发现优质的 AI Agent 技能

Run agents and commands in **isolated VM environments** using Docker Desktop's sandbox feature. Each sandbox gets its own lightweight VM with filesystem isolation, network proxy controls, and workspace mounting via virtiofs.

Process tabular data (CSV, TSV, JSON, JSON Lines) using standard command-line tools and Python. No external dependencies required beyond Python 3.

Build, test, document, and debug HTTP APIs from the command line. Covers the full API lifecycle: scaffolding endpoints, testing with curl, generating OpenAPI docs, mocking services, and debugging.

Control WiiM and LinkPlay audio devices from the command line.

You have a new skill called `linux-desktop` that lets you control the Linux GUI. Work through each step below to learn how it works. Run all commands from the skill directory.

来自 openclaw/skills 技能

You are an expert at FFmpeg commands and the FFHub.io cloud transcoding API. Help users process video/audio files by generating the right FFmpeg command and executing it via the FFHub API.

- The user provides a Mealie base URL (e.g., `https://mealie.example.com`) and/or an API token and asks to add/import a recipe, create or modify a meal plan, fetch a shopping list, or query existing recipes. - The user wants to automate meal‑planning tasks from the command line or through a script.

Voice conversation with AI via Whisper STT + ElevenLabs TTS. Records audio, transcribes, generates spoken responses, stores transcripts. For v1: CLI-based commands only (no real-time conversation mode — that's v1.1). Focus on transcript storage/search + TTS/STT wrappers.

Local-first CRM for tracking leads, deals, follow-ups, and pipeline. Uses SQLite with WAL mode, CLI via Commander.

> ⛔ **HARD RULE: NEVER edit `openclaw.json` model fields directly.** > Always use this skill's commands. No exceptions. Ever.

This reference defines the expected command surface for the Python-first skill wrapper: