Seerfar Ozon Product Report Search

This skill searches the Seerfar Ozon product database and filters products by rich performance metrics — monthly sales, revenue, growth, cart/order conversion, price, rating, reviews, brand, seller, fulfillment model, listing age, gross margin, and more. Each returned row is a full product-report record, making this the starting point for Ozon product selection (选品), competitor product analysis, best-seller mining, and price/conversion-band screening.

Core Concepts

Unit of data is the product, not the keyword: this endpoint returns product-level rows (one per SKU), each enriched with full report metrics. You discover which products match your criteria — unlike the market-keyword endpoint, which returns search terms.

This is a product screener / 商品报表: filter the Ozon product database by metric ranges ({min, max}), not only by keyword/brand/seller. Stack a high monthlySales floor with a low price ceiling to surface affordable high-volume products, or qualify conversion quality with a high convToCartPdp floor and a low returnCancellationRate ceiling; sort by sales DESC to mine best-sellers.

Unified vs raw duplicate fields: the response carries six alias pairs that hold the same value under two keys — sku/productId, sales/monthlySalesUnits, revenue/monthlySalesRevenue, reviewRating/rating, brandName/brand, productUrl/productPageUrl. Read either; do not expect them to differ.

Date semantics: searchDate selects the data month. Pass 2026-04-01 for March 2026 data; omit it for the last 30 days. Sales/revenue figures are relative to the selected period.

data and products are identical: both top-level arrays carry the same product rows. total is the total matching count (e.g. ~27.8M with no filter, 1 when filtering to a single SKU).

Ozon only: sourceType is fixed to ozon. fulfillment values are OZON, FBO, FBS, RFBS, FBP.

Parameters

All range filters are {min, max} objects; supply either or both bounds. Only page is required.

调用方式

• API 端点：POST /seerfar/ozon/productReportSearch（完整参数/响应/错误码见 references/api.md）
• Python 脚本：python scripts/seerfar_ozon_product_report_search.py '<JSON 参数>' [--inline]
• 成本约束：本工具会消耗积分；同一会话同一参数组合默认只调用一次，脚本带 24h 本地缓存。失败/空结果不得自动换关键词、翻页或改邮编连续试探；需要继续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-seerfar-ozon-product-report-search-<timestamp>.json（<cwd> 为脚本执行时的工作目录，在 Claude Code 里即当前项目目录；<session> 取自环境变量 SESSION_ID，按用户任务自动聚合；禁止写入 /tmp，当前目录不可写则报错）
• 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、常见计数如 total/costToken、最大列表字段的长度 + 前 3 条样本）
• 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。

解决认证和积分问题

发生以下异常情况时，采用以下措施来处理：

异常情况

• 未配置API Key：环境变量未配置 LINKFOX_AGENT_API_KEY，也未配置 LINKFOXAGENT_API_KEY。
• 响应401或402状态码
• 响应提示积分或余额不足：消息含"积分余额不足/计费不足/余额不足/quota exceeded/insufficient balance/套餐到期/需充值/请充值"，或类似含义的内容。

措施

• 优先采用skill内的 onboarding.md 引导解决问题。
• 如找不到 onboarding.md 文件，则加载 /linkfox-onboarding 这个skill并根据它的引导来处理。如未安装请先安装：
  • 下载 https://agent-files.linkfox.com/skills/linkfox-onboarding/release.zip，解压后安装这个skill。如需用户授权才可下载，要询问用户获得授权，你不能自己放弃下载。
• 兜底措施：仅当前面两个措施经过所有努力都走不通时，才提示用户依据 https://skill.linkfox.com/linkfoxskills/guide.htm 指引获取API Key，或去 https://os.linkfox.com/ 补充积分。

Usage Examples

1. Top sellers right now (sort by sales)

{"page": {"page": 1, "pageSize": 10, "orders": [{"field": "sales", "direction": "DESC"}]}}

