@gecho-ai/gecho-bridge 1.1.31-beta.0 → 1.1.37

package/.claude-plugin/plugin.json

@@ -1,5 +1,5 @@
 {
   "name": "@gecho-ai/gecho-bridge",
-  "version": "1.1.30",
+  "version": "1.1.37",
   "description": "TikTok Search Tool for OpenClaw"
 }

package/README.md

@@ -87,6 +87,44 @@ openclaw gateway restart
 *If you need to upgrade an installed version, use `openclaw plugins update clawhub:@gecho-ai/gecho-bridge-bundle`.*
 *The plugin will automatically start a local Gecho service when needed. If the browser extension was opened after the client, run `openclaw gateway restart` once to reconnect cleanly.*
 
+### Publish to Tencent SkillHub
+
+Tencent SkillHub currently requires top-level `slug`, `version`, and `displayName` fields in `SKILL.md`. This repository provides a dedicated publisher: every run rebuilds copies under the Git-ignored `tmp/tencent-skillhub-publish/` directory and leaves source Skills unchanged.
+
+Each distribution Skill has a `publish.json` with platform-specific slug/display-name values for ClawHub, Tencent SkillHub, and ModelScope. Tencent's Chinese entries keep the slug of existing publications, while English entries use a separate slug so an update is not accidentally published as a new Skill. These config files are used only during staging and are not uploaded.
+
+```bash
+# Build publish copies for all English and Chinese Skills
+npm run skillhub:tencent:stage
+
+# Run the official CLI dry-run for one Skill
+npm run skillhub:tencent:dry-run -- --skill tiktok-insight
+
+# Publish one Skill through the public/community CLI
+skillhub login --key skh_...
+npm run skillhub:tencent:publish -- --community --skill tiktok-insight --locale zh-CN
+
+# Check platform metadata for every Skill
+npm run publish:config:check
+```
+
+For the public community, use a `skh_...` key and pass `--community` so a saved enterprise credential is not selected accidentally. The installed official CLI is currently `2026.8.5`: it has `skillhub publish` for the public community, but not the `skillhub skill publish` command shown in the website screenshot.
+
+An enterprise key (`sk-ent-...`) can log in to the enterprise source and search or install team Skills, but the current team upload endpoint requires a web login session. Team publication must therefore be completed in the [enterprise publishing console](https://skillhub.cn/enterprise/dashboard/publish); this repository can still generate and validate the upload copy.
+
+```bash
+# Public/community publication: log in with skh_... first
+skillhub login --key skh_...
+npm run skillhub:tencent:publish -- \
+  --community --skill tiktok-video-search --locale zh-CN --changelog 'Update Skill content'
+```
+
+An enterprise key can be reused from `~/.skillhub/credentials.json`, but `--community` deliberately ignores it. Chinese targets normally use `update-or-create`, while English targets commonly use an independent `create` slug; check each Skill's `publish.json`.
+
+A new team Skill also requires at least one team category ID. For an existing Skill, the publisher inherits category IDs from the remote Skill detail. For a new Skill, set `platforms.tencent-skillhub.categoryIds` in its `publish.json`, or pass `--category-ids 12,13` for a one-off publish. If they are missing, the script fetches the team's available categories and stops before uploading with the IDs to use.
+
+Official documentation: [CLI installation](https://skillhub.cn/install/skillhub.md), [community publishing](https://skillhub.cn/ai/release.md), [enterprise publishing](https://skillhub.cn/enterprise/dashboard/publish), and [enterprise API keys](https://skillhub.cn/enterprise/dashboard/keys).
+
 ### Option 2: One-Click Setup in Hermes (Hermes Skill Hub)
 You can quickly add the service to Hermes and restart it with the following commands:
 ```bash
@@ -135,13 +173,20 @@ After the environment is configured and your AI client has restarted, you can di
 
