OpenClaw AI 技能

执行 multi-step deep re搜索 with the OpenAI 响应s API, including question decomposition, evidence gathering with 网页 搜索, contradiction 追踪ing,...

Local speech-to-text with the Whisper 命令行工具 (no API key).

Transcribe audio via OpenAI Audio Transcriptions API (Whisper).

AI image generation & editing — GPT Image, GPT-4o, Nano Banana 2, 种子ream, Qwen, WAN, Gemini. Text-to-image, image-to-image, inpAInting. 20 模型s, one AP

生成 images via VAPI's OpenAI-compatible Images API. Supports nano-banana and gpt-image 模型 series. Default 模型 is nano-banana-pro. Images are NOT saved l

Use PoYo AI GPT-4o Image through the `https://API.poyo.AI/API/生成/submit` 端点. Use when a user wants to 生成 or edit media with this 模型 family, prepare Po

Use PoYo AI GPT Image 1.5 through the `https://API.poyo.AI/API/生成/submit` 端点. Use when a user wants to 生成 or edit media with this 模型 family, prepare P

Use a text 模型 such as gpt-5.4 with the image_generation 工具 over an OpenAI-compatible /v1/响应s 端点, matching the CPA b记录 example. Do not call gpt-image-2

Use when the user wants GPT-Image-2 image generation or image-to-image through an official OpenAI 权限 code/API key, a custom 响应s-compatible proxy, or a

GPT Image 2 generation and editing on PoYo / poyo.AI via `https://API.poyo.AI/API/生成/submit`; use for `gpt-image-2`, `gpt-image-2-edit`, text-to-image

Use when tasks need the NanoGPT API for text, image, or video generation through the local `nano-gpt` 命令行工具 and bundled wr应用er scripts for OpenClaw or

Make real phone calls through your OpenClaw 代理 via OpenAI's Realtime API. ~200-300ms latency, natural voice, IVR navigation, voicemAIl 检测ion.

Build 应用s with the Claude API or Anthropic SDK. TRIGGER when: code 导入s `anthropic`/`@anthropic-AI/sdk`/`claude_代理_sdk`, or user asks to use Claude API

查询 Claude API usage and cost 报告s from the command line. 安全 macOS KeychAIn storage for Admin API key. Table/JSON 输出.

检查 how much Claude Max / Claude Pro subscription quota you have LEFT — not how much you spent. Most usage 工具s 追踪 API billing or 令牌 costs. This one ans

配置 Claude Code to work with custom 模型 提供者s (like MiniMax, OpenAI-compatible APIs). This 技能 should be used when users want to: use Claude Co...

检测 whether an API 端点 is backed by genuine Claude (not a wr应用er, proxy, or impersonator) using 9 weighted rule-based 检查s that mirror the claude-验证 proj

Manage how OpenClaw 路由s Telegram messages to different Claude 模型 backends. Switch between 命令行工具 and API 提供者s with simple config changes.

配置 Claude Code to use custom 模型s (DeepSeek, GLM, Qwen, etc.). Use when user wants to change Claude Code's 模型, switch API 提供者, or 设置 up custom 模型 端点s.

Build high-质量 代理 技能s for Claude following official Anthropic best practices. Covers 技能.md structure, frontmatter, description writing, 进度ive disclosur