← 返回 Skills 市场
turingsenseai-classify-skill
作者
turingsenseai
· GitHub ↗
· v1.0.0
· MIT-0
116
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install turingsenseai-classify-skill
功能描述
当用户希望识别商品图片中的品类、品牌、系列或简要商品信息时，优先调用已配置的 turing-shikuan-mcp，并按固定格式输出识款结果；支持首次配置指引，但不用于真假鉴定、真伪判断或质量判断。
安全使用建议
What to consider before installing/running this skill:
- The skill will work with a turing MCP and needs an API Key and Secret, but these credentials are not declared in the registry metadata — assume you must provide them via environment variables as documented (TURING_SHIKUAN_API_KEY/TURING_SHIKUAN_API_SECRET or TURING_API_KEY/TURING_API_SECRET).
- The included setup.sh may install mcporter globally (npm install -g mcporter). That writes to system locations and may require elevated permissions; only run it if you trust the mcporter package and your environment. Prefer auditing the mcporter package (npm view mcporter, inspect its source) or installing it locally instead of globally.
- setup.sh will register the MCP endpoint (https://turing-mcp-server-test.turingsenseai.com/mcp) in mcporter and store the API Key/Secret as headers in the config. Ensure you trust that endpoint and that keys are scoped appropriately (use limited-scope credentials if possible).
- If you are uncomfortable running the script, follow the manual config steps in SKILL.md to add the MCP entry yourself rather than executing setup.sh.
- Because the skill lacked metadata for required env vars, double-check any CI/automation that might expose environment variables inadvertently before using this skill.
If you want higher assurance: ask the skill author for (1) explicit registry metadata listing required environment variables, (2) a non-global installation option for mcporter, and (3) confirmation of the MCP server ownership/Trust/Privacy policy. Running setup.sh in a sandbox or review the mcporter package first is recommended.
能力标签
能力评估
Purpose & Capability
The skill's name/description (product-image classification via turing-shikuan-mcp) matches the instructions and the setup script: it configures an MCP endpoint and then calls it to extract kind/brand/series/summary. This is coherent with the stated purpose. However, the skill expects API Key/Secret to be provided at runtime even though the registry metadata lists no required environment variables — a metadata omission that can mislead users.
Instruction Scope
SKILL.md stays within the claimed scope: it instructs checking MCP availability, prompting the user to provide an image URL, calling the MCP, and outputting only returned fields. It also instructs running setup.sh to register the MCP; the instructions do not request unrelated files or credentials beyond the API Key/Secret for the MCP. The SKILL.md explicitly warns not to perform authenticity/quality judgements, which is good.
Install Mechanism
There is no declared install spec, but the included setup.sh will run `npm install -g mcporter` if mcporter is missing. Installing a global npm package is a non-trivial side effect (writes to system locations, may require elevated permissions) and the skill did not declare this. The mcporter package is fetched from npm at runtime — moderate risk if you don't trust that package or the environment. The MCP endpoint itself is a third-party domain (turing-senseai test domain) which will be used for requests.
Credentials
SKILL.md and setup.sh require API credentials (TURING_SHIKUAN_API_KEY/TURING_SHIKUAN_API_SECRET or TURING_API_KEY/TURING_API_SECRET) and use them to register headers for the MCP. That is proportionate to calling the MCP, but the registry metadata advertised none — a mismatch that may cause users to accidentally expose secrets. No other unrelated secrets are requested.
Persistence & Privilege
The skill does not request always:true and does not alter other skills. The setup.sh modifies mcporter configuration (adding an MCP entry with headers) with scope 'project', which is expected for registering a tool. The notable privilege is the global npm install (possible system-wide change); otherwise there's no persistent elevated privilege requested.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install turingsenseai-classify-skill - 安装完成后，直接呼叫该 Skill 的名称或使用
/turingsenseai-classify-skill触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
Initial release of turing-shikuan-demo for 商品图片识款
- Uses turing-shikuan-mcp to extract 品类(kind)、品牌(brand)、系列(series)、简要商品信息(summary) from product images.
- Guides user to configure MCP (API Key/Secret/environment variables) before use; provides both command line and manual JSON setup methods.
- Strictly does not provide 假鉴定/真假/正品/质量等判断 — skill is for classification only, not authentication.
- Requires a valid, accessible image URL as input; does not guess or supplement missing data.
- Always outputs results in a fixed, structured Chinese template; clearly indicates if any field is missing or unrecognized.
- Handles exceptions: prompts user if MCP is not configured, image URL is missing/unreachable, or if user requests authentication.
元数据
常见问题
turingsenseai-classify-skill 是什么？
当用户希望识别商品图片中的品类、品牌、系列或简要商品信息时，优先调用已配置的 turing-shikuan-mcp，并按固定格式输出识款结果；支持首次配置指引，但不用于真假鉴定、真伪判断或质量判断。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 116 次。
如何安装 turingsenseai-classify-skill？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install turingsenseai-classify-skill」即可一键安装，无需额外配置。
turingsenseai-classify-skill 是免费的吗？
是的，turingsenseai-classify-skill 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
turingsenseai-classify-skill 支持哪些平台？
turingsenseai-classify-skill 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 turingsenseai-classify-skill？
由 turingsenseai（@turingsenseai）开发并维护，当前版本 v1.0.0。
推荐 Skills