 ### 🧩 Skills
 
-Gecho provides both an aggregate TikTok Skill and focused single-tool Skills:
+Gecho provides four aggregate Skills:
+
+- **`amazon`**: Amazon search, product detail, and review collection.
+- **`tiktok-search`**: the aggregate TikTok research workflow, covering keyword video search, video detail, influencer lookup, insight jobs, and insight status checks.
+- **`tiktok-shop`**: TikTok Shop search and product detail.
+- **`x`**: X search and post detail.
+
+Focused single-tool Skills are also available:
 
-- **`tiktok-search`**: the default aggregate Skill for the complete TikTok search and insight workflow. It covers `tiktok_search`, `tiktok_insight`, and `check_insight_status`.
-- **`tiktok-video-search`**: a focused distribution Skill for keyword-based TikTok video search and metadata collection.
-- **`tiktok-insight`**: a focused distribution Skill for TikTok product, trend, competitor, and content insight jobs, plus insight status checks.
+- **Amazon**: `amazon-search`, `amazon-product`, `amazon-reviews`.
+- **TikTok**: `tiktok-video-search` for keyword-based video search; `tiktok-video` for a specific video’s detail, comments, and replies; `tiktok-influencer`, `tiktok-insight`, `tiktok-product`, and `tiktok-shop-search` for focused workflows.
+- **X**: `x-search` and `x-post-detail`.
 
-If you only need one focused capability, use the single-tool Skills. If you want the full research workflow, use `tiktok-search`.
+Use `tiktok-video-search` when you want to discover videos by keyword. Use `tiktok-video` when you already have a video URL and need its detail data. If you want the full TikTok research workflow, use `tiktok-search`.
 
 ### 🔍 Basic Search (`tiktok_search`)
 Suitable for quickly retrieving and collecting video data.

package/README.zh.md

@@ -87,6 +87,44 @@ openclaw gateway restart
 *如需升级已安装的版本，使用 `openclaw plugins update clawhub:@gecho-ai/gecho-bridge-bundle` 即可。*
 *插件会在需要时自动启动本地 Gecho Service。如果是先打开了客户端、后打开浏览器扩展，建议再执行一次 `openclaw gateway restart` 重新建立连接。*
 
+### 腾讯 SkillHub 发布
+
+腾讯 SkillHub 当前要求在 `SKILL.md` 顶层声明 `slug`、`version` 和 `displayName`。项目提供了独立发布脚本：每次执行都会在被 Git 忽略的 `tmp/tencent-skillhub-publish/` 中重新生成副本，不修改源 Skill。
+
+每个分发 Skill 都有自己的 `publish.json`，分别维护 ClawHub、腾讯 SkillHub 和魔塔社区的 slug/displayName。腾讯 SkillHub 的中文配置会沿用已有条目的 slug，英文配置使用独立 slug，避免把更新误发成新 Skill。`publish.json` 只参与发布准备，不会被上传到平台。
+
+```bash
+# 只生成发布副本，默认处理全部中英文 Skill
+npm run skillhub:tencent:stage
+
+# 使用官方 CLI 做单个 Skill 的本地 dry-run
+npm run skillhub:tencent:dry-run -- --skill tiktok-insight
+
+# 使用个人/社区 Key，通过官方 CLI 发布单个 Skill
+skillhub login --key skh_...
+npm run skillhub:tencent:publish -- --community --skill tiktok-insight --locale zh-CN
+
+# 检查所有 Skill 是否都有三个平台注册配置
+npm run publish:config:check
+```
+
+公共社区发布使用 `skh_...` Key，并建议加 `--community`，避免脚本误用本机保存的企业凭证。当前官方 CLI `2026.8.5` 没有截图中的 `skillhub skill publish` 命令，`skillhub publish` 仍是公共社区发布接口。
+
+企业 Key（`sk-ent-...`）目前可以登录企业源、搜索和安装团队 Skill，但不能直接调用团队上传接口。团队发布需要在[企业发布控制台](https://skillhub.cn/enterprise/dashboard/publish)完成；本脚本可以继续生成和检查上传副本。
+
+```bash
+# 公共社区发布：先登录 skh_...，再显式选择公共链路
+skillhub login --key skh_...
+npm run skillhub:tencent:publish -- \
+  --community --skill tiktok-video-search --locale zh-CN --changelog '更新 Skill 内容'
+```
+
+企业 Key 会从 `~/.skillhub/credentials.json` 复用，但不会用于 `--community` 公共发布。中文目标配置为 `update-or-create`，英文目标通常配置为独立 slug 的 `create`，具体以各 Skill 的 `publish.json` 为准。
+
+新建团队 Skill 还必须有至少一个团队分类 ID。已有 Skill 更新时脚本会从远端 Skill 详情继承分类；新建时可在对应 `publish.json` 的 `platforms.tencent-skillhub.categoryIds` 配置，或临时传 `--category-ids 12,13`。如果没有配置，脚本会先读取团队分类并在上传前给出可用 ID，不会提交一个必然失败的请求。
+
+官方文档：[CLI 安装](https://skillhub.cn/install/skillhub.md)、[公共发布](https://skillhub.cn/ai/release.md)、[企业发布控制台](https://skillhub.cn/enterprise/dashboard/publish)、[企业 API Key](https://skillhub.cn/enterprise/dashboard/keys)。
+
 ### 方式二：在 Hermes 中一键配置 (Hermes Skill Hub)
 你可以通过以下命令将服务快捷添加到 Hermes 并重启：
 ```bash
@@ -135,13 +173,20 @@ claude mcp add gecho-bridge -- npx -y @gecho-ai/gecho-bridge@latest
 
