Skills from clawhub.ai/zjianru

High-risk skill: monitors and runs local jobs and reports results via Telegram. It executes arbitrary configured commands with `subprocess.run`, reads `~/.openclaw/openclaw.json` for `botToken`, and posts to `https://api.telegram.org/bot{token}/sendMessage`.

Provides a multi-engine CLI web search and content extractor for AI agents. It reads API keys (`TAVILY_API_KEY`, `EXA_API_KEY`, `SERPER_API_KEY`, `SERPAPI_API_KEY`), instructs running `node {baseDir}/scripts/search.mjs`, and calls external endpoints such as `https://api.tavily.com/search` and `https://api.exa.ai/search`.

This skill orchestrates safe restarts of the OpenClaw Gateway, preserving context in `~/.openclaw/restart-guard/restart-context.md`, backing up `~/.openclaw/openclaw.json`, spawning a detached guardian, and sending notifications. It executes shell commands (`subprocess.run(..., shell=True)`), reads env vars like `GATEWAY_AUTH_TOKEN`/`TELEGRAM_BOT_TOKEN`, and makes network calls (e.g., `https://api.telegram.org/bot{token}/sendMessage`, `http://{host}:{port}/tools/invoke`), creating elevated risk.