Temu Product Search & Detail

This skill supports a dependency-aware Temu product research workflow:

1. Resolve a marketplace with the site-list endpoint when regionId is unknown.
2. Resolve category IDs with the category-list endpoint when category filtering is needed.
3. Search products with the selected regionId and optional catIds.
4. Fetch one product's detail and recent 30-day history with goodsId and the same regionId.

Use only the endpoints needed for the user's request. Read references/api.md for the complete request and response contract.

The site-list and category-list endpoints are helper capabilities for product search/detail. A request solely for site or category metadata must not auto-select this skill. If the user explicitly invokes this skill, or the skill is already active, siteList or categoryList may be called independently.

Core Concepts

• regionId is the marketplace selector. Get a non-null positive value from sites[].regionId; never substitute siteId.
• catIds is an array of category IDs for product search. Populate it only from non-null categories[].catId values.
• parentCatId drills into the category tree. Omit it for top-level categories, then reuse a returned catId to fetch children.
• Search returns paginated items[]; detail requires one goodsId and returns goods plus history[].
• The default marketplace is regionId: 211 (United States). Search defaults are page: 1, size: 20, and matchMode: 2.
• Treat every paid endpoint call as a new charge. Within one session, call the same endpoint and parameters only once unless the user explicitly asks to repeat it.

Shared Product Response Fields

Endpoints

/geekbi/temu/goodsSearch

Request

• regionId, page, and size select the marketplace and page; page * size must not exceed 10,000.
• keyword matches the title. matchMode is 1 for strict or 2 for fuzzy matching.
• catIds accepts an integer array, for example [984, 982].
• Range filters use paired *Min / *Max fields for sales units, revenue, price, supply price, rating, reviews, inventory, and growth rates. Never make a minimum greater than its maximum.
• status is an integer array whose elements may be 1 (normal), 2 (out of stock), or 3 (delisted). hostingMode is 1 (full-managed) or 2 (semi-managed).
• onSaleTimeMin/Max and mallOpenTimeMin/Max use ISO-8601 date-time strings.
• order is asc or desc; use only a sort field accepted by the API or returned in an established workflow.

Response

/geekbi/temu/goodsDetail

Request

• goodsId is required and normally comes from search items[].goodsId.
• Pass the same regionId used for search so prices and market context remain consistent. Only reuse a non-empty items[].goodsId; stop and explain when search returns no valid ID.

Response

/geekbi/temu/siteList

Request

• Site list takes {}. Use only a non-null positive regionId from its response, never siteId.

Response

/geekbi/temu/categoryList

Request

• Category list takes {} for top-level categories or {"parentCatId": <catId>} for children; reuse only a non-null valid catId.

Response

Common Response Rules

Business fields may be absent or null. Preserve the runtime JSON type and do not replace missing values with zero.

Helper Endpoint Rule

• For a helper-only request, call only the requested free helper endpoint. Do not invoke goodsSearch or goodsDetail unless the user also requests product data.

调用方式

• API 端点：POST /geekbi/temu/goodsSearch、POST /geekbi/temu/goodsDetail、POST /geekbi/temu/siteList 或 POST /geekbi/temu/categoryList（完整参数/响应/错误码见 references/api.md）
• Python 脚本：python scripts/geekbi_temu_goods_search.py '<JSON 参数>' [--inline]、python scripts/geekbi_temu_goods_detail.py '<JSON 参数>' [--inline]、python scripts/geekbi_temu_site_list.py '{}' [--inline] 或 python scripts/geekbi_temu_category_list.py '<JSON 参数>' [--inline]
• 成本约束：商品搜索/详情各消耗 15 算力/次，站点/品类列表不消耗算力；同一端点同一参数组合使用 24h 本地缓存，同一会话默认只调用一次。--inline 不绕过缓存；--no-cache 会跳过缓存读写并强制真实请求，付费端点可能再次扣算力，免费辅助端点仅强制刷新。失败、空结果或瞬时错误不得自动重试，也不得自动换关键词、翻页或改站点连续试探；需要继续付费检索时先向用户说明会产生额外消耗。

输出策略（脚本默认行为）：

• 始终将完整响应写入 <cwd>/linkfox/<YYYY-MM-DD>/<session>/data/linkfox-geekbi-temu-product-<timestamp>.json；<cwd> 为脚本执行时的当前工作目录，<session> 取自环境变量 SESSION_ID，未设置时自动生成。当前目录不可写时直接报错，不回退到用户目录或系统临时目录。
• 响应体 ≤ 8 KB：落盘后把完整 JSON 打印到 stdout
• 响应体 > 8 KB：落盘后 stdout 只输出摘要（顶层字段、errcode/errmsg/total 等公共字段、最大业务列表的长度 + 前 3 条样本；忽略 columns 渲染元数据）
• 加 --inline 强制全量打印到 stdout（同样落盘）