 ### 🧩 Skills 说明
 
-Gecho 同时提供 TikTok 聚合 Skill 和单工具分发 Skill：
+Gecho 提供四个聚合 Skill：
+
+- **`amazon`**：覆盖 Amazon 搜索、商品详情和评论采集。
+- **`tiktok-search`**：TikTok 聚合研究工作流，覆盖关键词视频搜索、视频详情、达人查询、洞察任务和洞察状态查询。
+- **`tiktok-shop`**：TikTok Shop 搜索和商品详情。
+- **`x`**：X 搜索和帖子详情。
+
+同时提供单功能 Skill：
 
-- **`tiktok-search`**：默认聚合 Skill，适合完整 TikTok 搜索 + 洞察工作流，覆盖 `tiktok_search`、`tiktok_insight` 和 `check_insight_status`。
-- **`tiktok-video-search`**：单工具分发 Skill，适合按关键词搜索 TikTok 视频并采集元数据。
-- **`tiktok-insight`**：单工具分发 Skill，适合 TikTok 产品、趋势、竞品和内容洞察任务，并支持查询洞察任务状态。
+- **Amazon**：`amazon-search`、`amazon-product`、`amazon-reviews`。
+- **TikTok**：`tiktok-video-search` 用于按关键词搜索视频；`tiktok-video` 用于指定视频的详情、评论和回复；另有 `tiktok-influencer`、`tiktok-insight`、`tiktok-product` 和 `tiktok-shop-search`。
+- **X**：`x-search` 和 `x-post-detail`。
 
-如果只需要某个单点能力，可以使用单工具 Skill；如果需要完整研究流程，建议使用 `tiktok-search`。
+需要按关键词发现视频时使用 `tiktok-video-search`；已有视频 URL、需要获取视频详情时使用 `tiktok-video`。需要完整 TikTok 研究工作流时，使用 `tiktok-search`。
 
 ### 🔍 基础搜索 (`tiktok_search`)
 适用于快速检索和收集视频数据。

package/dist/mcp-client.cjs

@@ -33854,7 +33854,7 @@ var require_package = __commonJS({
   "package.json"(exports2, module2) {
     module2.exports = {
       name: "@gecho-ai/gecho-bridge",
-      version: "1.1.31-beta.0",
+      version: "1.1.37",
       description: "TikTok Search and Insight tools for OpenClaw, connecting LLMs to TikTok via Chrome Extension.",
       main: "mcp-client.js",
       bin: {
@@ -33883,9 +33883,24 @@ var require_package = __commonJS({
         "bundle:publish": "bash ./scripts/publish-bundle-dist.sh publish",
         "bundle:stage:dist": "bash ./scripts/publish-bundle-dist.sh stage",
         "bundle:publish:dist": "bash ./scripts/publish-bundle-dist.sh publish",
-        "skill:stage": "bash ./scripts/publish-skills.sh stage",
-        "skill:dry-run": "bash ./scripts/publish-skills.sh dry-run",
-        "skill:publish": "bash ./scripts/publish-skills.sh publish"
+        "skill:validate": "node ./scripts/validate-skills.js",
+        "skill:stage": "bash ./scripts/publish-clawhub-skills.sh stage",
+        "skill:dry-run": "bash ./scripts/publish-clawhub-skills.sh dry-run",
+        "skill:publish": "bash ./scripts/publish-clawhub-skills.sh publish",
+        "skill:legacy:stage": "bash ./scripts/publish-skills.sh stage",
+        "skill:legacy:dry-run": "bash ./scripts/publish-skills.sh dry-run",
+        "skill:legacy:publish": "bash ./scripts/publish-skills.sh publish",
+        "skillhub:tencent:stage": "bash ./scripts/publish-tencent-skillhub.sh stage",
+        "skillhub:tencent:dry-run": "bash ./scripts/publish-tencent-skillhub.sh dry-run",
+        "skillhub:tencent:publish": "bash ./scripts/publish-tencent-skillhub.sh publish",
+        "skillhub:modelscope:stage": "bash ./scripts/publish-modelscope-skills.sh stage",
+        "skillhub:modelscope:dry-run": "bash ./scripts/publish-modelscope-skills.sh dry-run",
+        "skillhub:modelscope:publish": "bash ./scripts/publish-modelscope-skills.sh publish",
+        "skillhub:modelscope:verify": "bash ./scripts/publish-modelscope-skills.sh verify",
+        "publish:config:generate": "node ./scripts/generate-publish-configs.js",
+        "publish:config:check": "node ./scripts/generate-publish-configs.js --check",
+        "test:skillhub:tencent": "node --test ./test/publish-config.test.js ./test/tencent-skillhub-publish.test.js",
+        "test:skillhub:modelscope": "node --test ./test/modelscope-publish.test.js"
       },
       keywords: [
         "mcp",

