open-responses-agent-dev

Open Responses Agent Development Skill

Build autonomous agents using the Open Responses API via the HuggingFace Inference Providers router.

What is Open Responses?

Open Responses is an open-source API standard for autonomous agent development. It provides:

• Sub-agent loops: Multi-step workflows in a single request
• Reasoning visibility: Access to agent thinking (raw, summary, or encrypted)
• Semantic streaming: Structured events instead of raw tokens
• Provider-agnostic design: Single endpoint, provider selection via model suffix

Key Concepts

Single Unified Endpoint

https://router.huggingface.co/v1/responses

Provider Selection via Model Suffix

moonshotai/Kimi-K2-Instruct-0905:groq      # Groq (fast)
meta-llama/Llama-3.1-70B-Instruct:together # Together AI
meta-llama/Llama-3.1-70B-Instruct:nebius   # Nebius (EU)
meta-llama/Llama-3.1-70B-Instruct:auto     # Auto selection

Supported Platforms

This skill works with:

• Claude Code
• Cursor (via .cursor/rules/)
• OpenCode (via .opencode/ and AGENTS.md)
• Codex (via .codex/ and AGENTS.md)

Supported Languages

• TypeScript/JavaScript
• Python

Installation

Claude Code

Install this skill directly in Claude Code:

# Clone the repository
git clone https://github.com/OthmanAdi/open-responses-agent-skill.git

# Create skills directory if it doesn't exist
mkdir -p ~/.claude/skills

# Copy the skill
cp -r open-responses-agent-skill/skills/open-responses-agent-dev ~/.claude/skills/

# Restart Claude Code or reload skills

Or install via URL:

/skills add https://github.com/OthmanAdi/open-responses-agent-skill

Cursor

Copy the rules file:

mkdir -p .cursor/rules
cp open-responses-agent-skill/.cursor/rules/open-responses-agent.mdc .cursor/rules/

OpenCode / Codex

Copy the configuration files:

# For OpenCode
cp -r open-responses-agent-skill/.opencode .
cp open-responses-agent-skill/AGENTS.md .

# For Codex
cp -r open-responses-agent-skill/.codex .
cp open-responses-agent-skill/AGENTS.md .

Quick Start

TypeScript (Recommended - Using OpenAI SDK)

npm install openai
export HF_TOKEN=your-token

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://router.huggingface.co/v1",
  apiKey: process.env.HF_TOKEN,
});

const response = await client.responses.create({
  model: "moonshotai/Kimi-K2-Instruct-0905:groq",
  instructions: "You are a helpful assistant.",
  input: "Explain quantum entanglement in simple terms.",
});

console.log(response.output_text);

Python (Recommended - Using OpenAI SDK)

pip install openai
export HF_TOKEN=your-token

from openai import OpenAI
import os

client = OpenAI(
    base_url="https://router.huggingface.co/v1",
    api_key=os.environ.get("HF_TOKEN"),
)

response = client.responses.create(
    model="moonshotai/Kimi-K2-Instruct-0905:groq",
    instructions="You are a helpful assistant.",
    input="Explain quantum entanglement in simple terms.",
)

print(response.output_text)

Structure

open-responses-agent-dev/
├── .claude-plugin/          # Claude Code plugin configuration
├── .cursor/rules/           # Cursor rules
├── .opencode/               # OpenCode configuration
├── .codex/                  # Codex configuration
├── skills/                  # Skill definition
│   └── open-responses-agent-dev/
│       └── SKILL.md         # Main skill instructions
├── examples/                # Complete examples
│   ├── typescript/          # TypeScript examples
│   └── python/              # Python examples
├── templates/               # Production-ready templates
│   ├── typescript/          # TypeScript starter
│   └── python/              # Python starter
├── docs/                    # Documentation
│   ├── migration-guide.md   # Chat Completion → Open Responses
│   └── provider-comparison.md # Provider comparison
├── AGENTS.md                # Agent instructions (OpenCode/Codex)
└── README.md                # This file

Examples

Basic Agent

Simple request with reasoning visibility:

• examples/typescript/basic-agent.ts
• examples/python/basic_agent.py

Sub-Agent Loop

Multi-step workflows with tools:

• examples/typescript/sub-agent-loop.ts
• examples/python/sub_agent_loop.py

Multi-Provider

Provider switching via model suffix:

• examples/typescript/multi-provider.ts
• examples/python/multi_provider.py

Reasoning Visibility

Accessing agent thinking:

• examples/typescript/reasoning-visibility.ts
• examples/python/reasoning_visibility.py

Provider Suffixes

Suffix	Provider	Description	Reasoning
:groq	Groq	Fast inference	RAW
:together	Together AI	Open weight specialist	RAW
:nebius	Nebius AI	European infrastructure	RAW
:auto	Auto	Automatic selection	Varies