openai-responses
Build agentic AI applications with OpenAI's Responses API - the stateful successor to Chat Completions. Preserves reasoning across turns for 5% better multi-turn performance and 40-80% improved cache utilization. Use when: building AI agents with persistent reasoning, integrating MCP servers for external tools, using built-in Code Interpreter/File Search/Web Search, managing stateful conversations, implementing background processing for long tasks, or migrating from Chat Completions to gain polymorphic outputs and server-side tools.
npx skills add majiayu000/claude-skill-registry --skill openai-responses --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Instructs the agent to use OpenAI's Responses API to build agentic applications with stateful conversations, preserved reasoning across turns, and built-in server-side tools (Code Interpreter, File Search, Web Search, Image Generation, MCP) for external integrations. It emphasizes polymorphic outputs (including messages, reasoning, and tool calls) and improved cache utilization for cost and latency benefits.
How it works
The skill guides creating conversations with automatic state management via conversation IDs, then sending inputs through openai.responses.create with optional tools (codes, MCP, file search, web search, image generation). It describes server-side tool execution (e.g., Code Interpreter) and how the model can perform tool calls and return polymorphic outputs. It shows how to initiate conversations, perform multiple turns while preserving context, and optionally manage history manually. It includes examples showing how to poll background tasks, handle MCP tool calls, and inspect outputs such as mcp_call, reasoning, and final messages.
When to use it
Use when building AI agents with persistent reasoning, integrating MCP servers for external tools, leveraging built-in tools (Code Interpreter/File Search/Web Search), managing stateful conversations, handling long-running tasks via background processing, or migrating from Chat Completions to gain polymorphic outputs and server-side tools.
What it can touch
The skill references tools available within the API, including:
- Code Interpreter
- File Search
- Web Search
- Image Generation
- MCP (Model Context Protocol) for external tool integrations It demonstrates how to call these tools via the API and how outputs may include tool-call structures like mcp_call, code_interpreter_call, file_search_call, web_search_call, image_generation_call, etc.
Caveats
License is MIT. It notes dependencies (openai@5.19.1+ for Node.js) and environment considerations (server-side usage to protect API keys). It states that the API does not store authorization tokens and tokens must be provided with each request. It also documents timeout behavior for standard vs background modes and emphasizes that reasoning state is preserved across turns.
# OpenAI Responses API **Status**: Production Ready **Last Updated**: 2025-10-25 **API Launch**: March 2025 **Dependencies**: openai@5.19.1+ (Node.js) or fetch API (Cloudflare Workers) --- ## What Is the Responses API? The Responses API (`/v1/responses`) is OpenAI's unified interface for building agentic applications, launched in March 2025. It fundamentally changes how you interact with OpenAI models by providing **stateful conversations** and a **structured loop for reasoning and acting**. ### Key Innovation: Preserved Reasoning State Unlike Chat Completions where reasoning is discarded between turns, Responses **keeps the notebook open**. The model's step-by-step thought processes survive into the next turn, improving performance by approximately **5% on TAUBench** and enabling better multi-turn interactions. ### Why Use Responses Over Chat Completions? | Feature | Chat Completions | Responses API | Benefit | |---------|-----------------|---------------|---------| | **State Management** | Manual (you track history) | Automatic (conversation IDs) | Simpler code, less error-prone | | **Reasoning** | Dropped between turns | Preserved across turns | Better multi-turn performance | |
- What Is the Responses API?
- Key Innovation: Preserved Reasoning State
- Why Use Responses Over Chat Completions?
- Quick Start (5 Minutes)
- 1. Get API Key
- 2. Install SDK (Node.js)
- 3. Or Use Direct API (Cloudflare Workers)
- Responses vs Chat Completions: Complete Comparison
- When to Use Each
- Architecture Differences
- Performance Benefits
- Stateful Conversations
- Automatic State Management
- Manual State Management (Alternative)
Sign up at https://platform.openai.com/ Navigate to API Keys section Create new key and save securely export OPENAI_API_KEY="sk-proj-..." npm install openai
What does the openai-responses skill do?
Build agentic AI applications with OpenAI's Responses API - the stateful successor to Chat Completions. Preserves reasoning across turns for 5% better multi-turn performance and 40-80% improved cache utilization. Use when: building AI agents with persistent reasoning, integrating MCP servers for external tools, using built-in Code Interpreter/File Search/Web Search, managing stateful conversations, implementing background processing for long tasks, or migrating from Chat Completions to gain polymorphic outputs and server-side tools.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill openai-responses --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.