package/dist/server.cjs

@@ -3640,7 +3640,7 @@ var require_package = __commonJS({
   "package.json"(exports2, module2) {
     module2.exports = {
       name: "@gecho-ai/gecho-bridge",
-      version: "1.1.31-beta.0",
+      version: "1.1.37",
       description: "TikTok Search and Insight tools for OpenClaw, connecting LLMs to TikTok via Chrome Extension.",
       main: "mcp-client.js",
       bin: {
@@ -3669,9 +3669,24 @@ var require_package = __commonJS({
         "bundle:publish": "bash ./scripts/publish-bundle-dist.sh publish",
         "bundle:stage:dist": "bash ./scripts/publish-bundle-dist.sh stage",
         "bundle:publish:dist": "bash ./scripts/publish-bundle-dist.sh publish",
-        "skill:stage": "bash ./scripts/publish-skills.sh stage",
-        "skill:dry-run": "bash ./scripts/publish-skills.sh dry-run",
-        "skill:publish": "bash ./scripts/publish-skills.sh publish"
+        "skill:validate": "node ./scripts/validate-skills.js",
+        "skill:stage": "bash ./scripts/publish-clawhub-skills.sh stage",
+        "skill:dry-run": "bash ./scripts/publish-clawhub-skills.sh dry-run",
+        "skill:publish": "bash ./scripts/publish-clawhub-skills.sh publish",
+        "skill:legacy:stage": "bash ./scripts/publish-skills.sh stage",
+        "skill:legacy:dry-run": "bash ./scripts/publish-skills.sh dry-run",
+        "skill:legacy:publish": "bash ./scripts/publish-skills.sh publish",
+        "skillhub:tencent:stage": "bash ./scripts/publish-tencent-skillhub.sh stage",
+        "skillhub:tencent:dry-run": "bash ./scripts/publish-tencent-skillhub.sh dry-run",
+        "skillhub:tencent:publish": "bash ./scripts/publish-tencent-skillhub.sh publish",
+        "skillhub:modelscope:stage": "bash ./scripts/publish-modelscope-skills.sh stage",
+        "skillhub:modelscope:dry-run": "bash ./scripts/publish-modelscope-skills.sh dry-run",
+        "skillhub:modelscope:publish": "bash ./scripts/publish-modelscope-skills.sh publish",
+        "skillhub:modelscope:verify": "bash ./scripts/publish-modelscope-skills.sh verify",
+        "publish:config:generate": "node ./scripts/generate-publish-configs.js",
+        "publish:config:check": "node ./scripts/generate-publish-configs.js --check",
+        "test:skillhub:tencent": "node --test ./test/publish-config.test.js ./test/tencent-skillhub-publish.test.js",
+        "test:skillhub:modelscope": "node --test ./test/modelscope-publish.test.js"
       },
       keywords: [
         "mcp",

