skills.rest
jezweb
jezwebCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

openai-agents

Build AI agents with the OpenAI Agents SDK in JavaScript/TypeScript environments.

961|99|Updated Oct 20, 2025
Search Tags:#ai agents#tool calling#multi-agent#openai#human-in-the-loop#voice ai#guardrails
One-click install
npx skills add https://github.com/jezweb/claude-skills --skill openai-agents
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: openai-agents
Source: https://github.com/jezweb/claude-skills/tree/main/skills/openai-agents
Command: npx skills add https://github.com/jezweb/claude-skills --skill openai-agents

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @openai/agents, @openai/agents-realtime, zod, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill eliminates the complexity of building production-ready AI applications with text agents, voice agents, and multi-agent workflows, saving you weeks of development time.

Core Features & Use Cases

  • Text Agents: Create intelligent assistants with tools, structured outputs, and streaming responses.
  • Voice Agents: Build realtime conversational AI with WebRTC/WebSocket transports.
  • Use Case: Imagine you need to create a customer service system with automated routing between billing and technical specialists, complete with human approval for critical actions.

Quick Start

Create a basic text agent that can answer questions about weather using the get_weather tool with city and units parameters.

Error Prevention & Automation

  • 9 Documented Errors: Prevents common pitfalls like Zod schema type errors, MCP tracing failures, and infinite loops.

Frequently Asked Questions about openai-agents

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build AI agents with tool calling and structured outputs?▼

Build AI agents using the OpenAI Agents SDK with tool definitions validated by Zod schemas. Define tools with input/output types, pass them to the agent, and receive structured responses with strict typing and streaming support across Node.js, Cloudflare Workers, and Next.js environments.

Can I create voice agents with real-time conversations over WebRTC?▼

Yes. Build realtime voice agents using the OpenAI Agents Realtime SDK with WebRTC and WebSocket transports. Deploy conversational AI that handles streaming audio input/output with the same tool-calling and guardrails capabilities as text agents.

How do I set up multi-agent handoffs and human-in-the-loop approval?▼

Configure multi-agent workflows where agents hand off to specialists and implement human-in-the-loop approval for critical actions. Use input/output guardrails to validate decisions before execution and require human sign-off when needed.

What are common errors when building agents with Zod schemas and how do I prevent them?▼

Nine documented errors include Zod schema type mismatches, MCP tracing failures, and infinite loops. The Skill provides error prevention patterns and references to avoid these pitfalls when defining tool schemas, configuring streaming, and managing agent state.

Does this support production deployment across different JavaScript runtimes?▼

Yes. Deploy production-grade agents to Node.js, Cloudflare Workers, and Next.js with cross-runtime templates. The Skill includes implementation patterns for each environment, supporting text agents, voice agents, and multi-agent workflows at scale.

Can I use input and output guardrails to enforce validation rules on agent actions?▼

Input/output guardrails validate all agent decisions against defined rules before execution. Combine them with Zod schemas to enforce strict type checking, business logic constraints, and safety requirements across text and voice agents.

Related Skills

Explore More Software Engineering Skills→
inkbox-aiinkbox-ai
33

inkbox-ts

Integrate the Inkbox TypeScript SDK to manage identities, mailboxes, vaults, and contacts.

Official
Advanced
nexus-labs-automationnexus-labs-automation
7

human-in-the-loop

Capture human approval, clarification, and escalation events as structured spans in AI agent workflows.

Official
Advanced
ovachieverovachiever
52

openai-agents

Build OpenAI Agents SDK workflows for text, voice, and multi-agent handoffs.

Community
Advanced
ksopylaksopyla

agent-tools-and-platforms

Provide guidance on agentic platforms, orchestration frameworks, agent runtimes, memory libraries, voice frameworks, durable execution engines, sandboxing solutions, and protocol standards.

Community
Advanced