getErrorMessage());
             return TaskHeartbeatVo.notAlive(fileTask.getModuleType(), status, "task is not running");
         }
-        BrandCrawlTaskEntity brandTask = selectBrandTask(taskId);
-        if (brandTask != null) {
-            String status = brandTask.getStatus();
-            if ("running".equalsIgnoreCase(status) || "pending".equalsIgnoreCase(status)) {
-                // 与 file 分支统一改为条件更新：整行 updateById 会拿读取快照覆盖并发写入的字段
-                // （客户端重启上报与品牌任务自身状态流转同时发生时）
-                int updated = brandCrawlTaskMapper.update(null, new LambdaUpdateWrapper<BrandCrawlTaskEntity>()
-                        .eq(BrandCrawlTaskEntity::getId, brandTask.getId())
-                        .in(BrandCrawlTaskEntity::getStatus, "running", "pending", "RUNNING", "PENDING")
-                        .set(BrandCrawlTaskEntity::getStatus, "cancelled")
-                        .set(BrandCrawlTaskEntity::getErrorMessage, safeReason)
-                        .set(BrandCrawlTaskEntity::getUpdatedAt, LocalDateTime.now()));
-                if (updated > 0) {
-                    log.warn("[task-interrupted] brand task marked cancelled by client restart taskId={} reason={}",
-                            taskId, safeReason);
-                    return TaskHeartbeatVo.notAlive(MODULE_BRAND, "cancelled", "marked cancelled");
-                }
-            }
-            log.info("[task-interrupted] brand task not in running/pending, skipped taskId={} status={}", taskId, status);
-            return TaskHeartbeatVo.notAlive(MODULE_BRAND, status, "task is not running");
+        TaskHeartbeatVo brandResult = brandTaskHeartbeatSpi.markInterrupted(taskId, safeReason);
+        if (brandResult != null) {
+            return brandResult;
         }
         log.debug("[task-interrupted] task not found (normal polling) taskId={}", taskId);
         return TaskHeartbeatVo.notAlive(null, null, "task not found");
     }
 
-    private BrandCrawlTaskEntity selectBrandTask(Long taskId) {
-        LambdaQueryWrapper<BrandCrawlTaskEntity> brandQuery = new LambdaQueryWrapper<BrandCrawlTaskEntity>()
-                .eq(BrandCrawlTaskEntity::getId, taskId)
-                .last("limit 1");
-        return brandCrawlTaskMapper.selectOne(brandQuery);
-    }
-
     private TaskHeartbeatVo touchFileTaskIfRunning(FileTaskEntity task, TaskHeartbeatRequest request) {
         if (task == null) {
             return null;
         }
         String moduleType = task.getModuleType() == null ? "" : task.getModuleType();
-        if (PublishTaskService.MODULE_TYPE.equals(moduleType)) {
-            publishTaskService.ensureTaskOwnedByCurrentInstance(task, "publish task heartbeat");
-        }
-        if (MODULE_SHOP_DATA_CRAWL.equals(moduleType)) {
-            shopDataCrawlTaskService.ensureTaskOwnedByCurrentInstance(task, "shop data crawl task heartbeat");
+        TaskModuleHeartbeatSpi handler = moduleHandlers.get(moduleType);
+        if (handler != null) {
+            handler.ensureOwnership(task, moduleType + " task heartbeat");
         }
         String status = task.getStatus();
         if (!STATUS_RUNNING.equals(status)) {
-            log.debug("[task-heartbeat] file task is not running (normal polling) taskId={ actualUserId={} moduleType={} status={}",
+            log.debug("[task-heartbeat] file task is not running (normal polling) taskId={} actualUserId={} moduleType={} status={}",
                     task.getId(), task.getUserId(), moduleType, status);
             return TaskHeartbeatVo.notAlive(moduleType, status, "task is not running");
         }
-        if (MODULE_SIMILAR_ASIN.equals(moduleType)) {
-            return touchSimilarAsinHeartbeat(task);
-        }
+
+        // updated_at 检查点：默认每次心跳刷新；实现声明了节流间隔的（similar-asin）按间隔刷新，
+        // 避免客户端密集心跳把库写满
+        Duration interval = handler == null ? null : handler.checkpointInterval();
+        if (isCheckpointDue(task, interval)) {
             LocalDateTime now = LocalDateTime.now();
             int updated = fileTaskMapper.update(null, new LambdaUpdateWrapper<FileTaskEntity>()
                     .eq(FileTaskEntity::getId, task.getId())
@@ -217,145 +180,37 @@ public class TaskHeartbeatService {
                     .set(FileTaskEntity::getUpdatedAt, now));
             if (updated <= 0) {
                 FileTaskEntity latest = fileTaskMapper.selectById(task.getId());
+                if (interval != null && latest != null && STATUS_RUNNING.equals(latest.getStatus())) {
+                    // 节流下的条件更新未命中：并发心跳刚刷新过，按最新实体继续（不判死）
+                    task = latest;
+                } else {
                     log.warn("[task-heartbeat] file task heartbeat update missed taskId={} actualUserId={} moduleType={} status={} latestStatus={}",
                             task.getId(), task.getUserId(), moduleType, status,
                             latest == null ? null : latest.getStatus());
-            return TaskHeartbeatVo.notAlive(moduleType, latest == null ? status : latest.getStatus(), "task is not running");
+                    return TaskHeartbeatVo.notAlive(moduleType, latest == null ? status : latest.getStatus(),
+                            "task is not running");
                 }
-        touchModuleHeartbeat(moduleType, task.getId(), request);
-        task.setUpdatedAt(now);
-        saveFileTaskCache(moduleType, task);
-        return TaskHeartbeatVo.alive(moduleType, STATUS_RUNNING);
-    }
-
-    private TaskHeartbeatVo touchSimilarAsinHeartbeat(FileTaskEntity task) {
-        LocalDateTime now = LocalDateTime.now();
-        long intervalMillis = Math.max(1_000L, similarAsinProperties.getDbTaskTouchIntervalMillis());
-        LocalDateTime cutoff = now.minus(Duration.ofMillis(intervalMillis));
-        boolean checkpointDue = task.getUpdatedAt() == null || !task.getUpdatedAt().isAfter(cutoff);
-
-        similarAsinTaskCacheService.touchTaskHeartbeat(task.getId());
-        taskFileJobService.touchRunningAssembleJobsIfStale(task.getId(), MODULE_SIMILAR_ASIN, intervalMillis);
-        if (!checkpointDue) {
-            saveFileTaskCache(MODULE_SIMILAR_ASIN, task);
-            return TaskHeartbeatVo.alive(MODULE_SIMILAR_ASIN, STATUS_RUNNING);
-        }
-
-        int updated = fileTaskMapper.update(null, new LambdaUpdateWrapper<FileTaskEntity>()
-                .eq(FileTaskEntity::getId, task.getId())
-                .eq(FileTaskEntity::getStatus, STATUS_RUNNING)
-                .and(wrapper -> wrapper.isNull(FileTaskEntity::getUpdatedAt)
-                        .or()
-                        .le(FileTaskEntity::getUpdatedAt, cutoff))
-                .set(FileTaskEntity::getUpdatedAt, now));
-        if (updated <= 0) {
-            FileTaskEntity latest = fileTaskMapper.selectById(task.getId());
-            if (latest == null || !STATUS_RUNNING.equals(latest.getStatus())) {
-                return TaskHeartbeatVo.notAlive(MODULE_SIMILAR_ASIN,
-                        latest == null ? task.getStatus() : latest.getStatus(), "task is not running");
-            }
-            task = latest;
             } else {
                 task.setUpdatedAt(now);
             }
-        saveFileTaskCache(MODULE_SIMILAR_ASIN, task);
-        return TaskHeartbeatVo.alive(MODULE_SIMILAR_ASIN, STATUS_RUNNING);
         }
 
-    private TaskHeartbeatVo touchBrandTaskIfRunning(BrandCrawlTaskEntity task, TaskHeartbeatRequest request) {
-        if (task == null) {
-            return null;
+        if (handler != null) {
+            handler.touchHeartbeat(task.getId(), request);
+            handler.onHeartbeat(task);
+            handler.saveTaskCache(task);
         }
-        String status = task.getStatus();
-        if (!BRAND_STATUS_RUNNING.equals(status)) {
-            log.warn("[task-heartbeat] brand task is not running taskId={} actualUserId={} status={}",
-                    task.getId(), task.getUserId(), status);
-            return TaskHeartbeatVo.notAlive(MODULE_BRAND, status, "task is not running");
-        }
-        int updated = brandCrawlTaskMapper.update(null, new LambdaUpdateWrapper<BrandCrawlTaskEntity>()
-                .eq(BrandCrawlTaskEntity::getId, task.getId())
-                .eq(BrandCrawlTaskEntity::getStatus, BRAND_STATUS_RUNNING)
-                .set(BrandCrawlTaskEntity::getUpdatedAt, LocalDateTime.now()));
-        if (updated <= 0) {
-            BrandCrawlTaskEntity latest = brandCrawlTaskMapper.selectById(task.getId());
-            log.warn("[task-heartbeat] brand task heartbeat update missed taskId={} actualUserId={} status={} latestStatus={}",
-                    task.getId(), task.getUserId(), status,
-                    latest == null ? null : latest.getStatus());
-            return TaskHeartbeatVo.notAlive(MODULE_BRAND, latest == null ? status : latest.getStatus(), "task is not running");
-        }
-        brandTaskProgressCacheService.touchHeartbeat(
-                task.getId(),
-                request == null ? null : request.getPhase(),
-                request == null ? null : request.getCurrent(),
-                request == null ? null : request.getTotal());
-        return TaskHeartbeatVo.alive(MODULE_BRAND, BRAND_STATUS_RUNNING);
+        return TaskHeartbeatVo.alive(moduleType, STATUS_RUNNING);
     }
 