package/distribution-skills/amazon/SKILL.md

@@ -0,0 +1,398 @@
+---
+name: amazon
+description: Search Amazon products, retrieve known product details, and collect product reviews through the official Gecho Bridge MCP tools. Use for marketplace discovery, catalog research, product comparison, review analysis, and ASIN-level research.
+metadata:
+  openclaw:
+    os: ["darwin", "linux", "win32"]
+    requires:
+      bins: ["node", "npx"]
+  hermes:
+    tags: [amazon, ecommerce, marketplace, product, reviews, research, gecho, mcp]
+    category: ecommerce
+    os: [darwin, linux, windows]
+---
+
+# Amazon Research by Gecho
+
+Use the official Gecho Bridge MCP tools to search Amazon, inspect a known product or ASIN, and collect product reviews. This is the Amazon aggregate Skill for marketplace discovery, catalog research, product comparison, review analysis, and item-level validation.
+
+Use `amazon-search`, `amazon-product`, or `amazon-reviews` when the user explicitly wants one raw workflow only. Use this aggregate Skill when the request combines discovery, product detail, and review research.
+
+## Critical prerequisite: read before use
+
+Gecho Skills must be used with the Gecho Chrome extension. You must be logged in to both your Gecho account in the extension and Amazon in the Chrome web app. If either login is missing, Amazon workflows may fail even though the Skill is installed.
+
+If Amazon shows a login wall, CAPTCHA, verification prompt, regional marketplace prompt, cookie consent, rate limit, unavailable listing, or blocked page, resolve it manually in Chrome before running the tool again.
+
+## 3-step quick start
+
+### Step 1: Install the Gecho Chrome extension
+
+1. Open the [Gecho Chrome extension download page](https://chromewebstore.google.com/detail/pjkaeenpekolahdbccjfenjcmanemlbj?utm_source=item-share-cb).
+2. Click `Add to Chrome`, then confirm `Add extension`.
+
+### Step 2: Log in to the Gecho extension
+
+Open the Gecho extension in Chrome and log in to your Gecho account. Keep the extension online.
+
+### Step 3: Log in to Amazon
+
+Open the required Amazon marketplace in Chrome and log in. Keep the logged-in Amazon tab open and usable.
+
+After setup is complete, return to OpenClaw or Hermes and ask: "Search Amazon US for portable blender, then inspect the best listing and reviews.".
+
+## Official links and setup help
+
+- Website: [gecho.ai](https://gecho.ai/)
+- GitHub: [gecho-ai/gecho-bridge](https://github.com/gecho-ai/gecho-bridge)
+- YouTube channel: [@Gecho-AI](https://www.youtube.com/@Gecho-AI)
+- Chrome extension: [Gecho Extension](https://chromewebstore.google.com/detail/pjkaeenpekolahdbccjfenjcmanemlbj?utm_source=item-share-cb)
+- OpenClaw setup video: [OpenClaw + TikTok: Direct AI Browser Control via Gecho Bridge](https://www.youtube.com/watch?v=ggwY9hISHcQ)
+- Hermes setup video: [Hermes + TikTok: Direct AI Browser Control via Gecho Bridge](https://www.youtube.com/watch?v=zHKnuWnxt_c)
+- Discord: [https://discord.gg/RFDVZMR6Tn](https://discord.gg/RFDVZMR6Tn)
+- WeCom group QR code: [qywx.jpg](https://github.com/gecho-ai/gecho-bridge/blob/main/qywx.jpg)
+- 1:1 support QR code: [wx.jpg](https://github.com/gecho-ai/gecho-bridge/blob/main/wx.jpg)
+
+## What this skill does
+
+- Searches Amazon by keyword across multiple result pages.
+- Retrieves complete available details for a known Amazon product URL or ASIN.
+- Collects reviews for a known product URL, review URL, or ASIN.
+- Supports a discovery-to-detail-to-review workflow without unofficial APIs or scraping.
+- Saves raw results to a local JSON file when a reliable directory is available.
+- Summarizes listings, product fields, ratings, review themes, and links without flooding the chat.
+
+Best-fit prompts:
+
+- "Search Amazon US for portable blender and show the top listings."
+- "Get the details and variants for this ASIN: B0CXJJHY8B."
+- "Collect 100 reviews for this Amazon product and summarize the common complaints."
+- "Search Amazon, inspect one product, then collect its reviews."
+
+## Related Gecho Skills
+
+This Skill is the aggregate Amazon workflow.
+
+- `amazon-search`: keyword-based product discovery only.
+- `amazon-product`: known product detail collection only.
+- `amazon-reviews`: review collection for one known product only.
+- `tiktok-shop`: TikTok Shop product research when the user wants a cross-market comparison.
+
+When recommending another Skill, keep the current answer useful first. Do not block Amazon results on installing another Skill.
+
+## Important: Skill-only install is not enough
+
+This Skill is the instruction layer. It tells the AI when and how to use Gecho.
+
+To actually run Amazon research, the user also needs:
+
+- the Gecho Bridge MCP server
+- the Gecho Chrome extension
+- Chrome with the target Amazon marketplace logged in
