发现优质的 AI Agent 技能
聚合 Claude Skills、LangChain、AutoGPT 等优质资源，助力开发者快速构建智能应用
Tool Orchestration Patterns Reference
Research-backed patterns inspired by NVIDIA ToolOrchestra, OpenAI Agents SDK, and multi-agent coordination research.
Quality Control Reference
Quality gates, code review process, and severity blocking rules. Enhanced with 2025 research on anti-sycophancy, heterogeneous teams, and OpenAI Agents SDK patterns.
OpenAI Agent Patterns Reference
Research-backed patterns from OpenAI's Agents SDK, Deep Research, and autonomous agent frameworks.
Create a secret
eas env:create --name OPENAI_API_KEY --value sk-xxx --environment production ```
May hit rate limits if too many calls
return openai.chat.completions.create(...) ``` **Correct (with rate limit):**
AI providers (one required for non-dry-run)
OPENAI_API_KEY: OpenAI provider key GEMINI_API_KEY: Google Gemini provider key ANTHROPIC_API_KEY: Anthropic Claude provider key OLLAMA_BASE_URL: Ollama base URL (optional, defaults to http://localhost:11434) TWITTERAPI_IO_KEY: twitterapi.io API key (preferred)
code-security-auditor
综合代码安全审计工具，集成 OWASP Top 10 检测、LLM 驱动分析、220+ 安全规则，对标 OpenAI Codex Security，完全本地运行
🔍 OpenAI Codex Security vs Code Security Auditor 深度对比
**对比日期**: 2026-03-07 **对比版本**: Codex Security (Research Preview) vs Code Security Auditor v1.0.0
Gemini Image Simple
Generate and edit images using **Gemini 3 Pro Image** via the OpenAI Python SDK and an OpenAI-compatible API endpoint.
OpenClaw Codex GPT-5.4 Enable Skill
把 `openai-codex/gpt-5.4` 用“配置层补丁”方式接入 OpenClaw，避免直接改安装目录或整包重编译。
OpenClaw Codex GPT-5.4 启用技能
这是一个可复用的 OpenClaw 技能，用来沉淀一套**低风险、可回滚、可验证**的流程：在**不重编译 OpenClaw** 的前提下，通过配置层补丁启用 `openai-codex/gpt-5.4`。
memclawz — Three-Speed Memory Skill
> **No API key required.** Unlike other memory solutions that need OpenAI/Google/Voyage API keys, memclawz runs entirely locally using embedded models. > > Install: `clawhub install yoniassia/memclawz`