-    private void touchModuleHeartbeat(String moduleType, Long taskId, TaskHeartbeatRequest request) {
-        switch (moduleType) {
-            case MODULE_PRODUCT_RISK -> {
-                productRiskTaskCacheService.touchTaskHeartbeat(taskId);
-            }
-            case PublishTaskService.MODULE_TYPE -> {
-                publishTaskService.touchHeartbeat(taskId, request);
-            }
-            case MODULE_PRICE_TRACK -> {
-                priceTrackTaskCacheService.touchTaskHeartbeat(taskId);
-            }
-            case MODULE_SHOP_MATCH -> {
-                shopMatchTaskCacheService.touchTaskHeartbeat(taskId);
-            }
-            case MODULE_PATROL_DELETE -> {
-                patrolDeleteTaskCacheService.touchTaskHeartbeat(taskId);
-            }
-            case MODULE_QUERY_ASIN -> {
-                queryAsinTaskCacheService.touchTaskHeartbeat(taskId);
-            }
-            case MODULE_SHOP_DATA_CRAWL -> shopDataCrawlTaskCacheService.touchTaskHeartbeat(taskId);
-            case MODULE_WITHDRAW -> {
-                withdrawTaskCacheService.touchTaskHeartbeat(taskId);
-            }
-            case MODULE_APPEARANCE_PATENT -> {
-                appearancePatentTaskCacheService.touchTaskHeartbeat(taskId);
-            }
-            case MODULE_SIMILAR_ASIN -> {
-                similarAsinTaskCacheService.touchTaskHeartbeat(taskId);
-            }
-            case MODULE_DELETE_BRAND -> {
-                deleteBrandTaskCacheService.saveProgress(taskId, buildDeleteBrandHeartbeatProgress(request), true);
-            }
-            case MODULE_COLLECT_DATA -> collectDataService.updateProgress(taskId, request);
-            default -> {
-            }
-        }
-    }
-
-    private Map<String, String> buildDeleteBrandHeartbeatProgress(TaskHeartbeatRequest request) {
-        String now = String.valueOf(System.currentTimeMillis());
-        Map<String, String> values = new LinkedHashMap<>();
-        values.put("last_heartbeat_at", now);
-        values.put("updated_at", now);
-        if (request != null) {
-            putIfPresent(values, "phase", request.getPhase());
-            putIfPresent(values, "current", request.getCurrent());
-            putIfPresent(values, "total", request.getTotal());
-        }
-        return values;
-    }
-
-    private void saveFileTaskCache(String moduleType, FileTaskEntity task) {
-        switch (moduleType) {
-            case MODULE_PRODUCT_RISK -> productRiskTaskCacheService.saveTaskCache(task);
-            case MODULE_PRICE_TRACK -> priceTrackTaskCacheService.saveTaskCache(task);
-            case MODULE_SHOP_MATCH -> shopMatchTaskCacheService.saveTaskCache(task);
-            case MODULE_PATROL_DELETE -> patrolDeleteTaskCacheService.saveTaskCache(task);
-            case MODULE_QUERY_ASIN -> queryAsinTaskCacheService.saveTaskCache(task);
-            case MODULE_SHOP_DATA_CRAWL -> shopDataCrawlTaskCacheService.saveTaskCache(task);
-            case MODULE_WITHDRAW -> withdrawTaskCacheService.saveTaskCache(task);
-            case MODULE_DELETE_BRAND -> deleteBrandTaskCacheService.saveTaskCache(task);
-            default -> {
-            }
+    /** 无节流间隔时永远认为需要刷新（历史行为）；有间隔时按 updated_at 距上次刷新的时长判断。 */
+    private boolean isCheckpointDue(FileTaskEntity task, Duration interval) {
+        if (interval == null) {
+            return true;
         }
+        long intervalMillis = Math.max(1_000L, interval.toMillis());
+        LocalDateTime cutoff = LocalDateTime.now().minus(Duration.ofMillis(intervalMillis));
+        return task.getUpdatedAt() == null || !task.getUpdatedAt().isAfter(cutoff);
     }
 
     /** 把终态同步到模块缓存与进度快照；失败只告警，不影响中断接口本身的成功语义。 */
@@ -372,10 +227,13 @@ public class TaskHeartbeatService {
         }
     }
 
-    private void putIfPresent(Map<String, String> values, String key, Object value) {
-        if (value != null) {
-            values.put(key, String.valueOf(value));
+    private void saveFileTaskCache(String moduleType, FileTaskEntity task) {
+        if (moduleType == null) {
+            return;
+        }
+        TaskModuleHeartbeatSpi handler = moduleHandlers.get(moduleType);
+        if (handler != null) {
+            handler.saveTaskCache(task);
         }
     }
-
 }

backend-java/src/main/java/com/nanri/aiimage/modules/task/service/TransientPayloadDeleteOrchestrator.java

@@ -2,6 +2,7 @@ package com.nanri.aiimage.modules.task.service;
 
 import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.fasterxml.jackson.databind.ObjectMapper;
+import com.nanri.aiimage.common.util.PendingDeleteJournal;
 import com.nanri.aiimage.modules.file.service.object.RustfsObjectStorageService;
 import com.nanri.aiimage.modules.task.mapper.TaskChunkMapper;
 import com.nanri.aiimage.modules.task.mapper.TaskScopeStateMapper;
@@ -10,8 +11,11 @@ import com.nanri.aiimage.modules.task.model.entity.TaskScopeStateEntity;
 import lombok.extern.slf4j.Slf4j;
 import org.springframework.beans.factory.annotation.Qualifier;
 import org.springframework.beans.factory.annotation.Value;
+import org.springframework.boot.context.event.ApplicationReadyEvent;
+import org.springframework.context.event.EventListener;
 import org.springframework.stereotype.Service;
 
+import java.nio.file.Path;
 import java.util.ArrayList;
 import java.util.LinkedHashMap;
 import java.util.LinkedHashSet;
@@ -49,6 +53,12 @@ public class TransientPayloadDeleteOrchestrator {
     @Value("${aiimage.transient-storage.max-pending-deletes:1000}")
     private long maxPendingDeletes = 1000;
 
+    /**
+     * 待删对象的本地落盘日志：pending 集合重启即丢，未被引用的对象会一直残留在桶里。
+     * 入队追加、flush 后按「仍在 pending + 异步删除失败」收敛重写、启动时回放后重新走引用检查。
+     */
+    private volatile PendingDeleteJournal journal;
+
     public TransientPayloadDeleteOrchestrator(TransientPayloadStorageService transientPayloadStorageService,
                                               RustfsObjectStorageService rustfsObjectStorageService,
                                               TaskChunkMapper taskChunkMapper,
@@ -84,6 +94,7 @@ public class TransientPayloadDeleteOrchestrator {
                 }
                 if (pendingPointers.add(pointer)) {
                     accepted++;
+                    journal().record(pointer);
                 }
             }
             return accepted;
@@ -164,6 +175,7 @@ public class TransientPayloadDeleteOrchestrator {
     }
 
     private void deleteObjects(List<String> pointers) {
+        List<String> failed = new ArrayList<>();
         for (String pointer : pointers) {
             if (pointer == null || !pointer.startsWith(RUSTFS_POINTER_PREFIX)) {
                 // 本地/OSS 指针有实例归属与其它清理路径，不在此批量删除范围。
@@ -173,10 +185,50 @@ public class TransientPayloadDeleteOrchestrator {
             try {
                 rustfsObjectStorageService.deleteObject(objectKey);
             } catch (Exception ex) {
+                // 失败对象写回日志：进程重启后由启动回放重新提交，不再随进程消失
+                failed.add(pointer);
                 log.warn("[transient-payload] async delete failed objectKey={} err={}",
                         objectKey, ex.getMessage());
             }
         }
+        synchronized (pendingPointers) {
+            Set<String> keep = new LinkedHashSet<>(pendingPointers);
+            keep.addAll(failed);
+            journal().rewrite(keep);
+        }
+        if (!failed.isEmpty()) {
+            log.warn("[transient-payload] async delete failed, kept in journal count={}", failed.size());
+        }
+    }
+
+    /**
+     * 启动回放：把上次进程残留的待删对象重新入队，等待下一轮 flush 做引用检查与物理删除
+     * （引用检查幂等，重复提交同一对象只入队一次）。
+     */
+    @EventListener(ApplicationReadyEvent.class)
+    public void replayJournalOnStartup() {
+        List<String> pointers = journal().readAll();
+        if (pointers.isEmpty()) {
+            return;
+        }
+        int accepted = submitDeletes(pointers);
+        log.info("[transient-payload] 启动回放待删对象 journalSize={} accepted={} pending={}",
+                pointers.size(), accepted, pendingCount());
+    }
+
+    private PendingDeleteJournal journal() {
+        PendingDeleteJournal local = journal;
+        if (local != null) {
+            return local;
+        }
+        synchronized (pendingPointers) {
+            if (journal == null) {
+                Path path = Path.of(System.getProperty("java.io.tmpdir"),
+                        "aiimage-delete-journal", "transient-payload-pending.log");
+                journal = new PendingDeleteJournal(path, pendingLimit());
+            }
+            return journal;
+        }
     }
 
