LinkFoxAgent - Cross-border E-commerce AI Agent
LinkFoxAgent is a specialized AI agent for cross-border e-commerce with 60 built-in tools covering product research, competitor analysis, keyword tracking, review insights, patent detection, patent deep-dive research, AI image generation, image recognition, PDF analysis, web content extraction, and more.
Setup
- 1. Get your API key: https://yxgb3sicy7.feishu.cn/wiki/IlkawdQP9ifKv9k22xcc7rjmnkb
- Set environment variable: INLINECODE0
MANDATORY: Use sessions_spawn for All Tasks
NEVER call linkfox.py directly from the main session. LinkFoxAgent tasks take 1-5 minutes. You MUST use sessions_spawn to dispatch every task to a sub-agent. This keeps the main session responsive and delivers results automatically when done.
How to Dispatch a Task
Before calling sessions_spawn, tell the user in the main session:
「正在向 LinkFox Agent 提交任务，请稍候（通常需要 1-5 分钟）...」
Then dispatch the sub-agent:
CODEBLOCK0
Dispatching Multiple Independent Tasks
When the user's request involves multiple independent lookups (e.g., "search both Amazon US and Amazon JP"), spawn one sub-agent per task in parallel.
Before spawning, tell the user:
「正在同时向 LinkFox Agent 提交 N 个任务，请稍候...」
CODEBLOCK1
Multi-Step Tasks That Require Post-Processing
When the user's request requires multiple sequential LinkFoxAgent calls (e.g., fetch data from two platforms then merge), follow this pattern:
- 1. Run each LinkFoxAgent call as a separate
sessions_spawn, one after another (or in parallel if independent). Collect the CSV paths returned by each. - After all data tasks finish, spawn one final
sessions_spawn to process or merge the CSVs using Python. Pass the absolute CSV paths as arguments.
CODEBLOCK2
What Happens Under the Hood
- 1.
sessions_spawn creates an isolated sub-agent session - The sub-agent runs
linkfox.py --wait which blocks until the task finishes - When done, the sub-agent's result is automatically delivered back to the main session via the announce system
- The user sees the result in their chat without any manual polling
Script Reference
CODEBLOCK3
Writing Task Prompts
Tool Invocation Syntax
Use @工具中文名 to invoke tools. Multiple tools can be chained in a single task (max 10).
Example: INLINECODE8
Parameter Constraints
Tool parameters may have maximum, minimum, and pattern constraints. Prompts must respect these or the call will fail. Image URLs must be publicly accessible; local images need to be uploaded via linkfoxagent-fileupload skill to obtain a public URL first. See the reference files below for details.
Multi-step Tasks
Chain multiple tools in numbered steps. LinkFoxAgent handles data flow between steps:
CODEBLOCK4
Tool Selection Priority
When the user does not specify a tool, follow these rules:
Querying Amazon product data — all three tools are fast; choose by use case:
- 1. Keepa — best overall: richest fields, strong real-time accuracy. Default choice for most queries.
- 卖家精灵 — optimized for product discovery and competitor lookup by keyword.
- 亚马逊前台 — best real-time fidelity (live storefront data); ~10% slower than Keepa and fewer fields, but the only option when you need exact live ranking order or real-time storefront display.