读数据建议：先看摘要判断是否足够；需要具体字段时优先用 jq或ConvertFrom-Json 从保存的 json 文件按需抽取，避免整份 JSON 进入上下文。

解决认证和算力问题

发生以下异常情况时，采用 references/onboarding.md 引导解决问题：

异常情况

• 未配置 API Key：环境变量未配置 LINKFOX_AGENT_API_KEY 或兼容键 LINKFOXAGENT_API_KEY。
• 响应 401 或 402 状态码。
• 响应提示算力或余额不足：消息含“算力余额不足/计费不足/余额不足/quota exceeded/insufficient balance/套餐到期/需充值/请充值”或类似含义。

Usage Examples

# Search products; catIds contains categories[].catId values
python scripts/geekbi_temu_goods_search.py '{"regionId":211,"keyword":"dress","catIds":[984],"page":1,"size":20,"sort":"sold","order":"desc"}'

# Reuse items[].goodsId and the same regionId
python scripts/geekbi_temu_goods_detail.py '{"goodsId":"<goodsId from search>","regionId":211}'

# Resolve regionId (use sites[].regionId, not siteId)
python scripts/geekbi_temu_site_list.py '{}'

# Browse top-level categories, then fetch one node's children
python scripts/geekbi_temu_category_list.py '{}'
python scripts/geekbi_temu_category_list.py '{"parentCatId":100}'

Display Rules

1. State the selected marketplace (regionId, site name, and currency when available), filters, page, and page size.
2. For search results, show product title, goodsId, price range, total/monthly sales, revenue, rating, reviews, inventory, shop, category, status, and listing time when returned.
3. For detail, group the current product profile separately from the 30-day history. Label the date/time field used for each historical row.
4. Treat rate fields as percentages only when the response contract identifies them as rates; do not silently convert units.
5. Keep category IDs and product IDs available for follow-up calls, but do not expose internal siteId as a marketplace selector.
6. Preserve null and unavailable fields rather than inventing values.

Important Limitations

• This skill intentionally packages four related endpoints as one product workflow. Keep each script's API_PATH and business-list summary behavior independent.
• This skill researches public Temu market data; it does not access or operate a seller account.
• Search is limited to the first 10,000 matching records, with size from 1 to 200.
• Detail requires an exact goodsId; it cannot infer one from a title alone.
• Monetary values follow the selected site's currency, except supply-price filters explicitly documented in CNY.
• Product fields and history fields may be missing or null; upstream extensions can appear under extraFields.
• Do not automatically call site/category helpers when the user already supplied valid IDs.
• Do not retry a paid endpoint with changed filters after an empty result without user confirmation.

User Expression & Scenario Quick Reference

Applicable: “用 GeekBI 查 Temu 商品”, “看这个 goodsId 的详情和近 30 天趋势”, “查 Temu 商品供货价和库存”, “比较日/周/月销量或销售额增长”, “linkfox-geekbi-temu-product”.

Adjacent: Within an active product search/detail workflow, use the free helper endpoints when a requested site or category has no known ID. When the user explicitly invokes this skill, or this skill is already active, either helper may also be called independently without calling a paid product endpoint. Category-name resolution may require starting at the top level and following parentCatId through multiple levels. When the user supplies a product URL but no goodsId, extract only if the ID is unambiguous; otherwise ask for it.

Not an automatic trigger: A request solely to enumerate Temu sites, regionId values, categories, or catId values, unless the user explicitly invokes this skill or the skill is already active.

Not applicable: Generic Temu keyword/category/price/rating/sales filtering without GeekBI-specific fields (use linkfox-temu-product-query); Temu seller-center orders, inventory updates, listing publication, fulfillment, account authorization, advertising operations, or research on another marketplace. Unless the user explicitly requests cross-source comparison, do not call both paid product-data skills.

算力消耗规则

商品搜索和商品详情每次各消耗 15 算力；站点列表和品类列表不消耗算力。

用户会为商品搜索和商品详情支付算力。高频调用、翻页或从搜索继续查询多个商品详情前，必须让用户决定是否继续。

Feedback:

Auto-detect and report feedback via the Feedback API when any of the following apply:

1. The functionality or purpose described in this skill does not match actual behavior
2. The skill's results do not match the user's intent
3. The user expresses dissatisfaction or praise about this skill
4. Anything you believe could be improved

Call the feedback API as specified in references/api.md. Do not interrupt the user's flow.

For more high-quality, professional cross-border e-commerce skills, visit LinkFox Skills.