     /** 批量 IN 反查两张引用表，返回仍被引用的指针集合（查询异常抛给调用方）。 */

backend-java/src/main/java/com/nanri/aiimage/modules/task/spi/BrandTaskHeartbeatSpi.java

@@ -0,0 +1,25 @@
+package com.nanri.aiimage.modules.task.spi;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.vo.TaskHeartbeatVo;
+
+/**
+ * 品牌检测任务（brand_crawl_tasks，独立于 biz_file_task）的心跳/中断扩展点。
+ *
+ * <p>2026-09 全维度审查 G5：品牌任务有自己的表与进度缓存，原先由 {@code TaskHeartbeatService}
+ * 直接注入 brand 模块的 Mapper 与缓存服务，构成 task → 业务的依赖。这里把品牌侧逻辑整体
+ * 交回 brand 模块实现，task 侧只按接口调用。
+ */
+public interface BrandTaskHeartbeatSpi {
+
+    /**
+     * 品牌任务心跳：任务不存在返回 null；非 running 返回 notAlive；running 则条件更新
+     * updated_at 并刷新进度缓存。
+     */
+    TaskHeartbeatVo heartbeat(Long taskId, TaskHeartbeatRequest request);
+
+    /**
+     * 客户端重启中断：running/pending → cancelled（条件更新），返回心跳结果；任务不存在返回 null。
+     */
+    TaskHeartbeatVo markInterrupted(Long taskId, String reason);
+}

backend-java/src/main/java/com/nanri/aiimage/modules/task/spi/BrandTaskStaleRepairSpi.java

@@ -0,0 +1,22 @@
+package com.nanri.aiimage.modules.task.spi;
+
+import java.time.LocalDateTime;
+import java.util.List;
+
+/**
+ * 品牌检测任务的陈旧修复扩展点（2026-09 全维度审查 G5）。
+ *
+ * <p>brand_crawl_tasks 有独立的表与状态字面量（pending/running/failed），原先由
+ * {@code StaleTaskRepairService} 直接操作品牌模块的 Mapper，构成 task → 业务的依赖。
+ * 现由品牌模块实现本接口，task 侧只按接口调用。
+ */
+public interface BrandTaskStaleRepairSpi {
+
+    /**
+     * 把 pending/running 且 updated_at 早于 cutoff 的品牌任务置为失败（条件更新，幂等）。
+     *
+     * @param limit 单轮处理上限
+     * @return 实际标记失败的任务 id（空表示本轮无陈旧任务）
+     */
+    List<Long> failStaleBrandTasks(LocalDateTime cutoff, int limit);
+}

backend-java/src/main/java/com/nanri/aiimage/modules/task/spi/CollectDataItemCleanupSpi.java

@@ -0,0 +1,20 @@
+package com.nanri.aiimage.modules.task.spi;
+
+import java.util.List;
+
+/**
+ * 采集明细行的历史清理扩展点（2026-09 全维度审查 G5）。
+ *
+ * <p>任务历史清理时除了 task 自己的表，还要连带删除 biz_collect_data_item 的明细行；
+ * 该表属于采集模块，原先由 {@code ModuleHistoryCleanupService} 直接使用其 Mapper，
+ * 构成 task → 业务的依赖。现由采集模块实现本接口。
+ */
+public interface CollectDataItemCleanupSpi {
+
+    /**
+     * 删除这些任务名下的采集明细行。
+     *
+     * @return 删除行数
+     */
+    int deleteItemsByTaskIds(List<Long> taskIds);
+}

backend-java/src/main/java/com/nanri/aiimage/modules/task/spi/TaskModuleHeartbeatSpi.java

@@ -0,0 +1,57 @@
+package com.nanri.aiimage.modules.task.spi;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+
+import java.time.Duration;
+
+/**
+ * 任务心跳的模块侧扩展点（2026-09 全维度审查 G5：消除 task ↔ 业务模块的双向依赖）。
+ *
+ * <p>背景：{@code TaskHeartbeatService} 原先直接注入 12 个业务模块的 CacheService 并用 switch 分发，
+ * 与"业务模块实现 task 的 Handler SPI"形成真实循环依赖。现改为：task 侧只依赖本接口，
+ * 各业务模块各自实现一个 Bean（Spring 注入 List 后在 task 侧建索引）。
+ *
+ * <p>新增模块时：实现本接口并在注册表中登记模块类型；心跳契约测试会校验覆盖。
+ */
+public interface TaskModuleHeartbeatSpi {
+
+    /** 本实现负责的 moduleType（与 biz_file_task.module_type 一致）。 */
+    String moduleType();
+
+    /**
+     * 刷新模块自己的进度缓存/心跳（非 DB 写入）。
+     * 多数模块实现为 {@code cacheService.touchTaskHeartbeat(taskId)}。
+     */
+    void touchHeartbeat(Long taskId, TaskHeartbeatRequest request);
+
+    /**
+     * 心跳时回写模块的任务缓存（部分模块没有独立缓存，保持默认空实现）。
+     */
+    default void saveTaskCache(FileTaskEntity task) {
+    }
+
+    /**
+     * 心跳前的任务归属校验：publish / shop-data-crawl 在非本实例拥有时拒绝心跳，
+     * 其余模块默认放行。
+     */
+    default void ensureOwnership(FileTaskEntity task, String operation) {
+    }
+
+    /**
+     * 心跳时的额外模块侧动作（如 similar-asin 刷新结果文件组装作业的租约）。
+     */
+    default void onHeartbeat(FileTaskEntity task) {
+    }
+
+    /**
+     * biz_file_task.updated_at 检查点节流间隔。
+     *
+     * <p>返回非空时，DB 的 updated_at 只在距上次刷新超过该间隔后才更新
+     * （客户端心跳间隔远小于 stale 判定阈值时避免每次心跳都写库）；
+     * 默认 null 表示每次心跳都刷新（与历史行为一致）。
+     */
+    default Duration checkpointInterval() {
+        return null;
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/withdraw/controller/WithdrawTaskController.java

@@ -3,9 +3,9 @@ package com.nanri.aiimage.modules.withdraw.controller;
 import com.nanri.aiimage.common.api.ApiResponse;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCandidateAddRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
 import com.nanri.aiimage.modules.withdraw.model.dto.WithdrawCandidateClearRequest;
 import com.nanri.aiimage.modules.withdraw.model.dto.WithdrawCreateTaskRequest;
 import com.nanri.aiimage.modules.withdraw.model.dto.WithdrawSubmitResultRequest;

backend-java/src/main/java/com/nanri/aiimage/modules/withdraw/service/WithdrawResolveService.java

@@ -4,9 +4,9 @@ import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.nanri.aiimage.common.exception.BusinessException;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskCandidateAddRequest;
 import com.nanri.aiimage.modules.productrisk.model.dto.ProductRiskMatchShopsRequest;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskCandidateVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskMatchShopsVo;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskShopQueueItemVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskCandidateVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskMatchShopsVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskShopQueueItemVo;
 import com.nanri.aiimage.modules.withdraw.mapper.WithdrawShopCandidateMapper;
 import com.nanri.aiimage.modules.withdraw.model.entity.WithdrawShopCandidateEntity;
 import com.nanri.aiimage.modules.ziniao.model.vo.ZiniaoShopMatchResultVo;

backend-java/src/main/java/com/nanri/aiimage/modules/withdraw/service/WithdrawTaskHeartbeatSpi.java

@@ -0,0 +1,35 @@
+package com.nanri.aiimage.modules.withdraw.service;
+
+import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+
+/**
+ * 取款 的任务心跳实现（2026-09 全维度审查 G5）。
+ *
+ * <p>心跳逻辑（缓存刷新 / 任务缓存回写）从 task 模块收回本模块，task 侧只依赖 SPI 接口，
+ * 消除 task → 业务模块的编译期依赖。
+ */
+@Service
+@RequiredArgsConstructor
+public class WithdrawTaskHeartbeatSpi implements TaskModuleHeartbeatSpi {
+
+    private final WithdrawTaskCacheService cacheService;
+
+    @Override
+    public String moduleType() {
+        return "WITHDRAW";
+    }
+
+    @Override
+    public void touchHeartbeat(Long taskId, TaskHeartbeatRequest request) {
+        cacheService.touchTaskHeartbeat(taskId);
+    }
+
+    @Override
+    public void saveTaskCache(FileTaskEntity task) {
+        cacheService.saveTaskCache(task);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/withdraw/service/WithdrawTaskService.java