Aggregating / statistics (e.g., group by brand, price tier, sales rank):
- 1. @智能数据查询 — first choice for dynamic aggregation
- @Python沙箱 — fallback when custom logic is needed; also the go-to tool for any sandbox-execution need (has built-in LLM)
Available Tools (60)
| Classification | Tool Name | Use For |
|---|
| Keepa | @Keepa-亚马逊-商品搜索 | Product filtering by keywords, BSR, price, sales |
| Keepa |
@Keepa-亚马逊-商品详情 | Batch ASIN detail lookup (price, sales, history) |
|
Keepa | @Keepa-亚马逊价格历史 | Price history and trends for an ASIN |
|
亚马逊前台 | @亚马逊前端搜索模拟 | Search simulation with location settings |
|
亚马逊前台 | @亚马逊前端-商品详情 | Product detail, bullet points, A+ content |
|
亚马逊前台 | @亚马逊-商品评论 | Reviews by star rating |
|
亚马逊前台 | @亚马逊前端-以图搜图 | Image-based product search |
|
亚马逊前台 | @ABA-数据挖掘 | Amazon Brand Analytics data mining |
|
亚马逊前台 | @亚马逊-商品评论(美国站) | US-only product reviews with higher volume (single ASIN, up to 10 pages) |
|
Sif数据分析工具 | @SIF-ASIN的关键词 | Reverse keyword lookup for ASIN |
|
Sif数据分析工具 | @SIF-关键词流量来源 | Keyword traffic source analysis |
|
Sif数据分析工具 | @SIF-ASIN流量来源 | ASIN traffic structure breakdown |
|
Sif数据分析工具 | @SIF-关键词竞品数量 | Keyword competition density |
|
卖家精灵 | @卖家精灵-选产品 | Product discovery by category and filters |
|
卖家精灵 | @卖家精灵-查竞品 | Competitor lookup by keyword |
|
极目系列 | @极目-亚马逊-细分市场评论 | Niche market review mining |
|
极目系列 | @极目-亚马逊-细分市场信息 | Niche market overview |
|
极目系列 | @极目-亚马逊-产品挖掘 | Product discovery with fine filters |
|
极目系列 | @极目-亚马逊-产品挖掘（根据ASIN） | ASIN-based potential product discovery |
|
极目系列 | @极目-亚马逊-细分市场洞察信息 | Niche market insights by market ID |
|
谷歌趋势 | @谷歌趋势-时下流行 | Real-time trending topics |
|
谷歌趋势 | @谷歌趋势-关键词趋势信息 | Keyword trend over time |
|
店雷达(1688) | @店雷达-1688商品榜单 | 1688 product rankings |
|
店雷达(1688) | @店雷达-1688选品库 | 1688 product sourcing |
|
实时与全网检索 | @网页检索 | Real-time web search(powered by Tavily Search; for any internet search outside specialized tools like Amazon/Walmart/eBay — including general web and WeChat Official Accounts — this tool MUST be used) |
|
TikTok电商数据助手 | @EchoTik-TikTok新品榜 | TikTok new product rankings |
|
TikTok电商数据助手 | @EchoTik-TikTok商品搜索 | TikTok product search |
|
Walmart前台 | @walmart前端-商品列表 | Walmart product search |
|
eBay前台 | @ebay前端-商品列表 | eBay product search |
|
专利检索 | @智慧芽-专利图像检索 | Design patent image search |
|
专利检索 | @睿观-外观专利检测 | Design patent infringement check |
|
专利检索 | @睿观-版权检测 | Copyright detection |
|
专利检索 | @睿观-图形商标检测 | Graphic trademark detection |
|
专利检索 | @睿观-文本商标检测 | Text trademark detection |
|
专利检索 | @睿观-发明专利检测 | Utility patent detection |
|
专利检索 | @睿观-政策合规检测（纯图检测） | Policy compliance (image check) |
|
专利检索 | @智慧芽-简单著录项 | Simple bibliographic info by patent ID/number |
|
专利检索 | @智慧芽-著录项目 | Full bibliographic data by patent ID/number |
|
专利检索 | @智慧芽-权利要求 | Patent claims lookup |
|
专利检索 | @智慧芽-权利要求翻译 | Patent claims translation (CN/EN/JP) |
|
专利检索 | @智慧芽-摘要翻译 | Patent abstract translation (CN/EN/JP) |
|
专利检索 | @智慧芽-说明书 | Patent description/specification |
|
