1688 多供应商比价

Model: qwen-max | ¥0.15/call

1688货源电商运营跨境电商卖家增长AIMS平台chat-visible

1688 多供应商比价：聚焦Ft 1688场景，提供跨境电商/外贸可落地的策略、流程与模板，适用于卖家与运营团队在 AIMS 平台直接调用。关键词：1688货源 / 电商运营 / 跨境电商 / 卖家增长。

Calls: 1

Skill Documentation

摘要

本技能基于 GitHub 高分仓库二次封装，原项目描述：在 1688.com 上做多供应商比价。输入产品关键词 + 可选规格，输出 5-15 个供应商的对比表：单价、起订量 MOQ、供应商年限、响应率、成交笔数、近 30 天复购率、所在地。覆盖"找同款"、"找低价"、"找靠谱工厂"三种模式。当用户说"1688 比价"、"找供应商"、"查 1688 上 XX 产品多少钱"、"找同款货源"时触发。在 AIMS 平台可直接调用，获得对应策略与方法论支持。

核心能力

ft-1688 — 1688 多供应商比价

Find the cheapest, most reliable, or most specialized supplier on 1688 for a given product. Outputs a side-by-side comparison table so you can make a sourcing decision in 5 minutes instead of opening 20 browser tabs.

When to invoke

User is doing product sourcing and wants a quick market scan on 1688.
User has a target price and wants to find suppliers that can hit it.
User wants to validate a specific supplier's claims (MOQ, sales volume, response rate).

Do not invoke for non-1688 sourcing (Alibaba International, Made-in-China, Global Sources are different platforms with different supplier bases).

Three modes

| Mode | Trigger | Goal |

|------|---------|------|

| **Cheapest** | "找最低价"、"最便宜的 XX" | Sort by unit price, ignore supplier quality signals |

| **Most reliable** | "找靠谱工厂"、"长期合作供应商" | Weight years-on-platform, response rate, repeat-buyer rate |

| **Spec match** | "找能定做 XX 的"、"能开私模的" | Filter by capability tags (OEM, ODM, customization) |

Default mode is **most reliable** if the user does not specify.

Workflow

Step 1 — Find candidate listings

`web_search` query templates (use the right one for the mode):

Generic: `web_search "site:1688.com {keyword}"`
Cheap: `web_search "site:1688.com {keyword} 价格"`
Reliable: `web_search "site:1688.com {keyword} {年限}年以上"` or `site:1688.com {keyword} 实力商家"`
Spec: `web_search "site:1688.com {keyword} 可定制"` or `site:1688.com {keyword} OEM`

Run 2-3 rounds with different angles. Aim for 8-15 unique product page URLs. De-duplicate by supplier.

If `web_search` returns poor coverage of 1688 (common in some regions), fall back to the bundled script:

python scripts/1688_search.py "{keyword}" --mode reliable --limit 15

The script uses 1688's public m-search.1688.com endpoint which does not require login. Expect occasional 403s — retry with a different `User-Agent` or wait 30s.

Step 2 — Extract per-listing fields

For each candidate URL, visit the product page and extract:

| Field | Source | Required? |

|-------|--------|-----------|

| Title | page H1 | yes |

| Min price (¥) | price block | yes |

| Max price (¥) | price block (range) | yes |

| MOQ | specs table | yes |

| Unit | "件" / "个" / "kg" / "米" | yes |

| Supplier name | header | yes |

| Supplier years on 1688 | supplier badge | yes |

| Supplier response rate | supplier badge | yes |

| 30-day sales volume | product page | yes |

| Repeat-buyer rate (近 30 天回头率) | product page | yes |

| Location (province) | header | yes |

| Has trade assurance (实力商家 / 源头工厂) | badge | yes |

| Shipping from (ship-from city) | shipping section | no |

| Lead time | shipping section | no |

If you are using `web_search` URL fetch and a field is not visible in the snippet, do not fabricate it. Mark it "not visible" in the report.

Step 3 — Normalize and compare

Compute a normalized "unit cost" for each listing so they can be compared apples-to-apples:

Same unit, same currency (¥).
If MOQ is high, the effective per-piece cost may differ at different volumes — show both MOQ price and 10x MOQ price where visible.
Exclude shipping unless the user asks (most 1688 listings are FOB domestic, not relevant to export pricing).

Step 4 — Output the comparison table

Markdown table with one row per supplier, sorted by the chosen mode:

## 1688 比价 — {keyword} — 模式: {cheapest|reliable|spec}

抓取时间: {ISO timestamp} | 共 {N} 个供应商 | 排序依据: {criterion}

| # | 供应商 | 标题 | 单价 | MOQ | 起批量单价 | 30天成交 | 回头率 | 响应率 | 年限 | 产地 | 实力商家 |
|---|--------|------|------|-----|------------|----------|------

…(原始内容较长，已截断，完整版见源仓库)


## 常见问题
## 这个技能适合谁使用?
适合外贸业务员、采购、合规与运营团队，在「1688 多供应商比价」场景中需要标准化流程、模板与合规提醒时调用。

## 在 AIMS 平台上怎么用?
直接在对话里描述业务目标（如"帮我做 {cn_name} 的方案"），平台调用本 skill 返回可执行步骤与模板。

## 涉及合规的部分要注意什么?
本技能仅给方法论与红线提醒，不构成法律意见；跨境结汇、税务、知识产权等务必遵守中国与目标国法规，必要时咨询专业机构。
## 模型对照说明

**原仓库推荐模型**: 未强制绑定（Agent Skills 标准，可在 Claude Code / Codex / Cursor 等客户端运行，或对接任意大模型）
**AIMS 平台实际使用**: qwen-max（15 分）
**AIMS 平台可用替代**（待海外通道恢复后可切换）:
- 国内: qwen-plus / qwen-turbo / minimax-m1 / minimax-m2 / minimax-m3
- 海外（当前占位，待恢复）: gpt-4.1 / gpt-4o / claude-opus-4.7 / gemini-2.5-pro

**对齐度**: 中 — 该类技能以方法论 / 策略 / 流程为主，qwen-max 足以承载中文场景的推理与生成。
**切换模型实操**: `aims.update_skill --skill_id <sk> --model_name qwen-max --call_price_fen 15`
## 风险与说明

本技能为开源项目二次封装，仅提供方法论、策略与流程建议，**不构成投资、法律或税务意见**。
涉及经营数据与用户信息时，请遵守《数据安全法》《个人信息保护法》及相关平台规则，做好脱敏与授权。

**供应链提示**: 比对供应商时以平台官方数据为准，警惕刷单与虚假交易信号；下单前务必核验资质与样品。
## 详细使用方式（安装与验证）

**在 AIMS 平台使用**: 直接调用本 skill 获得外贸实战方法论与模板；原始仓库 https://github.com/Elvinaskill/foreign-trade-suite。

Open AIMS China Panel