@@ -10,7 +10,7 @@ import com.nanri.aiimage.common.exception.BusinessCodes;
 import com.nanri.aiimage.common.exception.BusinessException;
 import com.nanri.aiimage.config.TaskPressureProperties;
 import com.nanri.aiimage.modules.file.service.oss.OssStorageService;
-import com.nanri.aiimage.modules.productrisk.model.vo.ProductRiskDashboardVo;
+import com.nanri.aiimage.common.model.vo.ProductRiskDashboardVo;
 import com.nanri.aiimage.modules.task.mapper.FileResultMapper;
 import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
 import com.nanri.aiimage.modules.task.model.entity.FileResultEntity;

backend-java/src/main/resources/db/V123__drop_prefix_indexes.sql

@@ -0,0 +1,46 @@
+-- V123: 删除 biz_file_result 上被复合索引完全覆盖的两个单列索引
+--
+-- 判定原则（同 V121）：被删索引的列是另一索引的**严格前缀**，
+-- 所有能用它的查询都能被保留的复合索引覆盖，删除只减少写入侧的索引维护与 buffer pool 占用。
+--
+-- 1) biz_file_result.idx_task_id (task_id)
+--    被 V38 的 idx_file_result_task_module_id (task_id, module_type, id) 覆盖。
+-- 2) biz_file_result.idx_module_type (module_type)
+--    被 V34 的 idx_biz_file_result_module_user_created (module_type, user_id, created_at)
+--    与 V42 的 idx_file_result_patrol_history (module_type, user_id, created_at, id) 覆盖。
+--    两处均为 V1 建表时的遗留索引（当时还没有任何复合索引）。
+--
+-- 风险：DROP INDEX 为 INPLACE 在线操作，但会立即改变执行计划；两者均为「被覆盖」判定。
+-- 回滚：
+--   ALTER TABLE biz_file_result ADD INDEX idx_task_id (task_id);
+--   ALTER TABLE biz_file_result ADD INDEX idx_module_type (module_type);
+
+SET @db_name = DATABASE();
+
+-- 1) biz_file_result.idx_task_id
+SET @idx_exists := (
+  SELECT COUNT(*) FROM information_schema.STATISTICS
+  WHERE TABLE_SCHEMA = @db_name AND TABLE_NAME = 'biz_file_result'
+    AND INDEX_NAME = 'idx_task_id'
+);
+SET @sql := IF(@idx_exists > 0,
+  'ALTER TABLE biz_file_result DROP INDEX idx_task_id',
+  'SELECT 1'
+);
+PREPARE stmt FROM @sql;
+EXECUTE stmt;
+DEALLOCATE PREPARE stmt;
+
+-- 2) biz_file_result.idx_module_type
+SET @idx_exists := (
+  SELECT COUNT(*) FROM information_schema.STATISTICS
+  WHERE TABLE_SCHEMA = @db_name AND TABLE_NAME = 'biz_file_result'
+    AND INDEX_NAME = 'idx_module_type'
+);
+SET @sql := IF(@idx_exists > 0,
+  'ALTER TABLE biz_file_result DROP INDEX idx_module_type',
+  'SELECT 1'
+);
+PREPARE stmt FROM @sql;
+EXECUTE stmt;
+DEALLOCATE PREPARE stmt;

backend-java/src/main/resources/db/V124__drop_dedupe_total_data_country_index.sql

@@ -0,0 +1,27 @@
+-- V124: 删除 biz_dedupe_total_data.idx_country（函数列索引，永不生效）
+--
+-- 背景（2026-09 全维度审查 C6）：该表按国家的筛选走 FIND_IN_SET(?, IFNULL(country,''))，
+-- 是函数列比较，MySQL 无法使用 country 上的普通索引；V91 建的 idx_country 从未被任何执行计划
+-- 使用过，只增加写入维护与 buffer pool 占用。
+--
+-- 前缀关键字搜索本身就吃 uk_data_value (data_value) 唯一索引（本次同时把关键字由
+-- %kw% 前置通配改为前缀匹配，见 DedupeTotalDataService.page），因此无需新增索引。
+--
+-- 风险：DROP INDEX 为在线 INPLACE 操作；country 仍用于结果展示，仅去掉索引。
+-- 回滚：
+--   ALTER TABLE biz_dedupe_total_data ADD INDEX idx_country (country);
+
+SET @db_name = DATABASE();
+
+SET @idx_exists := (
+  SELECT COUNT(*) FROM information_schema.STATISTICS
+  WHERE TABLE_SCHEMA = @db_name AND TABLE_NAME = 'biz_dedupe_total_data'
+    AND INDEX_NAME = 'idx_country'
+);
+SET @sql := IF(@idx_exists > 0,
+  'ALTER TABLE biz_dedupe_total_data DROP INDEX idx_country',
+  'SELECT 1'
+);
+PREPARE stmt FROM @sql;
+EXECUTE stmt;
+DEALLOCATE PREPARE stmt;

backend-java/src/test/java/com/nanri/aiimage/ArchitectureBoundaryTest.java

@@ -34,7 +34,9 @@ class ArchitectureBoundaryTest {
      * 本次以实测值 119 为准恢复告警能力；经 git diff 核验，当日的 task 模块改动
      * （引用计数按 task 收敛、快照 upsert 缓存、心跳条件更新、锁续期重试等）未新增任何业务依赖。
      */
-    private static final int TASK_TO_BUSINESS_BASELINE = 119;
+    // 2026-09-14 心跳/陈旧修复/历史清理 SPI 化后实测 6（G5：task 只依赖 task.spi 下接口，
+// 业务模块各自实现）；继续按棘轮收敛，不允许回升
+    private static final int TASK_TO_BUSINESS_BASELINE = 6;
 
     private static volatile JavaClasses cached;
 

backend-java/src/test/java/com/nanri/aiimage/common/util/PendingDeleteJournalTest.java

@@ -0,0 +1,90 @@
+package com.nanri.aiimage.common.util;
+
+import org.junit.jupiter.api.Test;
+import org.junit.jupiter.api.io.TempDir;
+
+import java.nio.charset.StandardCharsets;
+import java.nio.file.Files;
+import java.nio.file.Path;
+import java.util.List;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertFalse;
+import static org.junit.jupiter.api.Assertions.assertTrue;
+
+/**
+ * 待删对象日志：入队追加、启动回放、收敛重写。对应审查项 D13（删除补偿队列重启即丢）。
+ */
+class PendingDeleteJournalTest {
+
+    @TempDir
+    Path tempDir;
+
+    @Test
+    void recordThenReadAllReturnsAdmittedKeysInOrder() {
+        PendingDeleteJournal journal = new PendingDeleteJournal(tempDir.resolve("nested/queue.log"), 100);
+
+        journal.record("a/1.bin");
+        journal.record("b/2.bin");
+        journal.record("a/1.bin");
+
+        assertEquals(List.of("a/1.bin", "b/2.bin"), journal.readAll());
+    }
+
+    @Test
+    void rewriteKeepsOnlyRemainingAndDropsFileWhenEmpty() throws Exception {
+        Path path = tempDir.resolve("queue.log");
+        PendingDeleteJournal journal = new PendingDeleteJournal(path, 100);
+        journal.record("keep.bin");
+        journal.record("done.bin");
+
+        journal.rewrite(List.of("keep.bin"));
+        assertEquals(List.of("keep.bin"), journal.readAll());
+
+        journal.rewrite(List.of());
+        assertFalse(Files.exists(path), "队列收敛为空后日志文件应删除");
+        assertEquals(List.of(), journal.readAll());
+    }
+
+    @Test
+    void missingFileYieldsEmptyInsteadOfThrowing() {
+        PendingDeleteJournal journal = new PendingDeleteJournal(tempDir.resolve("not-created.log"), 10);
+        assertEquals(List.of(), journal.readAll());
+    }
+
+    @Test
+    void blankAndMultilineRecordsAreSanitizedOrIgnored() throws Exception {
+        Path path = tempDir.resolve("queue.log");
+        PendingDeleteJournal journal = new PendingDeleteJournal(path, 10);
+
+        journal.record(null);
+        journal.record("   ");
+        journal.record("bad\nkey.bin");
+        journal.record(" good.bin ");
+
+        assertEquals(List.of("bad key.bin", "good.bin"), journal.readAll());
+        // 单行一条记录：换行被替换，不会破坏行结构
+        assertEquals(2, Files.readAllLines(path, StandardCharsets.UTF_8).size());
+    }
+
+    @Test
+    void corruptedLineIsIgnoredWhileValidOnesSurvive() throws Exception {
+        Path path = tempDir.resolve("queue.log");
+        Files.createDirectories(tempDir);
+        Files.writeString(path, "\n  \nvalid.bin\n", StandardCharsets.UTF_8);
+
+        PendingDeleteJournal journal = new PendingDeleteJournal(path, 10);
+
+        assertEquals(List.of("valid.bin"), journal.readAll());
+    }
+
+    @Test
+    void rewriteRespectsMaxEntriesCap() {
+        PendingDeleteJournal journal = new PendingDeleteJournal(tempDir.resolve("queue.log"), 2);
+
+        journal.rewrite(List.of("k1.bin", "k2.bin", "k3.bin"));
+
+        assertEquals(List.of("k1.bin", "k2.bin"), journal.readAll());
+        assertTrue(journal.readAll().size() <= 2);
+    }
+}

backend-java/src/test/java/com/nanri/aiimage/modules/deletebrand/service/TaskModuleCoverageTest.java