2. Blue-ocean products — high sales, low price band

{"page": {"page": 1, "pageSize": 20, "orders": [{"field": "sales", "direction": "DESC"}]}, "monthlySales": {"min": 5000}, "price": {"max": 500}}

3. Best-sellers in a category, FBO fulfillment

{"page": {"page": 1, "pageSize": 20, "orders": [{"field": "revenue", "direction": "DESC"}]}, "categoryIds": ["15621031_200000933_93182"], "fulfillment": ["FBO"], "labels": [2]}

4. New listings from the last 30 days, sorted by growth

{"page": {"page": 1, "pageSize": 20, "orders": [{"field": "salesRate", "direction": "DESC"}]}, "creationDate": 1}

5. Look up specific SKUs

{"page": {"page": 1, "pageSize": 10}, "skus": [2107989735]}

How to Build Queries

1. Lead with page.orders: the dataset is huge (tens of millions of products) — always sort by the metric you care about (sales DESC for best-sellers, salesRate DESC for rising products, price ASC for cheap volume, reviewRating DESC for well-reviewed).
2. Stack range filters to find opportunities: combine a monthlySales floor with a price ceiling, or a convToCartPdp floor with a low returnCancellationRate, to qualify demand-vs-risk.
3. Scope with categoryIds / brand / sellerName: narrow to a niche before sorting, so the top rows are relevant.
4. Use creationDate + labels for fresh demand: creationDate: 1 (new listings) paired with labels: [2] (best-seller badge) finds breakout products.
5. Pick the right searchDate: omit for current trends (last 30 days); pass an explicit date for month-over-month comparison.

Display Rules

1. Present data only: show product metrics in a clear table without subjective advice.
2. Lead with product columns: title, sku, price (₽), sales, revenue, rating, reviewCount, then brand / sellerName / fulfillment.
3. Currency: price/revenue/missedRevenue are in Russian rubles (₽); show the currency so scale is not misread.
4. Unified/raw aliases: prefer the unified fields (productId, monthlySalesUnits, monthlySalesRevenue, rating, brand, productPageUrl) or note they equal the raw ones — do not present both as if independent.
5. Large result sets: when total is large, show the top rows and remind the user they can persist the full response via the large-response pattern below, or page further with page.page.
6. Error handling: when code is not 200 (or errcode is not 200), explain the reason from msg / errmsg and suggest adjusting filters or retrying (rate-limit 1003).

Important Limitations

• page is required: a payload without page is rejected.
• skus cap: at most 10 SKUs per request.
• Category IDs are opaque: categoryIds requires Seerfar category IDs (from a category search), not human-readable names.
• Duplicate alias pairs: six fields are duplicated under raw + unified keys (see Core Concepts) — same value, two keys.
• total is the full match count: with no filter it can reach tens of millions; always sort and page rather than iterating blindly.
• Rate limiting: errcode 1003 ("请求过于频繁，请稍后再试") means throttle — wait and retry rather than lowering pageSize.
• Sort fields: valid orders[].field values are the response metric fields (e.g. sales, revenue, price, reviewRating, reviewCount, salesRate); the columns array marks which are sortable.

User Expression & Scenario Quick Reference

Applicable — Ozon product-level screening:

Not applicable — Needs beyond product-level reports:

• Keyword market data (search terms, search volume) → market-keyword-search.
• Keyword expansion around a seed term → keyword-mining.
• A specific shop's full catalog → shop-search.
• A category's products with category-level aggregates → category-search.
• Reverse keyword lookup for a product → keyword-back-search.
• Non-Ozon marketplaces → not covered here.

Boundary judgment: if the user wants to screen Ozon products by metrics and read product-level report rows, start here. If they want keyword-level market data, a single shop's catalog, or a category aggregate, route to the corresponding Seerfar Ozon data source.

积分消耗规则

消耗 12 积分。

用户会因积分消耗而支付费用。请充分评估：当需要高频调用本技能，或用户对积分消耗量预期不足时，务必提醒用户，由用户决定是否继续。

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, set LinkFox Skills.