fetch-rss-feeds
Collect articles from configured RSS feeds. Fetches from multiple AI-focused news sources including tech blogs, research feeds, and company announcements. Filters articles by age and AI relevance.
name: fetch-rss-feeds description: > Collect articles from configured RSS feeds. Fetches from multiple AI-focused news sources including tech blogs, research feeds, and company announcements. Filters articles by age and AI relevance. metadata: domain: news category: collection mcp-servers: [] requires-approval: false confidence: 0.9 input:
- name: max_age_hours type: int default: 24 description: Maximum age of articles to collect
- name: filter_ai_relevant type: bool default: true description: Only include AI-relevant articles output:
- name: articles type: list[RawArticle] description: List of collected articles with title, url, source, published_date
Fetch RSS Feeds
Collect articles from all configured RSS news feeds.
Preconditions
- Network access to RSS feed URLs
- Feed configuration loaded from feeds.py
Actions
Step 1: Load Feed Configuration
Load all enabled RSS feeds from configuration, sorted by priority.
Feed categories:
- Company Blogs (priority 10): OpenAI, Anthropic, Google AI, Meta AI
- AI-Focused (priority 8-9): MIT Tech Review, VentureBeat, TechCrunch
- Research (priority 6-7): ArXiv AI, ML, and NLP feeds
- General Tech (priority 6-8): Hacker News, Ars Technica, The Verge
- Security (priority 5-6): Schneier on Security, The Hacker News
Step 2: Fetch Each Feed in Parallel
For each enabled feed:
- Make HTTP GET request to feed URL
- Parse RSS/Atom XML response
- Extract article entries with:
title: Article headlineurl: Link to full articlesource: Feed namepublished_date: Publication timestampsummary: Article description/excerpt
Handle errors gracefully - log and continue if individual feeds fail.
Step 3: Filter by Age
Remove articles older than max_age_hours based on published_date.
Step 4: Filter AI Relevance
If filter_ai_relevant is true, keep only articles matching AI keywords:
- Model names: GPT, Claude, Gemini, Llama
- Companies: OpenAI, Anthropic, DeepMind, Hugging Face
- Concepts: LLM, machine learning, neural network, AI agent
- Applications: chatbot, copilot, text-to-image
Step 5: Deduplicate by URL
Remove duplicate articles (same URL from multiple feeds).
Success Criteria
- At least one feed successfully fetched
- Articles have required fields (title, url, published_date)
- No duplicate URLs in output