AI Agent Instructions

You are a financial news analyst for Alpha Stocks Insight. Your task: Generate a professional stock analysis article. INPUT: You will receive: - A stock ticker (e.g., NASDAQ:NVDA) - Recent news headlines and data for that stock - Today's date OUTPUT: Return a JSON object matching this exact schema: { "id": <next_available_integer>, "slug": "<kebab-case-title>", "title": "<compelling headline, 60-100 chars>", "author": "<one of: Marcus Chen, Sarah Mitchell, James Rivera, Emma Thornton, David Park, Lisa Wang, Robert Kim, Michael Torres>", "authorBio": "<one-sentence bio for the author>", "date": "<e.g., April 20, 2026>", "time": "<e.g., 9:42 AM ET>", "dateISO": "<ISO 8601 datetime>", "teaser": "<2-3 sentence summary, 150-200 chars>", "content": "<FULL article in Markdown, 600-1200 words, with ## headers>", "tickers": [{"exchange": "NASDAQ", "symbol": "NVDA"}], "category": "<one of: Earnings Preview, Earnings, Stock Analysis, Technology, Semiconductors, Cloud Computing, AI & Technology, Big Tech, Government & Defense, AI Infrastructure>", "featured": false, "readTime": "<X min read>", "tags": ["tag1", "tag2", "tag3"] } RULES: - Content must include sections: analysis, key metrics, risks, conclusion - Always end content with: "*This analysis is for informational purposes only.*" - Use factual, measured language — not hype - Include specific numbers and data points from the provided news - Do NOT invent financial figures — only use what was provided - Do NOT provide a buy/sell recommendation

# agents/generate_articles.py import json, os, datetime, httpx import anthropic client = anthropic.Anthropic(api_key=os.environ["ANTHROPIC_API_KEY"]) finnhub_key = os.environ["FINNHUB_API_KEY"] # 1. Load ticker list with open("data/tickers.json") as f: tickers = json.load(f) # 2. Load existing articles (to get next ID) with open("data/articles.json") as f: articles = json.load(f) next_id = max(a["id"] for a in articles) + 1 # 3. For each ticker, fetch news + generate article for ticker in tickers[:3]: # limit to 3 per run sym = ticker["symbol"] # Fetch latest news from Finnhub url = f"https://finnhub.io/api/v1/company-news?symbol={sym}&from={datetime.date.today()}&to={datetime.date.today()}&token={finnhub_key}" news = httpx.get(url).json()[:5] # top 5 headlines news_text = "\n".join([f"- {n['headline']}: {n['summary'][:200]}" for n in news]) # 4. Call Claude to generate article msg = client.messages.create( model="claude-opus-4-6", max_tokens=2000, system=SYSTEM_PROMPT, # paste the prompt from Step 2 messages=[{ "role": "user", "content": f"Generate an article for {ticker['exchange']}:{sym}\n\nRecent news:\n{news_text}\n\nToday: {datetime.date.today()}" }] ) # 5. Parse and append article article = json.loads(msg.content[0].text) article["id"] = next_id next_id += 1 articles.insert(0, article) # 6. Save updated articles.json with open("data/articles.json", "w") as f: json.dump(articles, f, indent=2) print(f"Generated {len(tickers[:3])} new articles.")

name: Generate Daily Articles on: schedule: - cron: '0 7 * * 1-5' # 7 AM UTC, weekdays workflow_dispatch: # manual trigger jobs: generate: runs-on: ubuntu-latest steps: - uses: actions/checkout@v4 - uses: actions/setup-python@v5 with: { python-version: '3.12' } - run: pip install anthropic httpx - run: python agents/generate_articles.py env: ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }} FINNHUB_API_KEY: ${{ secrets.FINNHUB_API_KEY }} - name: Commit and push new articles run: | git config user.name "AI Agent" git config user.email "agent@alphastocksinsight.com" git add data/articles.json git commit -m "chore: add daily AI-generated articles" git push # Vercel auto-deploys on push to main ✓