MiniMax Coding Plan

该 Skill 封装本地 bash/Python 包装脚本，提供 MiniMax Coding Plan 的原生网页搜索与图片理解两个工具，结果以 JSON 输出；凭据优先取 MINIMAX_API_KEY，否则读取 OpenClaw 授权配置中的 minimax-portal 记录。

yjli-new

热度 90.13

最近更新 2026-05-17 13:37:34

分类：搜索依赖：代码执行 / 读写文件 / 第三方 API应用场景：研究 / 数据处理

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add minimax-coding-plan

Metadata

name: minimax-coding-plan
description: MiniMax Coding Plan native web search and image understanding for OpenClaw. Use when the user specifically wants MiniMax-native search or image analysis, or when an active MiniMax plan should handle vision instead of the generic media stack.
metadata: {"openclaw":{"emoji":"🧠"}}

MiniMax Coding Plan

Use the local wrapper:

bash {baseDir}/scripts/minimax-plan.sh <tool> [args...]

Available tools:

web_search --query "..."
understand_image --prompt "..." --image-source /path/to/file-or-url

Recommended routing:

Use this skill when the user explicitly wants MiniMax-native search or MiniMax-native image understanding.
For ordinary web research, tavily-search is still the general-purpose default.
If the main minimax-portal text model path is not exposed as a reliable image route in the current OpenClaw setup, use understand_image.

Examples:

bash {baseDir}/scripts/minimax-plan.sh web_search --query "MiniMax M2.5 release note"
bash {baseDir}/scripts/minimax-plan.sh understand_image --prompt "Describe the UI in this screenshot" --image-source /tmp/screen.png
bash {baseDir}/scripts/minimax-plan.sh understand_image --prompt "Extract the visible text" --image-source https://example.com/image.png

Notes:

This wrapper first uses MINIMAX_API_KEY when present. Otherwise it looks for an existing minimax-portal OAuth profile in OPENCLAW_AUTH_PROFILES_JSON, OPENCLAW_AGENT_DIR, OPENCLAW_HOME, or ~/.openclaw.
Output is JSON. For image analysis, read content first. If base_resp.status_code != 0, explain the MiniMax API error plainly.
Supported image formats: JPEG, PNG, WebP.