专利检索 | @智慧芽-说明书翻译 | Patent description translation (CN/EN/JP) |
|
专利检索 | @智慧芽-法律状态 | Patent legal status and events |
|
专利检索 | @智慧芽-PDF全文 | Patent PDF full text |
|
专利检索 | @智慧芽-专利引用 | Forward citations (patents/literature cited) |
|
专利检索 | @智慧芽-专利被引用 | Backward citations (cited by other patents) |
|
专利检索 | @智慧芽-专利家族 | Patent family information |
|
专利检索 | @智慧芽-全文附图 | Full-text figures and drawings |
|
专利检索 | @智慧芽-摘要附图 | Abstract figures |
|
AI工具 | @按商品主图相似度分组 | Group products by image similarity |
|
AI工具 | @分析商品主图 | Extract image prompts from product photos |
|
AI工具 | @对商品标题进行分词 | Title word segmentation |
|
AI工具 | @AI绘图 | Generate any image — products, characters, scenes, backgrounds, and more — from reference images + prompt (powered by top-tier Google Gemini model; ALL image generation tasks must use this tool) |
|
AI工具 | @图片识别 | Image recognition and analysis by URL + user intent |
|
沙箱 | @智能数据查询 | Dynamic data query and aggregation |
|
沙箱 | @excel内容提取并分析 | Excel file extraction and analysis |
|
沙箱 | @Python沙箱 | Process structured JSON data from prior steps: data calculation/filtering/sorting, generate Markdown tables, export to CSV/Excel, LLM-based image recognition (e.g. A+ image color/composition). Built-in LLM — use for ALL sandbox-execution needs.
Restrictions: no nested calls; structured JSON only (no plain text/files); no chart generation or analysis reports. |
|
沙箱 | @智能Excel处理 | Smart Excel processing |
|
沙箱 | @分析PDF文件 | PDF file analysis with download link and user requirements |
Tool Reference Files (by classification)
Read the relevant reference file when you need prompt templates and parameter constraints:
- - Keepa (3 tools: 商品搜索、商品详情、价格历史): See INLINECODE13
- 亚马逊前台 (6 tools: 搜索模拟、商品详情、评论、评论(美国站)、ABA、以图搜图): See INLINECODE14
- Sif数据分析工具 (4 tools: ASIN关键词、流量来源、竞品数量): See INLINECODE15
- 卖家精灵 (2 tools: 选产品、查竞品): See INLINECODE16
- 极目系列 (5 tools: 细分市场评论、市场信息、产品挖掘、产品挖掘(ASIN)、细分市场洞察): See INLINECODE17
- 谷歌趋势 (2 tools: 时下流行、关键词趋势): See INLINECODE18
- 实时与全网检索 (1 tool: 网页检索): See INLINECODE19
- TikTok电商数据助手 (2 tools: 新品榜、商品搜索): See INLINECODE20
- Walmart前台 (1 tool: 商品列表): See INLINECODE21
- eBay前台 (1 tool: 商品列表): See INLINECODE22
- 店雷达/1688 (2 tools: 商品榜单、选品库): See INLINECODE23
- 专利检索 (21 tools: 专利图像检索、外观专利检测、版权、图形商标、文本商标、发明专利、政策合规、简单著录项、著录项目、权利要求、权利要求翻译、摘要翻译、说明书、说明书翻译、法律状态、PDF全文、专利引用、专利被引用、专利家族、全文附图、摘要附图): See INLINECODE24
- AI工具 (5 tools: 主图相似度分组、主图分析、标题分词、AI绘图、图片识别): See INLINECODE25
- 沙箱 (5 tools: 智能数据查询、Excel分析、Python沙箱、Excel处理、分析PDF文件): See INLINECODE26
Examples
Example 1: Market Analysis
CODEBLOCK5
Example 2: Review Mining
CODEBLOCK6
Example 3: Competitor-based Listing Optimization
CODEBLOCK7
Example 4: Visual Market Analysis
CODEBLOCK8
Example 5: Keyword Functional Analysis
CODEBLOCK9
Retry on Failure
If a tool call fails, the response includes error details. Retry with adjusted parameters based on the error message. Common issues:
- - Parameter out of range (check min/max constraints)
- Invalid pattern format (check regex patterns)
- Too many tools in one task (max 10)