@@ -0,0 +1,112 @@
+package com.nanri.aiimage.modules.deletebrand.service;
+
+import com.nanri.aiimage.common.module.TaskModuleRegistry;
+import com.nanri.aiimage.modules.appearancepatent.service.AppearancePatentTaskService;
+import com.nanri.aiimage.modules.brand.service.BrandTaskService;
+import com.nanri.aiimage.modules.shopdatacrawl.service.ShopDataCrawlTaskService;
+import com.nanri.aiimage.modules.similarasin.service.SimilarAsinTaskService;
+import com.nanri.aiimage.config.ModuleCleanupProperties;
+import com.nanri.aiimage.config.TaskFileJobConfig;
+import org.junit.jupiter.api.Test;
+
+import java.util.List;
+import java.util.Map;
+import java.util.Set;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertNotNull;
+import static org.junit.jupiter.api.Assertions.assertTrue;
+import static org.mockito.Mockito.mock;
+
+/**
+ * 模块清单覆盖面自检（2026-09 全维度审查 G6）。
+ *
+ * 新增工具模块此前要同时改 5 处枚举，其中 4 处没有自检，漏改表现为"功能静默不生效"。
+ * 本测试把「注册表 = 单一来源」这件事钉住：数据驱动的三处必须与注册表完全一致，
+ * 委派式判死巡检线必须为注册表标记的每个模块登记动作。
+ * 新增模块时若漏改任一处，这里会红。
+ */
+class TaskModuleCoverageTest {
+
+    /** 当前线上模块清单（新增模块时同时更新：注册表 + 本清单 + 对应 Handler） */
+    private static final Set<String> EXPECTED_MODULES = Set.of(
+            "PUBLISH", "DEDUPE", "SPLIT", "CONVERT", "DELETE_BRAND", "PRODUCT_RISK_RESOLVE",
+            "PRICE_TRACK", "SHOP_MATCH", "PATROL_DELETE", "QUERY_ASIN", "WITHDRAW",
+            "APPEARANCE_PATENT", "SIMILAR_ASIN", "COLLECT_DATA", "SHOP_DATA_CRAWL", "BRAND");
+
+    @Test
+    void registryCoversEveryKnownModuleWithLabel() {
+        assertEquals(EXPECTED_MODULES, TaskModuleRegistry.moduleTypes());
+
+        Map<String, String> labels = TaskModuleRegistry.labels();
+        assertEquals(EXPECTED_MODULES.size(), labels.size());
+        for (String moduleType : EXPECTED_MODULES) {
+            String label = labels.get(moduleType);
+            assertNotNull(label, "模块缺少中文名: " + moduleType);
+            assertTrue(!label.isBlank(), "模块中文名不能为空: " + moduleType);
+        }
+        assertEquals("跟价", TaskModuleRegistry.labelOf("PRICE_TRACK"));
+        assertEquals("UNKNOWN_MODULE", TaskModuleRegistry.labelOf("UNKNOWN_MODULE"), "未知模块回退原值");
+    }
+
+    @Test
+    void resultFileJobWhitelistMatchesRegistry() {
+        assertEquals(TaskModuleRegistry.resultFileJobModuleTypes(), TaskFileJobConfig.RESULT_FILE_JOB_MODULE_TYPES);
+        // 纯表格处理模块不产出结果文件作业
+        assertTrue(!TaskFileJobConfig.RESULT_FILE_JOB_MODULE_TYPES.contains("DEDUPE"));
+        assertTrue(!TaskFileJobConfig.RESULT_FILE_JOB_MODULE_TYPES.contains("SPLIT"));
+        assertTrue(!TaskFileJobConfig.RESULT_FILE_JOB_MODULE_TYPES.contains("CONVERT"));
+        assertTrue(TaskFileJobConfig.RESULT_FILE_JOB_MODULE_TYPES.contains("SHOP_DATA_CRAWL"));
+    }
+
+    @Test
+    void ageCleanupListMatchesRegistry() {
+        ModuleCleanupProperties properties = new ModuleCleanupProperties();
+
+        assertEquals(TaskModuleRegistry.ageCleanupModuleTypes(), Set.copyOf(properties.getModuleTypes()));
+        // 店铺数据采集靠每日工作簿自身保留期，不参与按天清理
+        assertTrue(!properties.getModuleTypes().contains("SHOP_DATA_CRAWL"));
+        assertTrue(properties.getModuleTypes().contains("DEDUPE"));
+    }
+
+    /**
+     * 只装配 delegatedStaleChecks 需要的四个业务服务，其余依赖留空——
+     * 这样测试聚焦"巡检线登记了哪些模块"，不牵扯任务锁/缓存等无关依赖。
+     */
+    private static DeleteBrandStaleTaskService serviceForDelegatedChecks() {
+        return new DeleteBrandStaleTaskService(
+                null, null, null, null, null, null, null, null, null, null, null, null, null, null, null,
+                null,
+                mock(BrandTaskService.class),
+                mock(AppearancePatentTaskService.class),
+                mock(SimilarAsinTaskService.class),
+                null, null, null, null,
+                mock(ShopDataCrawlTaskService.class));
+    }
+
+    @Test
+    void delegatedStaleCheckHandlersCoverRegistryFlag() {
+        DeleteBrandStaleTaskService service = serviceForDelegatedChecks();
+
+        Map<String, Runnable> handlers = service.delegatedStaleChecks();
+
+        assertEquals(TaskModuleRegistry.delegatedStaleCheckModuleTypes(), handlers.keySet(),
+                "注册表标记了 delegatedStaleCheck 的模块必须都在巡检线登记动作（反之亦然）");
+        for (Runnable action : handlers.values()) {
+            assertNotNull(action);
+        }
+    }
+
+    @Test
+    void delegatedStaleCheckIncludesModulesThatHadOwnScheduler() {
+        DeleteBrandStaleTaskService service = serviceForDelegatedChecks();
+
+        List<String> keys = List.copyOf(service.delegatedStaleChecks().keySet());
+
+        // 商品管理采集 2026-09 从自带 @Scheduled 并入巡检线（P1-8 双实例判死盲区），不得回退
+        assertTrue(keys.contains("SHOP_DATA_CRAWL"), "商品管理采集必须留在 stale-check 巡检线内: " + keys);
+        assertTrue(keys.contains("SIMILAR_ASIN"));
+        assertTrue(keys.contains("APPEARANCE_PATENT"));
+        assertTrue(keys.contains("BRAND"));
+    }
+}

backend-java/src/test/java/com/nanri/aiimage/modules/shopdatacrawl/service/ShopDataCrawlItemRetentionServiceTest.java

@@ -0,0 +1,84 @@
+package com.nanri.aiimage.modules.shopdatacrawl.service;
+
+import com.nanri.aiimage.common.service.DistributedJobLockService;
+import com.nanri.aiimage.modules.shopdatacrawl.mapper.ShopDataCrawlItemMapper;
+import org.junit.jupiter.api.Test;
+
+import java.time.LocalDate;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.mockito.ArgumentMatchers.any;
+import static org.mockito.ArgumentMatchers.anyInt;
+import static org.mockito.ArgumentMatchers.anyString;
+import static org.mockito.ArgumentMatchers.eq;
+import static org.mockito.Mockito.mock;
+import static org.mockito.Mockito.never;
+import static org.mockito.Mockito.times;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.when;
+
+/**
+ * 采集明细表保留期清理（2026-09 全维度审查：该表此前无任何清理策略）。
+ */
+class ShopDataCrawlItemRetentionServiceTest {
+
+    private final ShopDataCrawlItemMapper itemMapper = mock(ShopDataCrawlItemMapper.class);
+    private final DistributedJobLockService jobLockService = mock(DistributedJobLockService.class);
+
+    private ShopDataCrawlItemRetentionService service() {
+        return new ShopDataCrawlItemRetentionService(itemMapper, jobLockService);
+    }
+
+    private void lockAvailable() {
+        when(jobLockService.tryLock(anyString(), any()))
+                .thenReturn(mock(DistributedJobLockService.LockHandle.class));
+    }
+
+    @Test
+    void deletesInBatchesUntilBatchNotFull() {
+        lockAvailable();
+        when(itemMapper.deleteOlderThanBatch(any(LocalDate.class), anyInt()))
+                .thenReturn(5000, 5000, 120);
+
+        service().purgeExpiredItems();
+
+        verify(itemMapper, times(3)).deleteOlderThanBatch(any(LocalDate.class), eq(5000));
+    }
+
+    @Test
+    void stopsAtBatchLimitPerRun() {
+        lockAvailable();
+        when(itemMapper.deleteOlderThanBatch(any(LocalDate.class), anyInt())).thenReturn(5000);
+
+        service().purgeExpiredItems();
+
+        verify(itemMapper, times(ShopDataCrawlItemRetentionService.MAX_BATCHES_PER_RUN))
+                .deleteOlderThanBatch(any(LocalDate.class), anyInt());
+    }
+
+    @Test
+    void skipsWhenAnotherInstanceHoldsLock() {
+        when(jobLockService.tryLock(anyString(), any())).thenReturn(null);
+
+        service().purgeExpiredItems();
+
+        verify(itemMapper, never()).deleteOlderThanBatch(any(LocalDate.class), anyInt());
+    }
+
+    @Test
+    void deleteFailureIsLoggedAndDoesNotThrow() {
+        lockAvailable();
+        when(itemMapper.deleteOlderThanBatch(any(LocalDate.class), anyInt()))
+                .thenThrow(new IllegalStateException("db down"));
+
+        service().purgeExpiredItems();
+
+        verify(itemMapper).deleteOlderThanBatch(any(LocalDate.class), anyInt());
+    }
+
+    @Test
+    void retentionDaysHasLowerBound() {
+        ShopDataCrawlItemRetentionService service = service();
+        assertEquals(30, service.effectiveRetentionDays(), "默认保留 30 天");
+    }
+}

backend-java/src/test/java/com/nanri/aiimage/modules/task/service/ModuleHistoryCleanupKeysetPaginationTest.java

