Multi Agent AI Briefing — AI skill for Claude Code

Multi-Agent Ai-Briefing

A FastAPI application that scrapes Hacker News headlines and uses multiple AI models to process them.

AI Agents

• OpenAI gpt-4o-mini → summarizes headlines
• Claude claude-sonnet-4-6 → classifies headlines into AI, Technology, Business, or Other
• Claude falls back to claude-sonnet-4-5-20250929 if needed

This demonstrates a simple **multi-agent pattern**, where specialized AI agents handle different tasks and their outputs are combined into one briefing.

Hacker News
     ↓
Headlines
     ↓
┌─────────────────┐
│  OpenAI Agent   │ → Summary
├─────────────────┤
│  Claude Agent   │ → Category
└─────────────────┘
     ↓
Final Briefing

Project Structure

multi-agent-ai-briefing/
├── .env
├── pyproject.toml
├── uv.lock
└── src/
     ├── __init__.py
     ├── main.py
     ├── scraper.py
     ├── summarizer.py
     └── classifier.py

Setup

Install the dependencies defined in `pyproject.toml`:

uv sync

Create `.env`:

OPENAI_API_KEY=your_openai_api_key
ANTHROPIC_API_KEY=your_anthropic_api_key

Run

uv run main:app --reload

Open the API documentation:

http://127.0.0.1:8000/docs

Then call:

GET /briefing

Initialize a FastAPI Project with `uv`

uv init project-name
cd project-name
uv add fastapi uvicorn

Create:

src/main.py

from fastapi import FastAPI

app = FastAPI()


@app.get("/")
async def root():
    """Return a welcome message."""
    return {"message": "Hello World"}

Run:

uv run uvicorn main:app --reload

Open:

http://127.0.0.1:8000/docs