| name | fetching-ai-news |
| description | Fetches AI news and vibe-coding stories from free sources and exposes a bundled MCP tool. Use when asked to scrape AI news, gather recent AI headlines, build a digest, or send the digest to Telegram. |
| compatibility | Requires Python 3.10+ and pip install -r references/requirements.txt. |
| license | MIT |
Fetching AI News
Fetches, classifies, and ranks AI News and Vibe Coding stories with a bundled reference implementation.
When to Use
- User asks to fetch or scrape recent AI news
- User wants AI News Digest or Vibe Coding digest
- User wants to preview digest content before sending
- User wants to send the generated digest to Telegram
- User needs a reusable AI-news MCP capability without paid APIs
Instructions
- Install dependencies from the bundled reference package.
cd skills/fetching-ai-news
pip install -r references/requirements.txt
- Start the bundled MCP server.
cd skills/fetching-ai-news
python -m references.src.mcp_server
-
Call run_ai_news_digest(hours=13, max_per_category=6, send=false).
-
Summarize counts plus top links from items.ai and items.vibe_coding.
-
Only use send=true after confirming Telegram credentials are configured.
References
Workflow Checklist
Common Pitfalls
- Do not write a new scraper when the bundled reference implementation already fits
- Do not set
send=true unless user explicitly wants Telegram delivery
- Do not assume Telegram credentials exist; check environment first
- Do not treat empty results as failure; check
raw_count first
- Run the server from
skills/fetching-ai-news, not repo root