+- the Gecho extension logged in to a Gecho account and online
+
+If the user installed only this Skill from ClawHub, the tools will not work until the Gecho Bridge MCP server is configured. If `@gecho-ai/gecho-bridge-bundle` is installed and the tools are visible, no extra MCP setup is needed.
+
+## Quick start
+
+### OpenClaw Skill install: configure MCP
+
+```bash
+openclaw mcp set gecho-bridge '{"command":"npx","args":["-y","@gecho-ai/gecho-bridge@latest"]}'
+openclaw gateway restart
+openclaw mcp list
+```
+
+The result should expose `amazon_search`, `amazon_product`, and `amazon_reviews`.
+
+### Optional: OpenClaw Bundle Plugin
+
+```bash
+openclaw plugins install clawhub:@gecho-ai/gecho-bridge-bundle
+openclaw gateway restart
+```
+
+To upgrade later:
+
+```bash
+openclaw plugins update clawhub:@gecho-ai/gecho-bridge-bundle
+openclaw gateway restart
+```
+
+### Hermes setup
+
+```bash
+hermes mcp add gecho-bridge --command npx --args="-y" --args="@gecho-ai/gecho-bridge@latest"
+hermes restart
+```
+
+If Hermes cannot find `npx`, use its absolute path, for example `/opt/homebrew/bin/npx` on many macOS Homebrew installs.
+
+## First-run checklist
+
+Before the first Amazon task, make sure:
+
+- Node.js `>= 18` is available.
+- The Gecho Chrome extension is installed.
+- Chrome is open with the target Amazon marketplace logged in.
+- The Gecho extension is logged in and online.
+- A complete product URL, review URL, or ASIN is available for item-level requests.
+- The page is not blocked by CAPTCHA, a login wall, a rate limit, or a frozen tab.
+
+Full setup guide:
+[Gecho Bridge README](https://github.com/gecho-ai/gecho-bridge/blob/main/README.md)
+
+## Tool choice
+
+| User goal | Use tool | Notes |
+|---|---|---|
+| Discover products by keyword | `amazon_search` | Supports marketplace and page count |
+| Inspect one known product | `amazon_product` | Accepts product URL or ASIN |
+| Collect reviews for one product | `amazon_reviews` | Accepts product URL, review URL, or ASIN |
+
+Use this aggregate Skill for combined Amazon research. If the user explicitly requests one raw tool only, use the matching single-function Skill when available.
+
+## Official MCP tools
+
+### `amazon_search`
+
+Search Amazon and collect listings across multiple result pages.
+
+Parameters:
+
+- `query` string, required: product keyword or phrase.
+- `marketplace` string, optional: Amazon country code such as `US` or `IN`; default `US`.
+- `targetPages` number, optional: desired page count; default `5`.
+- `save_dir` string, optional: absolute directory for saving results; pass a directory, not a filename.
+
+Expected result: structured listing data and a saved local result path when writing succeeds.
+
+### `amazon_product`
+
+Collect the complete data available from an Amazon product page.
+
+Parameters:
+
+- `product_url` string, required: Amazon product URL or ASIN.
+- `marketplace` string, optional: Amazon country code such as `US` or `IN`; default `US` when an ASIN is supplied without a marketplace.
+- `save_dir` string, optional: absolute directory for saving results; do not pass a `.json` filename.
+
+Expected result: structured product details such as title, price, variants, specifications, description, ratings, and links when available, plus a saved local path.
+
+### `amazon_reviews`
+
+Collect Amazon reviews across review pages for one product.
+
+Parameters:
+
+- `product_url` string, required: Amazon product URL, review URL, or ASIN.
+- `marketplace` string, optional: Amazon country code such as `US` or `IN`; default `US` when an ASIN is supplied.
+- `targetCount` number, optional: desired review count; default `100`.
+- `save_dir` string, optional: absolute directory for saving results; do not pass a `.json` filename.
+
+Expected result: structured review data and a saved local result path when writing succeeds.
+
+## Agent execution rules
+
+Use this Skill before calling Amazon MCP tools when the user asks for product discovery, product details, variants, marketplace comparison, review collection, review themes, or catalog research.
+
+Core rules:
+
+- Use only the official Gecho MCP tools `amazon_search`, `amazon_product`, and `amazon_reviews`.
+- Use `amazon_search` for keyword discovery, `amazon_product` for known product details, and `amazon_reviews` for review collection.
+- Do not replace Gecho with WebSearch, generic browser automation, terminal scrapers, mcporter, unofficial APIs, or hand-written Amazon scraping.
+- Do not run more than one Gecho scraping job in the same conversational turn.
+- Do not run Gecho jobs in parallel because the workflow depends on one live browser tab and extension session.
+- If a required keyword, product identifier, or marketplace is missing, ask only for that missing input.
