literature-report_skill

Overview

This skill is an automated literature-report system that finds and summarizes the latest top‑tier research daily. It uses an LLM for semantic filtering and generates bilingual (Chinese/English) summaries, then pushes reports to enterprise messengers like Feishu, DingTalk, or WeChat.

How this skill works

The system periodically queries configured sources (RSS, PubMed, targeted journals), applies AI‑assisted relevance scoring, and extracts key points from selected papers. For each paper it produces a concise bilingual abstract, metadata (title, authors, journal, link), and an optional highlight list, then sends the compiled report to configured push endpoints or stores it for review.

When to use it

• Set up daily automated updates for a research group
• When you want instant bilingual summaries of new papers in a niche topic
• To maintain a daily digest of top‑journal publications for grant or lab meetings
• Integrate with team chat tools (Feishu/DingTalk/WeChat) for scheduled push
• Quickly validate whether recent literature impacts an ongoing experiment or manuscript

Best practices

• Provide a clear, specific research topic and sample keywords to reduce noise
• Limit max_papers per day to a manageable number (3–10) to keep digests concise
• Use a dedicated LLM API key stored in environment variables; do not commit keys to repos
• Enable push only for verified user IDs and test delivery with a single target before broad rollout
• Periodically review and refine journal lists and query strings to match evolving interests

Example use cases

• A PI receives a 5‑paper bilingual digest each morning for lab meeting preparation
• A postdoc configures a narrow topic to detect emerging methods in a subfield
• A translational team tracks clinical‑relevant advances across Nature/Science and PubMed
• An editorial office automates candidate article monitoring and shortlists items for internal review
• An industry research group pushes daily summaries to WeChat for distributed teams