← 返回 Skills 市场
hongjiahao371-pixel

MiniMax MCP Tools

作者 hongjiahao371-pixel · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
115
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install minimax-mcp-tools
功能描述
提供联网搜索和图片内容理解服务,支持关键词搜索和基于描述请求的图片分析。
安全使用建议
Before installing: (1) Confirm the MiniMax service and domain (https://platform.minimaxi.com and https://api.minimaxi.com) are legitimate and trustworthy. (2) Do NOT run curl | sh against https://astral.sh without inspecting the script first — download and review the installer contents locally, or obtain uvx from an official package repository. (3) Verify the provenance of the 'clawhub install minimax-mcp' package (who publishes it, is it signed?). (4) Treat the MINIMAX_API_KEY as a secret: create a least-privilege token if the service supports it, and store it in your platform's secret store rather than embedding it in plain text if possible. (5) Be aware the skill will read local image files you point it at; only supply images you are willing to transmit to the MiniMax API. (6) Ask the publisher for a homepage or source repository if you need more assurance — the registry metadata currently lacks links or author details.
功能分析
Type: OpenClaw Skill Name: minimax-mcp-tools Version: 1.0.0 The skill bundle provides documentation and configuration instructions for integrating MiniMax MCP tools (web search and image understanding) into OpenClaw. It directs users to install the legitimate 'uv' tool and configure an MCP server using a specific package (minimax-coding-plan-mcp). While the SKILL.md contains a hardcoded local file path (/Users/js/.local/bin/uvx) which is a configuration error for most users, there is no evidence of malicious intent, data exfiltration, or harmful prompt injection.
能力评估
Purpose & Capability
The described capabilities (web_search and understand_image) match the SKILL.md instructions: the skill uses a local MCP bridge (uvx) and a MiniMax API. However the skill metadata declares no required environment variables or primary credential while the SKILL.md explicitly requires a MINIMAX_API_KEY and MINIMAX_API_HOST in ~/.openclaw/mcp.json. That mismatch is unexpected and should have been reflected in the registry metadata.
Instruction Scope
The runtime instructions tell the user/agent to install uvx via a remote installer (curl -LsSf https://astral.sh/uv/install.sh | sh), to edit ~/.openclaw/mcp.json to add an API key and host, and to allow reading local image file paths for image analysis. Installing and running a remote install script and placing API keys in a local config are reasonable for this tool but expand scope beyond simple guidance; the remote installer step is high-risk if the domain/script is not verified. The SKILL.md also omits explicit guidance about safe handling of the API key (no mention of least-privilege keys or secret storage).
Install Mechanism
This is an instruction-only skill, so there is no automated install spec, but the documentation instructs running a curl | sh installer from https://astral.sh — a third-party URL. The evaluator cannot verify that URL or script contents; running such installers without review is risky. The rest of the installation (editing mcp.json, clawhub install minimax-mcp) references commands whose origin isn't documented; users should confirm those packages' provenance before running them.
Credentials
Metadata lists no required env vars or primary credential, yet SKILL.md instructs storing MINIMAX_API_KEY and MINIMAX_API_HOST in the MCP config. The need for a single API key is expected for this service, but it should be declared in metadata. There are no unrelated credentials requested, and local image file access is justified by the feature, but the missing declaration of the API key is an inconsistency that affects trust and automated permission checks.
Persistence & Privilege
The skill does not request always:true and does not request system-wide privileges. It asks the user to add an entry to the user's ~/.openclaw/mcp.json and to restart the OpenClaw gateway — standard for adding an MCP plugin. There is no request to modify other skills' configs or global system settings in the instructions.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install minimax-mcp-tools
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /minimax-mcp-tools 触发
  4. 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
初始版本
元数据
Slug minimax-mcp-tools
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

MiniMax MCP Tools 是什么?

提供联网搜索和图片内容理解服务,支持关键词搜索和基于描述请求的图片分析。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 115 次。

如何安装 MiniMax MCP Tools?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install minimax-mcp-tools」即可一键安装,无需额外配置。

MiniMax MCP Tools 是免费的吗?

是的,MiniMax MCP Tools 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

MiniMax MCP Tools 支持哪些平台?

MiniMax MCP Tools 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 MiniMax MCP Tools?

由 hongjiahao371-pixel(@hongjiahao371-pixel)开发并维护,当前版本 v1.0.0。

💬 留言讨论