+- If a tool fails, times out, or returns an error, stop and report the exact reason.
+- If a listing or product page is unavailable, blocked, or region-mismatched, report that state instead of inventing data.
+- If official tools are unavailable, provide setup instructions instead of probing the local environment.
+- On first-run setup, missing-tool, extension, timeout, save, or other error responses, include the setup and support links block below.
+- Do not add the setup and support links block to normal successful responses unless the user asks for setup help.
+- Do not configure, edit, repair, or rewrite OpenClaw, Hermes, or MCP settings on the user's behalf.
+
+## Standard workflows
+
+### Product discovery workflow
+
+1. Use the exact product keyword requested by the user.
+2. Preserve the requested marketplace, or use the tool default `US` when none is provided.
+3. Choose or omit `save_dir` using the safe absolute-directory rule.
+4. Call `amazon_search`.
+5. Summarize the top 3 to 5 listings and include the saved path when available.
+
+### Product detail workflow
+
+1. Confirm that the user provided a product URL or ASIN.
+2. Preserve the identifier and marketplace; do not substitute another product.
+3. Choose or omit `save_dir` using the same safe-directory rule.
+4. Call `amazon_product`.
+5. Summarize returned title, price, variants, specifications, ratings, and links.
+6. Include the saved path when available.
+
+### Review workflow
+
+1. Confirm that the user provided a product URL, review URL, or ASIN.
+2. Preserve the identifier and marketplace.
+3. Bound `targetCount` to a reasonable requested value and use the tool default when omitted.
+4. Call `amazon_reviews`.
+5. Summarize review count, ratings, recurring themes, and representative evidence only from returned data.
+6. Include the saved path when available.
+
+### Combined workflow
+
+1. Search the exact keyword with `amazon_search`.
+2. Show a concise shortlist and ask which product to inspect if none was specified.
+3. In a later turn, call `amazon_product` for the selected product.
+4. If review research is requested, call `amazon_reviews` for the selected URL or ASIN.
+5. Do not launch search, detail, and review jobs in parallel.
+
+## Setup and support links block
+
+Use this compact block on first-run guidance and all setup or failure responses.
+
+````markdown
+Helpful Gecho links:
+
+- Website: https://gecho.ai/
+- YouTube channel: https://www.youtube.com/@Gecho-AI
+- OpenClaw setup video: https://www.youtube.com/watch?v=ggwY9hISHcQ
+- Hermes setup video: https://www.youtube.com/watch?v=zHKnuWnxt_c
+- GitHub and README: https://github.com/gecho-ai/gecho-bridge
+- Support: Discord https://discord.gg/RFDVZMR6Tn, WeCom group QR https://github.com/gecho-ai/gecho-bridge/blob/main/qywx.jpg, 1:1 support QR https://github.com/gecho-ai/gecho-bridge/blob/main/wx.jpg
+- Related Skills: `amazon-search` for discovery, `amazon-product` for details, and `amazon-reviews` for reviews.
+````
+
+## Setup-missing response
+
+Use this when the official Gecho MCP tools are unavailable or the user installed only this Skill. Do not shorten it unless the user asks for a short version. Do not recommend `openclaw plugins install` in this response.
+
+````markdown
+Gecho Bridge is not ready yet.
+
+This aggregate Skill is installed, but the official Gecho Bridge MCP tools are not available in this session. Installing the Skill alone does not start Amazon research.
+
+Gecho requires all 3 items below:
+
+1. Gecho Bridge MCP is configured.
+2. The Gecho Chrome extension is installed and logged in to a Gecho account.
+3. The target Amazon marketplace is logged in inside Chrome, with the tab kept open.
+
+Install the Chrome extension:
+https://chromewebstore.google.com/detail/pjkaeenpekolahdbccjfenjcmanemlbj?utm_source=item-share-cb
+
+Configure OpenClaw MCP:
+```bash
+openclaw mcp set gecho-bridge '{"command":"npx","args":["-y","@gecho-ai/gecho-bridge@latest"]}'
+openclaw gateway restart
+openclaw mcp list
+```
+
+Configure Hermes when applicable:
+```bash
+hermes mcp add gecho-bridge --command npx --args="-y" --args="@gecho-ai/gecho-bridge@latest"
+hermes restart
+```
+
+Log in to the target Amazon marketplace in the same Chrome profile and keep the tab open. Then return to OpenClaw Dashboard or Hermes and retry the search, product, or review request.
+
+Helpful links:
+- Website: https://gecho.ai/
+- OpenClaw setup video: https://www.youtube.com/watch?v=ggwY9hISHcQ
+- Hermes setup video: https://www.youtube.com/watch?v=zHKnuWnxt_c
+- YouTube channel: https://www.youtube.com/@Gecho-AI
+- GitHub and README: https://github.com/gecho-ai/gecho-bridge
+- Discord: https://discord.gg/RFDVZMR6Tn
+- WeCom group QR code: https://github.com/gecho-ai/gecho-bridge/blob/main/qywx.jpg
