--- name: ai-news-digest description: "Daily AI News Digest Bot that aggregates RSS feeds from 20 AI news sources, uses AI to filter for relevance and importance, compiles top stories into digestible formats, and delivers via Discord/Telegram. Use when: (1) building automated AI news monitoring systems, (2) setting up daily/periodic AI news digests for teams or communities, (3) creating production-ready RSS aggregation with AI filtering, (4) delivering curated AI news content to Discord or Telegram channels. Targets 5 top stories daily from comprehensive source monitoring." --- # AI News Digest Bot A production-ready system that automatically aggregates AI news from 20 high-quality sources, filters articles using AI for relevance and importance, and delivers polished daily digests to Discord and Telegram channels. ## Overview This skill implements a complete AI news pipeline: 1. **RSS Aggregation**: Monitors 20 carefully selected AI news sources 2. **AI Filtering**: Uses LLM to score articles for relevance, importance, novelty, and quality 3. **Digest Compilation**: Formats top stories for optimal platform delivery 4. **Automated Delivery**: Sends to Discord/Telegram with proper formatting and chunking Perfect for AI teams, communities, and professionals who need consistent, high-quality AI news curation. ## Quick Start ### Basic Usage ```bash # Run complete pipeline with default settings python3 scripts/run_digest.py run # Generate Discord and Telegram formats python3 scripts/run_digest.py run --formats discord telegram # Test individual components python3 scripts/run_digest.py test rss python3 scripts/run_digest.py status ``` ### Configuration Set environment variables: ```bash export OPENAI_API_KEY="your-openai-key" export DISCORD_WEBHOOK="https://discord.com/api/webhooks/your-webhook" export TELEGRAM_BOT_TOKEN="your-bot-token" export TELEGRAM_CHAT_ID="your-chat-id" ``` ### Daily Automation Add to crontab for daily 9 AM UTC delivery: ```bash 0 9 * * * cd /path/to/skill && python3 scripts/run_digest.py run >> logs/digest.log 2>&1 ``` ## Core Components ### 1. RSS Aggregation (`scripts/rss_aggregator.py`) Monitors 20 premium AI news sources including: - **Research**: OpenAI, Anthropic, Google AI, DeepMind, MIT Tech Review - **Industry**: TechCrunch AI, VentureBeat AI, The Information - **Technical**: Ars Technica, IEEE Spectrum, Papers With Code - **Ethics & Safety**: AI Ethics Lab, AI Safety News **Features:** - Parallel feed processing with respectful rate limiting - Automatic deduplication by content hash - Configurable age filtering (default: 24 hours) - Error handling with graceful degradation - Caching for reliability and debugging **Usage:** ```bash # Fetch all sources python3 scripts/rss_aggregator.py --sources references/sources.json --max-age 24 # Custom output python3 scripts/rss_aggregator.py --output today_articles.json ``` ### 2. AI Filtering (`scripts/ai_filter.py`) Uses OpenAI GPT models to score each article on: - **Relevance** (0-10): How relevant to AI developments - **Importance** (0-10): Significance for the AI field - **Novelty** (0-10): How new or unique the information is - **Quality** (0-10): Writing quality and informativeness **Features:** - Intelligent fallback scoring when AI unavailable - Category classification (research/product/funding/partnership/etc.) - Rate limiting and error handling - Batch processing with progress tracking **Usage:** ```bash # Filter articles with minimum score threshold python3 scripts/ai_filter.py articles.json --min-score 6.0 --max-articles 50 # Get top 5 stories only python3 scripts/ai_filter.py articles.json --top-stories 5 # Use different model python3 scripts/ai_filter.py articles.json --model gpt-4 ``` ### 3. Digest Compilation (`scripts/digest_compiler.py`) Transforms filtered articles into platform-optimized formats: **Discord Format:** - Markdown formatting with embeds - Category grouping with emojis - 2000-character chunking - Rich metadata display **Telegram Format:** - Simplified markdown - 4096-character limits - Inline links and formatting - Mobile-optimized layout **Additional Formats:** - Email (HTML with styling) - Markdown (clean, shareable) - Plain text (universal compatibility) **Usage:** ```bash # Compile for Discord python3 scripts/digest_compiler.py filtered.json --format discord # Multiple formats python3 scripts/digest_compiler.py filtered.json --format email --output digest.html ``` ### 4. Complete Pipeline (`scripts/run_digest.py`) Orchestrates the entire process with: - **Configuration management** via environment variables - **Error handling** and recovery - **Automatic delivery** to configured platforms - **Logging** and monitoring - **Flexible scheduling** support **Usage:** ```bash # Full pipeline python3 scripts/run_digest.py run # Skip delivery (generate files only) python3 scripts/run_digest.py run --no-delivery # Deliver existing digest python3 scripts/run_digest.py deliver discord --file output/digest_discord_latest.txt # System status python3 scripts/run_digest.py status ``` ## Configuration ### Environment Variables ```bash # AI Filtering (required for quality scoring) OPENAI_API_KEY=sk-your-key-here