@@ -9,7 +9,7 @@ import com.nanri.aiimage.config.InstanceMetadata;
 import com.nanri.aiimage.config.ModuleCleanupProperties;
 import com.nanri.aiimage.config.StorageProperties;
 import com.nanri.aiimage.config.TransientStorageProperties;
-import com.nanri.aiimage.modules.collectdata.mapper.CollectDataItemMapper;
+import com.nanri.aiimage.modules.task.spi.CollectDataItemCleanupSpi;
 import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataItemEntity;
 import com.nanri.aiimage.modules.file.service.object.RustfsObjectStorageService;
 import com.nanri.aiimage.modules.file.service.oss.OssStorageService;
@@ -90,7 +90,7 @@ class ModuleHistoryCleanupKeysetPaginationTest {
     private TaskResultPayloadMapper taskResultPayloadMapper;
     private TaskScopeStateMapper taskScopeStateMapper;
     private TaskChunkMapper taskChunkMapper;
-    private CollectDataItemMapper collectDataItemMapper;
+    private CollectDataItemCleanupSpi collectDataItemMapper;
     private DistributedJobLockService lockService;
     private TransientPayloadStorageService storage;
     private RustfsObjectStorageService rustfs;
@@ -129,7 +129,7 @@ class ModuleHistoryCleanupKeysetPaginationTest {
         taskResultPayloadMapper = mock(TaskResultPayloadMapper.class);
         taskScopeStateMapper = mock(TaskScopeStateMapper.class);
         taskChunkMapper = mock(TaskChunkMapper.class);
-        collectDataItemMapper = mock(CollectDataItemMapper.class);
+        collectDataItemMapper = mock(CollectDataItemCleanupSpi.class);
         StringRedisTemplate redisTemplate = mock(StringRedisTemplate.class);
         ValueOperations<String, String> valueOperations = mock(ValueOperations.class);
         when(redisTemplate.opsForValue()).thenReturn(valueOperations);

backend-java/src/test/java/com/nanri/aiimage/modules/task/service/ModuleHistoryCleanupLoggingTest.java

@@ -13,7 +13,7 @@ import com.nanri.aiimage.config.InstanceMetadata;
 import com.nanri.aiimage.config.ModuleCleanupProperties;
 import com.nanri.aiimage.config.StorageProperties;
 import com.nanri.aiimage.config.TransientStorageProperties;
-import com.nanri.aiimage.modules.collectdata.mapper.CollectDataItemMapper;
+import com.nanri.aiimage.modules.task.spi.CollectDataItemCleanupSpi;
 import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataItemEntity;
 import com.nanri.aiimage.modules.file.service.object.RustfsObjectStorageService;
 import com.nanri.aiimage.modules.file.service.oss.OssStorageService;
@@ -92,7 +92,7 @@ class ModuleHistoryCleanupLoggingTest {
     private TaskResultPayloadMapper taskResultPayloadMapper;
     private TaskScopeStateMapper taskScopeStateMapper;
     private TaskChunkMapper taskChunkMapper;
-    private CollectDataItemMapper collectDataItemMapper;
+    private CollectDataItemCleanupSpi collectDataItemMapper;
     private DistributedJobLockService lockService;
     private TransientPayloadStorageService storage;
     private RustfsObjectStorageService rustfs;
@@ -128,7 +128,7 @@ class ModuleHistoryCleanupLoggingTest {
         taskResultPayloadMapper = mock(TaskResultPayloadMapper.class);
         taskScopeStateMapper = mock(TaskScopeStateMapper.class);
         taskChunkMapper = mock(TaskChunkMapper.class);
-        collectDataItemMapper = mock(CollectDataItemMapper.class);
+        collectDataItemMapper = mock(CollectDataItemCleanupSpi.class);
         StringRedisTemplate redisTemplate = mock(StringRedisTemplate.class);
         ValueOperations<String, String> valueOperations = mock(ValueOperations.class);
         when(redisTemplate.opsForValue()).thenReturn(valueOperations);

backend-java/src/test/java/com/nanri/aiimage/modules/task/service/ModuleHistoryCleanupPayloadCleanupTest.java

@@ -9,7 +9,7 @@ import com.nanri.aiimage.config.InstanceMetadata;
 import com.nanri.aiimage.config.ModuleCleanupProperties;
 import com.nanri.aiimage.config.StorageProperties;
 import com.nanri.aiimage.config.TransientStorageProperties;
-import com.nanri.aiimage.modules.collectdata.mapper.CollectDataItemMapper;
+import com.nanri.aiimage.modules.task.spi.CollectDataItemCleanupSpi;
 import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataItemEntity;
 import com.nanri.aiimage.modules.file.service.object.RustfsObjectStorageService;
 import com.nanri.aiimage.modules.file.service.oss.OssStorageService;
@@ -86,7 +86,7 @@ class ModuleHistoryCleanupPayloadCleanupTest {
     private TaskResultPayloadMapper taskResultPayloadMapper;
     private TaskScopeStateMapper taskScopeStateMapper;
     private TaskChunkMapper taskChunkMapper;
-    private CollectDataItemMapper collectDataItemMapper;
+    private CollectDataItemCleanupSpi collectDataItemMapper;
     private DistributedJobLockService lockService;
     private TransientPayloadStorageService storage;
     private RustfsObjectStorageService rustfs;
@@ -122,7 +122,7 @@ class ModuleHistoryCleanupPayloadCleanupTest {
         taskResultPayloadMapper = mock(TaskResultPayloadMapper.class);
         taskScopeStateMapper = mock(TaskScopeStateMapper.class);
         taskChunkMapper = mock(TaskChunkMapper.class);
-        collectDataItemMapper = mock(CollectDataItemMapper.class);
+        collectDataItemMapper = mock(CollectDataItemCleanupSpi.class);
         StringRedisTemplate redisTemplate = mock(StringRedisTemplate.class);
         ValueOperations<String, String> valueOperations = mock(ValueOperations.class);
         when(redisTemplate.opsForValue()).thenReturn(valueOperations);

backend-java/src/test/java/com/nanri/aiimage/modules/task/service/TaskHeartbeatServiceTest.java

@@ -1,50 +1,48 @@
 package com.nanri.aiimage.modules.task.service;
 
+import com.baomidou.mybatisplus.core.MybatisConfiguration;
 import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
 import com.baomidou.mybatisplus.core.conditions.update.LambdaUpdateWrapper;
-import com.baomidou.mybatisplus.core.MybatisConfiguration;
 import com.baomidou.mybatisplus.core.metadata.TableInfoHelper;
 import com.nanri.aiimage.config.SimilarAsinProperties;
-import com.nanri.aiimage.modules.appearancepatent.service.AppearancePatentTaskCacheService;
-import com.nanri.aiimage.modules.brand.mapper.BrandCrawlTaskMapper;
-import com.nanri.aiimage.modules.brand.service.BrandTaskProgressCacheService;
 import com.nanri.aiimage.modules.collectdata.service.CollectDataService;
-import com.nanri.aiimage.modules.deletebrand.service.DeleteBrandTaskCacheService;
-import com.nanri.aiimage.modules.patroldelete.service.PatrolDeleteTaskCacheService;
-import com.nanri.aiimage.modules.pricetrack.service.PriceTrackTaskCacheService;
-import com.nanri.aiimage.modules.productrisk.service.ProductRiskTaskCacheService;
+import com.nanri.aiimage.modules.collectdata.service.CollectDataTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.publish.service.PublishTaskHeartbeatSpi;
 import com.nanri.aiimage.modules.publish.service.PublishTaskService;
-import com.nanri.aiimage.modules.queryasin.service.QueryAsinTaskCacheService;
 import com.nanri.aiimage.modules.shopdatacrawl.service.ShopDataCrawlTaskCacheService;
+import com.nanri.aiimage.modules.shopdatacrawl.service.ShopDataCrawlTaskHeartbeatSpi;
 import com.nanri.aiimage.modules.shopdatacrawl.service.ShopDataCrawlTaskService;
-import com.nanri.aiimage.modules.shopmatch.service.ShopMatchTaskCacheService;
 import com.nanri.aiimage.modules.similarasin.service.SimilarAsinTaskCacheService;
+import com.nanri.aiimage.modules.similarasin.service.SimilarAsinTaskHeartbeatSpi;
 import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
 import com.nanri.aiimage.modules.task.model.dto.TaskHeartbeatRequest;
 import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
 import com.nanri.aiimage.modules.task.model.vo.TaskHeartbeatVo;
-import com.nanri.aiimage.modules.withdraw.service.WithdrawTaskCacheService;
+import com.nanri.aiimage.modules.task.spi.BrandTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
 import org.apache.ibatis.builder.MapperBuilderAssistant;
 import org.junit.jupiter.api.BeforeAll;
-import org.junit.jupiter.api.Test;
-import org.junit.jupiter.api.extension.ExtendWith;
-import org.mockito.InjectMocks;
 import org.mockito.InOrder;
-import org.mockito.Mock;
-import org.mockito.junit.jupiter.MockitoExtension;
+import org.junit.jupiter.api.Test;
 
 import java.time.LocalDateTime;
+import java.util.List;
 
 import static org.junit.jupiter.api.Assertions.assertEquals;
 import static org.junit.jupiter.api.Assertions.assertTrue;
 import static org.mockito.ArgumentMatchers.any;
 import static org.mockito.ArgumentMatchers.isNull;
-import static org.mockito.Mockito.verify;
 import static org.mockito.Mockito.inOrder;
+import static org.mockito.Mockito.mock;
 import static org.mockito.Mockito.never;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.verifyNoInteractions;
 import static org.mockito.Mockito.when;
 