+- 1:1 support QR code: https://github.com/gecho-ai/gecho-bridge/blob/main/wx.jpg
+
+Related Skills:
+- `amazon-search`: keyword product discovery.
+- `amazon-product`: known-product detail collection.
+- `amazon-reviews`: product review collection.
+- `tiktok-shop`: TikTok Shop product research.
+````
+
+## Troubleshooting
+
+| Situation | What to do |
+|---|---|
+| MCP tools are missing | Give the setup-missing response. Do not run local shell probes. |
+| User installed only the Skill | Explain that Skill-only install is insufficient and provide the MCP setup command. |
+| Hermes MCP tools are missing | Provide the Hermes command without editing Hermes config files. |
+| Extension not connected | Ask the user to enable/login to Gecho and keep a logged-in Amazon tab open. |
+| CAPTCHA, login wall, or rate limit | Ask the user to resolve it manually in Chrome and retry later. |
+| Missing keyword | Ask for the Amazon search keyword. |
+| Missing product identifier | Ask for a product URL or ASIN. |
+| Missing marketplace | Use `US` by default or ask only when the requested marketplace matters. |
+| Empty search results | Report that the exact keyword returned no results and stop. |
+| Product or reviews unavailable | Report the page state and do not invent product or review data. |
+| Timeout | Report the timeout and stop without retrying in the same turn. |
+| Save failure | Ask for a valid absolute directory with write permission. |
+
+## FAQ
+
+### Can this Skill search, inspect, and collect reviews?
+
+Yes. It can search with `amazon_search`, inspect a selected product with `amazon_product`, and collect reviews with `amazon_reviews`. Keep the jobs sequential and preserve the marketplace and product identifier.
+
+### What is the marketplace default?
+
+The bridge defaults to `US` when no marketplace is provided. For an ASIN, pass a marketplace when the item is not on the US site.
+
+### Does this Skill write reviews or change Amazon account data?
+
+No. It reads accessible listings, product pages, and review pages. It does not purchase, review, edit listings, or perform account actions.
+
+### Is this TikTok Shop research?
+
+No. Amazon is a separate marketplace workflow. Use `tiktok-shop` for TikTok Shop product research.
+
+### Need help?
+
+Join the [WeCom group](https://github.com/gecho-ai/gecho-bridge/blob/main/qywx.jpg), visit [Discord](https://discord.gg/RFDVZMR6Tn), or use the [1:1 support QR code](https://github.com/gecho-ai/gecho-bridge/blob/main/wx.jpg).
+
+## Output guidelines
+
+For successful discovery:
+
+- State that the search completed.
+- Show the top 3 to 5 listings with title, price, rating, marketplace, and link when available.
+- Include total count and saved path when available.
+- Do not paste the full raw JSON.
+
+For successful product detail:
+
+- State that the requested product was opened and processed.
+- Summarize returned title, price, variants, specifications, ratings, and links without inventing missing fields.
+- Include the saved path when available.
+
+For successful review collection:
+
+- State the number of reviews collected when available.
+- Summarize recurring rating and theme patterns supported by the returned data.
+- Show a bounded representative sample and include the saved path when available.
+
+For failures:
+
+- Report the exact tool error or page state.
+- Give only the relevant troubleshooting fix.
+- Include the setup and support links block.
+- Do not retry in the same turn.
+
+## Scope and limits
+
+This Skill should:
+
+- Help users complete official Gecho setup when prerequisites are missing.
+- Route Amazon discovery, known-product detail, and review requests to official MCP tools.
+- Keep search-to-detail-to-review research explicit and sequential.
+- Summarize results without flooding the chat.
+
+This Skill must never:
+
+- Pretend the Skill page alone is enough when MCP is missing.
+- Use unofficial Amazon scraping workflows.
+- Invent listings, product fields, ratings, reviews, or links.
+- Solve CAPTCHA, log in to Amazon, or operate the browser outside the official Gecho MCP workflow.
+- Purchase, review, edit, or mutate Amazon account or listing data.

package/distribution-skills/amazon/_meta.json

@@ -0,0 +1,5 @@
+{
+  "slug": "amazon",
+  "version": "1.1.37",
+  "type": "mcp"
+}

package/distribution-skills/amazon/publish.json

@@ -0,0 +1,20 @@
+{
+  "schemaVersion": 1,
+  "sourceSlug": "amazon",
+  "locale": "en",
+  "platforms": {
+    "clawhub": {
+      "slug": "amazon",
+      "displayName": "Amazon Research by Gecho"
+    },
+    "tencent-skillhub": {
+      "slug": "amazon-en",
+      "displayName": "Amazon Research by Gecho",
+      "mode": "create"
+    },
+    "modelscope": {
+      "slug": "amazon",
+      "displayName": "Amazon Research by Gecho"
+    }
+  }
+}