← Back to Skills Marketplace
MiniMax Vision
by
hongjiahao371-pixel
· GitHub ↗
· v1.0.0
· MIT-0
164
Downloads
0
Stars
2
Active Installs
1
Versions
Install in OpenClaw
/install minimax-vision
Description
MiniMax 图片理解 MCP - 自动识别用户发送的图片
Usage Guidance
This skill appears to do exactly what it says: call a locally installed mcporter MCP to analyze images dropped into ~/.openclaw/media/inbound. Before installing, confirm that the mcporter binary and the minimax MCP are from a trusted source (the skill simply executes that binary). Also remove or update the hardcoded '/Users/js/.local/bin' PATH modification in minimax_vision.py (it's a developer leftover and may not exist on your system). Finally, ensure the inbound media directory permissions are appropriate and that you trust any MCP service mcporter will contact.
Capability Analysis
Type: OpenClaw Skill
Name: minimax-vision
Version: 1.0.0
The skill is designed for image recognition using the MiniMax MCP, but minimax_vision.py contains a hardcoded absolute path to a specific user's directory (/Users/js/.local/bin) within the environment PATH configuration. While likely a developer oversight, this is a security risk and a vulnerability that could lead to the execution of unintended binaries if the environment matches the hardcoded string.
Capability Assessment
Purpose & Capability
Name/description (image understanding) match the runtime instructions and included Python script: both call 'mcporter call minimax.understand_image' on images in the inbound media directory. No unrelated credentials, binaries, or install steps are requested.
Instruction Scope
SKILL.md instructs the agent to detect images in ~/.openclaw/media/inbound and call mcporter; the Python file implements the same behavior. The instructions do not ask the agent to read unrelated files, exfiltrate data, or contact unexpected endpoints.
Install Mechanism
There is no install spec (instruction-only plus a small helper script), so nothing is downloaded or written at install time. This is low-risk and proportional for this functionality.
Credentials
The skill requires no secrets or extra environment variables. The only oddity is the helper script prepending a hardcoded path (/Users/js/.local/bin) to PATH before invoking mcporter — likely a developer leftover. It does not request or expose credentials, but the hardcoded path is platform- and user-specific and should be removed or made configurable.
Persistence & Privilege
The skill is not always-enabled and is user-invocable; it does not request persistent system privileges or attempt to modify other skills' configs. Autonomous invocation is allowed by platform default but is not set to 'always'.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install minimax-vision - After installation, invoke the skill by name or use
/minimax-vision - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
MiniMax Vision skill initial release.
- Adds automatic image recognition using MiniMax MCP.
- Supports user-initiated image analysis with common image file types.
- Integrates with mcporter for backend processing.
- Returns analyzed results directly to the user upon image upload.
Metadata
Frequently Asked Questions
What is MiniMax Vision?
MiniMax 图片理解 MCP - 自动识别用户发送的图片. It is an AI Agent Skill for Claude Code / OpenClaw, with 164 downloads so far.
How do I install MiniMax Vision?
Run "/install minimax-vision" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is MiniMax Vision free?
Yes, MiniMax Vision is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does MiniMax Vision support?
MiniMax Vision is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created MiniMax Vision?
It is built and maintained by hongjiahao371-pixel (@hongjiahao371-pixel); the current version is v1.0.0.
More Skills