huangkiki/dailypaper-skills

用Agent skills打造我的论文流水线

What it solves

This project provides a set of custom skills for Claude Code to automate the tedious process of tracking, filtering, and summarizing academic papers. It eliminates the need to manually browse through multiple sources like arXiv and HuggingFace to find relevant research, while automating the creation of structured notes in Obsidian.

How it works

The system operates as a pipeline of Claude Code skills that integrate with external sources and local tools:

  1. Discovery: It fetches candidate papers from HuggingFace Daily, Trending, and the arXiv API, scoring them based on user-defined keywords to filter out irrelevant content.
  2. Curation: Claude analyzes the filtered list and categorizes papers into "Must Read," "Worth Watching," or "Can Skip."
  3. Synthesis: For "Must Read" papers, it generates structured notes covering methods, experiments, formulas, and limitations, then saves them as Markdown files in an Obsidian vault.
  4. Integration: It can optionally integrate with Zotero to search existing libraries, read by category, or batch-process papers.
  5. Visualization: A lightweight FastAPI-based web viewer is provided to browse the generated recommendations and notes in a browser.

Who it’s for

Researchers, students, and AI practitioners who use Claude Code and Obsidian to manage their literature review process and want to automate their daily paper screening and knowledge management.

Highlights

  • Automated Pipeline: One command (e.g., "Today's paper recommendations") triggers the full fetch-review-note process.
  • Obsidian Native: Automatically maintains a concept library, double-linked notes, and directory pages (MOCs) within Obsidian.
  • Zotero Support: Ability to read papers directly from Zotero categories or search by title.
  • GitHub Trending Integration: Tracks fast-growing open-source projects on GitHub and labels them by research relevance.
  • Structured Note Generation: Extracts key information from arXiv HTML, project pages, and PDFs, including image handling.

Related

  • Project
    juliye2025/evil-read-arxiv

    An automated research workflow that searches, recommends, and analyzes academic papers from arXiv and Semantic Scholar, integrating results directly into Obsidian.

  • Project
    luwill/research-skills

    A collection of Claude Code skills designed to automate academic research workflows, including exhaustive literature retrieval, verified proposal writing, and high-fidelity slide generation.

  • Project
    917Dhj/DeepPaperNote

    An agent skill that transforms complex academic papers into structured, evidence-based Obsidian notes to build a durable academic wiki.

  • Project
    AgriciDaniel/claude-obsidian

    A local-first knowledge system for Claude Code and Agent Skills hosts that turns source material into linked, source-cited Obsidian pages with strict provenance tracking.

  • Project
    Weizhena/Deep-Research-skills

    A structured research workflow for Claude Code, OpenCode, and Codex that enables systematic outline generation, deep web investigation, and automated report creation.