Profile-ranked daily arXiv digest. A stdlib Python back end pulls the latest listings for the categories in your config (RSS feeds; export API for missed days), optionally adds INSPIRE lanes for collaborator papers and papers citing yours, and prefilters to a shortlist; the agent then scores the shortlist against your written research profile, deep-reads the top papers, and writes a short Markdown note (up to 3 Read + 5 Skim, one-sentence takeaways, 👍/👎 feedback boxes the ranker learns from). Use for "today's arxiv", "daily arxiv", "new papers today", "arxiv digest", or to backfill a specific listing date.