-@ExtendWith(MockitoExtension.class)
+/**
+ * 心跳分发契约（G5 SPI 化后）：task 侧只按 moduleType 路由到 {@link TaskModuleHeartbeatSpi}，
+ * 各模块的行为由各自的 SPI 实现负责，因此这里同时覆盖"路由正确"与"实现行为正确"。
+ */
 class TaskHeartbeatServiceTest {
 
     @BeforeAll
@@ -54,35 +52,30 @@ class TaskHeartbeatServiceTest {
                 FileTaskEntity.class);
     }
 
-    @Mock private FileTaskMapper fileTaskMapper;
-    @Mock private BrandCrawlTaskMapper brandCrawlTaskMapper;
-    @Mock private ProductRiskTaskCacheService productRiskTaskCacheService;
-    @Mock private PublishTaskService publishTaskService;
-    @Mock private PriceTrackTaskCacheService priceTrackTaskCacheService;
-    @Mock private ShopMatchTaskCacheService shopMatchTaskCacheService;
-    @Mock private PatrolDeleteTaskCacheService patrolDeleteTaskCacheService;
-    @Mock private QueryAsinTaskCacheService queryAsinTaskCacheService;
-    @Mock private ShopDataCrawlTaskCacheService shopDataCrawlTaskCacheService;
-    @Mock private ShopDataCrawlTaskService shopDataCrawlTaskService;
-    @Mock private WithdrawTaskCacheService withdrawTaskCacheService;
-    @Mock private AppearancePatentTaskCacheService appearancePatentTaskCacheService;
-    @Mock private SimilarAsinTaskCacheService similarAsinTaskCacheService;
-    @Mock private SimilarAsinProperties similarAsinProperties;
-    @Mock private TaskFileJobService taskFileJobService;
-    @Mock private DeleteBrandTaskCacheService deleteBrandTaskCacheService;
-    @Mock private BrandTaskProgressCacheService brandTaskProgressCacheService;
-    @Mock private CollectDataService collectDataService;
+    private final FileTaskMapper fileTaskMapper = mock(FileTaskMapper.class);
+    private final TaskProgressSnapshotService taskProgressSnapshotService = mock(TaskProgressSnapshotService.class);
+    private final BrandTaskHeartbeatSpi brandTaskHeartbeatSpi = mock(BrandTaskHeartbeatSpi.class);
 
-    @InjectMocks private TaskHeartbeatService service;
+    private TaskHeartbeatService serviceWith(TaskModuleHeartbeatSpi... handlers) {
+        return new TaskHeartbeatService(fileTaskMapper, taskProgressSnapshotService, brandTaskHeartbeatSpi,
+                List.of(handlers));
+    }
+
+    private static FileTaskEntity runningTask(long taskId, String moduleType) {
+        FileTaskEntity task = new FileTaskEntity();
+        task.setId(taskId);
+        task.setModuleType(moduleType);
+        task.setStatus("RUNNING");
+        return task;
+    }
 
     @Test
     @SuppressWarnings("unchecked")
     void publishHeartbeatTouchesGenericTaskAndModuleProgress() {
         long taskId = 20142L;
-        FileTaskEntity task = new FileTaskEntity();
-        task.setId(taskId);
-        task.setModuleType(PublishTaskService.MODULE_TYPE);
-        task.setStatus("RUNNING");
+        FileTaskEntity task = runningTask(taskId, PublishTaskService.MODULE_TYPE);
+        PublishTaskService publishTaskService = mock(PublishTaskService.class);
+        TaskHeartbeatService service = serviceWith(new PublishTaskHeartbeatSpi(publishTaskService));
 
         TaskHeartbeatRequest request = new TaskHeartbeatRequest();
         request.setPhase("dispatching");
@@ -90,7 +83,6 @@ class TaskHeartbeatServiceTest {
         request.setTotal(100);
 
         when(fileTaskMapper.selectOne(any(LambdaQueryWrapper.class))).thenReturn(task);
-        when(brandCrawlTaskMapper.selectOne(any(LambdaQueryWrapper.class))).thenReturn(null);
         when(fileTaskMapper.update(isNull(), any(LambdaUpdateWrapper.class))).thenReturn(1);
 
         TaskHeartbeatVo result = service.heartbeat(taskId, request);
@@ -99,7 +91,7 @@ class TaskHeartbeatServiceTest {
         assertEquals(PublishTaskService.MODULE_TYPE, result.getModuleType());
         InOrder routingBeforeUpdate = inOrder(publishTaskService, fileTaskMapper);
         routingBeforeUpdate.verify(publishTaskService)
-                .ensureTaskOwnedByCurrentInstance(task, "publish task heartbeat");
+                .ensureTaskOwnedByCurrentInstance(task, PublishTaskService.MODULE_TYPE + " task heartbeat");
         routingBeforeUpdate.verify(fileTaskMapper).update(isNull(), any(LambdaUpdateWrapper.class));
         verify(publishTaskService).touchHeartbeat(taskId, request);
     }
@@ -108,10 +100,12 @@ class TaskHeartbeatServiceTest {
     @SuppressWarnings("unchecked")
     void shopDataCrawlHeartbeatChecksOwnerBeforeDatabaseUpdate() {
         long taskId = 20143L;
-        FileTaskEntity task = new FileTaskEntity();
-        task.setId(taskId);
-        task.setModuleType("SHOP_DATA_CRAWL");
-        task.setStatus("RUNNING");
+        FileTaskEntity task = runningTask(taskId, "SHOP_DATA_CRAWL");
+        ShopDataCrawlTaskService shopDataCrawlTaskService = mock(ShopDataCrawlTaskService.class);
+        ShopDataCrawlTaskCacheService cacheService = mock(ShopDataCrawlTaskCacheService.class);
+        TaskHeartbeatService service = serviceWith(
+                new ShopDataCrawlTaskHeartbeatSpi(cacheService, shopDataCrawlTaskService));
+
         TaskHeartbeatRequest request = new TaskHeartbeatRequest();
         request.setModuleType("SHOP_DATA_CRAWL");
         when(fileTaskMapper.selectOne(any(LambdaQueryWrapper.class))).thenReturn(task);
@@ -122,27 +116,27 @@ class TaskHeartbeatServiceTest {
         assertTrue(result.isAlive());
         InOrder order = inOrder(shopDataCrawlTaskService, fileTaskMapper);
         order.verify(shopDataCrawlTaskService)
-                .ensureTaskOwnedByCurrentInstance(task, "shop data crawl task heartbeat");
+                .ensureTaskOwnedByCurrentInstance(task, "SHOP_DATA_CRAWL task heartbeat");
         order.verify(fileTaskMapper).update(isNull(), any(LambdaUpdateWrapper.class));
-        verify(shopDataCrawlTaskCacheService).touchTaskHeartbeat(taskId);
-        verify(shopDataCrawlTaskCacheService).saveTaskCache(task);
-        verify(brandCrawlTaskMapper, never()).selectOne(any(LambdaQueryWrapper.class));
+        verify(cacheService).touchTaskHeartbeat(taskId);
+        verify(cacheService).saveTaskCache(task);
+        // SHOP_DATA_CRAWL 不需要查品牌任务表（历史行为：跳过品牌查找）
+        verifyNoInteractions(brandTaskHeartbeatSpi);
     }
 
     @Test
     @SuppressWarnings("unchecked")
     void collectDataHeartbeatForwardsProcessedKeywordProgress() {
         long taskId = 21016L;
-        FileTaskEntity task = new FileTaskEntity();
-        task.setId(taskId);
-        task.setModuleType("COLLECT_DATA");
-        task.setStatus("RUNNING");
+        FileTaskEntity task = runningTask(taskId, "COLLECT_DATA");
+        CollectDataService collectDataService = mock(CollectDataService.class);
+        TaskHeartbeatService service = serviceWith(new CollectDataTaskHeartbeatSpi(collectDataService));
+
         TaskHeartbeatRequest request = new TaskHeartbeatRequest();
         request.setCurrent(4);
         request.setTotal(19);
 
         when(fileTaskMapper.selectOne(any(LambdaQueryWrapper.class))).thenReturn(task);
-        when(brandCrawlTaskMapper.selectOne(any(LambdaQueryWrapper.class))).thenReturn(null);
         when(fileTaskMapper.update(isNull(), any(LambdaUpdateWrapper.class))).thenReturn(1);
 
