减肥打卡记录

减肥进度追踪助手。用户告诉 AI 今天的体重，AI 自动帮您记录并生成漂亮的图表，清晰展示减肥进度。支持中英文界面，macOS/Windows/Linux 多平台使用。当用户说"记录体重"、"今天体重"、"减肥打卡"、"体重多少"时自动触发。

This skill appears to be what it says: a small local web app that stores weight records in weight_history.json and serves jianfei.html. Things to consider before installing: 1) The included setup.sh and the SKILL.md ask you to run shell commands — inspect setup.sh (you can see it) before running it. 2) The instructions start a Python HTTP server without an explicit bind host; by default that can be reachable from the local network. If you want it strictly local, run: python3 -m http.server <PORT> --bind 127.0.0.1. 3) The skill will read/write weight_history.json and config.json in the skill directory — back them up if they contain data you care about. 4) Stopping the server uses kill/netstat/fuser commands; those are normal but review them if you have sensitive long-running processes. If you want extra assurance, run the HTML locally from a sandboxed machine or inspect the full jianfei.html to verify it does not contact external endpoints (the provided files appear to only load local JSON).

Type: OpenClaw Skill Name: weight-tracker Version: 1.0.1 The skill bundle provides a legitimate weight tracking utility but requires the AI agent to perform high-risk operations, including executing shell commands to manage a local HTTP server and using process-killing commands (e.g., 'kill', 'taskkill', 'lsof') to manage ports. While these capabilities are plausibly needed for the stated purpose of rendering a local HTML dashboard, the instructions in SKILL.md and the complexity of the setup.sh script grant the agent broad system-level interaction capabilities that could be abused if the agent's context is compromised.