OPENAI_MODEL=gpt-3.5-turbo # or gpt-4 for higher quality # Digest Settings DIGEST_MAX_AGE_HOURS=24 # How far back to look for articles DIGEST_MIN_SCORE=6.0 # Minimum AI score threshold DIGEST_MAX_ARTICLES=50 # Maximum articles to process DIGEST_TOP_STORIES=5 # Number of top stories in final digest DIGEST_FORMATS=discord,telegram # Output formats # Discord Delivery (optional) DISCORD_WEBHOOK=https://discord.com/api/webhooks/... # Telegram Delivery (optional) TELEGRAM_BOT_TOKEN=your-bot-token TELEGRAM_CHAT_ID=your-chat-id ``` ### Custom Configuration File Create `config.json` for advanced settings: ```json { "max_age_hours": 24, "min_score": 6.5, "max_articles": 30, "top_stories_count": 5, "openai_model": "gpt-3.5-turbo", "output_formats": ["discord", "telegram", "markdown"], "discord_webhook": "https://...", "telegram_bot_token": "...", "telegram_chat_id": "..." } ``` Use with: `python3 scripts/run_digest.py run --config config.json` ## Production Deployment ### Dependencies ```bash pip install -r requirements.txt ``` Required packages: - `feedparser` - RSS/Atom feed parsing - `requests` - HTTP client for APIs and webhooks - `python-dateutil` - Date/time handling ### Automated Scheduling **Cron (Linux/macOS):** ```bash # Daily at 9 AM UTC 0 9 * * * cd /path/to/ai-news-digest && python3 scripts/run_digest.py run # Every 6 hours 0 */6 * * * cd /path/to/ai-news-digest && python3 scripts/run_digest.py run ``` **Systemd (Linux):** ```bash # Create service and timer files sudo systemctl enable ai-news-digest.timer sudo systemctl start ai-news-digest.timer ``` **Docker:** ```bash docker-compose up --build # For one-time run # Or configure with cron for scheduled execution ``` See [references/deployment.md](references/deployment.md) for complete deployment guide. ### Monitoring **Health Checks:** ```bash # System status python3 scripts/run_digest.py status # Test components python3 scripts/run_digest.py test rss python3 scripts/run_digest.py test compile ``` **Log Monitoring:** ```bash tail -f logs/digest.log grep ERROR logs/digest.log ``` **File Management:** ```bash # Clean old cache files (7+ days) find cache/ -name "*.json" -mtime +7 -delete # Archive old digests tar -czf archive_$(date +%Y%m).tar.gz output/ ``` ## Output Examples ### Discord Output ``` ## 🤖 AI News Digest - March 6, 2026 📊 **5 top stories** • Average quality: 7.8/10 ### 🔬 Research ⚡ **OpenAI Announces GPT-5 with Reasoning Capabilities** 📈 Score: 9.2/10 | 📰 OpenAI Blog *Revolutionary breakthrough in AI reasoning shows 40% improvement in complex problem-solving tasks...* 🔗 [Read more](https://...) ### 🚀 Product 💰 **Anthropic Raises $4B Series D Led by Google** 📈 Score: 8.1/10 | 📰 TechCrunch *Major funding round values company at $60B as AI safety becomes...* --- *Generated at 09:15 UTC • Powered by AI News Digest Bot* ``` ### Telegram Output ``` *🤖 AI News Digest - March 6, 2026* 📊 5 top AI stories today ⚡ *1. OpenAI Announces GPT-5 with Reasoning* 📈 Score: 9.2/10 | 📰 OpenAI Blog Revolutionary breakthrough in AI reasoning... 🔗 [Read more](https://...) 💰 *2. Anthropic Raises $4B Series D* 📈 Score: 8.1/10 | 📰 TechCrunch Major funding round values company... 🔗 [Read more](https://...) _Generated at 09:15 UTC_ ``` ## Customization ### Adding News Sources Edit `references/sources.json`: ```json { "name": "Your AI Blog", "url": "https://yourblog.com", "rss": "https://yourblog.com/feed.xml", "category": "industry", "weight": 0.8, "description": "Industry insights and analysis" } ``` ### Custom Filtering Logic Modify `ai_filter.py` to adjust: - Scoring criteria and prompts - Fallback heuristics - Category classifications - Quality thresholds ### Platform Integration Add new delivery platforms by: 1. Creating delivery method in `run_digest.py` 2. Adding platform-specific formatting in `digest_compiler.py` 3. Handling platform limits and requirements ### Content Customization Adjust digest content by: - Modifying article selection in filtering phase - Customizing formatting templates - Adding/removing metadata fields - Changing emoji and styling ## Revenue Potential This skill targets $200-$2,000 monthly revenue through: 1. **SaaS for Teams**: AI news monitoring for companies 2. **Discord Bot Premium**: Advanced features for AI communities 3. **API Access**: Curated AI news data for developers 4. **Custom Implementations**: Branded solutions for organizations 5. **Newsletter Service**: Email delivery with premium tiers The system is production-ready for scaling and monetization with minimal additional development. ## Troubleshooting ### Common Issues **No articles found:** - Check RSS source availability - Verify network connectivity - Review source configurations **AI filtering fails:** - Confirm OpenAI API key and credits - Check rate limits and quotas - Review error logs for details **Delivery failures:** - Validate webhook URLs and tokens - Check message size limits - Test manual delivery commands **Performance issues:** - Monitor API usage and costs - Optimize source selection - Implement caching strategies See [references/deployment.md](references/deployment.md) for detailed troubleshooting and optimization guides.