         TaskHeartbeatVo result = service.heartbeat(taskId, request);
@@ -155,21 +149,39 @@ class TaskHeartbeatServiceTest {
     @SuppressWarnings("unchecked")
     void similarAsinHeartbeatUsesRedisWithoutRefreshingRecentDatabaseCheckpoint() {
         long taskId = 20581L;
-        FileTaskEntity task = new FileTaskEntity();
-        task.setId(taskId);
-        task.setModuleType("SIMILAR_ASIN");
-        task.setStatus("RUNNING");
+        FileTaskEntity task = runningTask(taskId, "SIMILAR_ASIN");
         task.setUpdatedAt(LocalDateTime.now());
+        SimilarAsinTaskCacheService cacheService = mock(SimilarAsinTaskCacheService.class);
+        TaskFileJobService taskFileJobService = mock(TaskFileJobService.class);
+        SimilarAsinProperties properties = mock(SimilarAsinProperties.class);
+        TaskHeartbeatService service = serviceWith(
+                new SimilarAsinTaskHeartbeatSpi(cacheService, taskFileJobService, properties));
+
         when(fileTaskMapper.selectOne(any(LambdaQueryWrapper.class))).thenReturn(task);
-        when(brandCrawlTaskMapper.selectOne(any(LambdaQueryWrapper.class))).thenReturn(null);
-        when(similarAsinProperties.getDbTaskTouchIntervalMillis()).thenReturn(120000L);
+        when(properties.getDbTaskTouchIntervalMillis()).thenReturn(120000L);
 
         TaskHeartbeatVo result = service.heartbeat(taskId, new TaskHeartbeatRequest());
 
         assertTrue(result.isAlive());
         assertEquals("SIMILAR_ASIN", result.getModuleType());
-        verify(similarAsinTaskCacheService).touchTaskHeartbeat(taskId);
+        verify(cacheService).touchTaskHeartbeat(taskId);
         verify(taskFileJobService).touchRunningAssembleJobsIfStale(taskId, "SIMILAR_ASIN", 120000L);
+        // 检查点未到期：不写 biz_file_task.updated_at
         verify(fileTaskMapper, never()).update(isNull(), any(LambdaUpdateWrapper.class));
     }
+
+    @Test
+    void unknownModuleStillHeartbeatsFileTaskWithoutModuleHandler() {
+        long taskId = 20150L;
+        FileTaskEntity task = runningTask(taskId, "UNREGISTERED_MODULE");
+        TaskHeartbeatService service = serviceWith();
+
+        when(fileTaskMapper.selectOne(any(LambdaQueryWrapper.class))).thenReturn(task);
+        when(fileTaskMapper.update(isNull(), any(LambdaUpdateWrapper.class))).thenReturn(1);
+
+        TaskHeartbeatVo result = service.heartbeat(taskId, new TaskHeartbeatRequest());
+
+        assertTrue(result.isAlive(), "未注册 SPI 的模块不应影响通用心跳（保持向后兼容）");
+        assertEquals("UNREGISTERED_MODULE", result.getModuleType());
+    }
 }

backend-java/src/test/java/com/nanri/aiimage/modules/task/service/TaskModuleHeartbeatSpiCoverageTest.java

@@ -0,0 +1,96 @@
+package com.nanri.aiimage.modules.task.service;
+
+import com.nanri.aiimage.config.SimilarAsinProperties;
+import com.nanri.aiimage.modules.appearancepatent.service.AppearancePatentTaskCacheService;
+import com.nanri.aiimage.modules.appearancepatent.service.AppearancePatentTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.collectdata.service.CollectDataService;
+import com.nanri.aiimage.modules.collectdata.service.CollectDataTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.deletebrand.service.DeleteBrandTaskCacheService;
+import com.nanri.aiimage.modules.deletebrand.service.DeleteBrandTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.patroldelete.service.PatrolDeleteTaskCacheService;
+import com.nanri.aiimage.modules.patroldelete.service.PatrolDeleteTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.pricetrack.service.PriceTrackTaskCacheService;
+import com.nanri.aiimage.modules.pricetrack.service.PriceTrackTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.productrisk.service.ProductRiskTaskCacheService;
+import com.nanri.aiimage.modules.productrisk.service.ProductRiskTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.publish.service.PublishTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.publish.service.PublishTaskService;
+import com.nanri.aiimage.modules.queryasin.service.QueryAsinTaskCacheService;
+import com.nanri.aiimage.modules.queryasin.service.QueryAsinTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.shopdatacrawl.service.ShopDataCrawlTaskCacheService;
+import com.nanri.aiimage.modules.shopdatacrawl.service.ShopDataCrawlTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.shopdatacrawl.service.ShopDataCrawlTaskService;
+import com.nanri.aiimage.modules.shopmatch.service.ShopMatchTaskCacheService;
+import com.nanri.aiimage.modules.shopmatch.service.ShopMatchTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.similarasin.service.SimilarAsinTaskCacheService;
+import com.nanri.aiimage.modules.similarasin.service.SimilarAsinTaskHeartbeatSpi;
+import com.nanri.aiimage.modules.task.spi.TaskModuleHeartbeatSpi;
+import com.nanri.aiimage.modules.withdraw.service.WithdrawTaskCacheService;
+import com.nanri.aiimage.modules.withdraw.service.WithdrawTaskHeartbeatSpi;
+import org.junit.jupiter.api.Test;
+
+import java.util.LinkedHashMap;
+import java.util.Map;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.mockito.Mockito.mock;
+
+/**
+ * 模块心跳 SPI 的覆盖面自检（2026-09 全维度审查 G5）。
+ *
+ * <p>心跳按 moduleType 路由，moduleType 字符串写错会让该模块的心跳静默失效
+ * （任务被当成无心跳 → 30 分钟后误判失败）。这里把"哪个实现负责哪个 moduleType"
+ * 钉死：新增模块漏实现或写错字符串都会红测试。
+ */
+class TaskModuleHeartbeatSpiCoverageTest {
+
+    private static Map<String, TaskModuleHeartbeatSpi> implementationsByModuleType() {
+        Map<String, TaskModuleHeartbeatSpi> impls = new LinkedHashMap<>();
+        register(impls, new AppearancePatentTaskHeartbeatSpi(mock(AppearancePatentTaskCacheService.class)));
+        register(impls, new CollectDataTaskHeartbeatSpi(mock(CollectDataService.class)));
+        register(impls, new DeleteBrandTaskHeartbeatSpi(mock(DeleteBrandTaskCacheService.class)));
+        register(impls, new PatrolDeleteTaskHeartbeatSpi(mock(PatrolDeleteTaskCacheService.class)));
+        register(impls, new PriceTrackTaskHeartbeatSpi(mock(PriceTrackTaskCacheService.class)));
+        register(impls, new ProductRiskTaskHeartbeatSpi(mock(ProductRiskTaskCacheService.class)));
+        register(impls, new PublishTaskHeartbeatSpi(mock(PublishTaskService.class)));
+        register(impls, new QueryAsinTaskHeartbeatSpi(mock(QueryAsinTaskCacheService.class)));
+        register(impls, new ShopDataCrawlTaskHeartbeatSpi(
+                mock(ShopDataCrawlTaskCacheService.class), mock(ShopDataCrawlTaskService.class)));
+        register(impls, new ShopMatchTaskHeartbeatSpi(mock(ShopMatchTaskCacheService.class)));
+        register(impls, new SimilarAsinTaskHeartbeatSpi(
+                mock(SimilarAsinTaskCacheService.class), mock(TaskFileJobService.class),
+                mock(SimilarAsinProperties.class)));
+        register(impls, new WithdrawTaskHeartbeatSpi(mock(WithdrawTaskCacheService.class)));
+        return impls;
+    }
+
+    private static void register(Map<String, TaskModuleHeartbeatSpi> impls, TaskModuleHeartbeatSpi impl) {
+        TaskModuleHeartbeatSpi exists = impls.put(impl.moduleType(), impl);
+        if (exists != null) {
+            throw new IllegalStateException("moduleType=" + impl.moduleType() + " 重复实现: "
+                    + exists.getClass().getName() + " / " + impl.getClass().getName());
+        }
+    }
+
+    @Test
+    void everyHeartbeatModuleHasImplementationWithExactModuleType() {
+        Map<String, TaskModuleHeartbeatSpi> impls = implementationsByModuleType();
+
+        assertEquals(12, impls.size(), "有任务心跳的模块都应有 SPI 实现，实际 " + impls.keySet());
+        for (String moduleType : new String[]{
+                "APPEARANCE_PATENT", "COLLECT_DATA", "DELETE_BRAND", "PATROL_DELETE", "PRICE_TRACK",
+                "PRODUCT_RISK_RESOLVE", "PUBLISH", "QUERY_ASIN", "SHOP_DATA_CRAWL", "SHOP_MATCH",
+                "SIMILAR_ASIN", "WITHDRAW"}) {
+            assertEquals(true, impls.containsKey(moduleType), "缺少心跳实现: " + moduleType);
+        }
+    }
+
+    @Test
+    void moduleTypesAreUppercaseAndUnique() {
+        for (TaskModuleHeartbeatSpi impl : implementationsByModuleType().values()) {
+            String moduleType = impl.moduleType();
+            assertEquals(moduleType.toUpperCase(), moduleType,
+                    "moduleType 必须与 biz_file_task.module_type 一致（全大写）: " + moduleType);
+        }
+    }
+}