dailypaper-skills
Picks fresh arXiv and HuggingFace papers each day, writes a critical review and saves Markdown notes with formulas.
- 1,356 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Scripts access the network and write files to the chosen notes folder.
- The installer copies files into the agent directory; Git sync of notes is a separate opt-in.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/dailypaper-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Scripts access the network and write files to the chosen notes folder.
- The installer copies files into the agent directory; Git sync of notes is a separate opt-in.
This is third-party code. Review the repository files before installing.
What it does
dailypaper-skills is a set of seven skills for reading research papers every day. A single phrase starts the chain: collect new work from HuggingFace Daily and arXiv, filter by your research interests, write a critical review of each paper and detailed notes on the key ones. The output is three kinds of linked Markdown files: a daily recommendation page with must-read, worth-a-look and skippable priorities, a paper note with method, formulas, figures and limits, and concept notes that link to each other. Sending an arXiv link alone is enough: the paper is read in full and saved to the configured folder, Obsidian vaults included. README example phrases are in Chinese and English.
Who it is for. Researchers and graduate students who follow the paper stream and keep notes in Obsidian or plain Markdown.
Good fit when
- A morning check of which new papers on your topic deserve reading.
- After a few days away, a summary of the last three days or week.
- An arXiv link arrives and needs a full note with formulas and limits.
- Concept notes should connect to papers already collected.
- A weekly or monthly review of trending GitHub projects.
Not a fit when
- A paper only needs a chat summary with no saved files.
- The agent cannot read and write files, run commands or reach the network.
- Results must appear on a schedule without a user prompt: the installer creates no scheduled jobs.
Example request
Recommend papers from the past week for my interests and save notes on the must-reads.Limitations
Python 3.10 or newer, Git and curl are required, plus Poppler for PDF reading. No separate API key is needed: selection and reviews use the agent's own model quota. Only the optional Jev scoring mode asks for `TYPESAFE_API_KEY`. The defaults reflect the author's interests: large-model RL, world models and dexterous hands, so replace them with your own. The authors state that passing install and script tests does not mean every client was verified end to end. Automatic Git commit and push of notes is off by default.
How to disable. Remove the copied skills from the agent directory, for example `~/.codex/skills/`, and the local repository clone. Notes in your folder stay; delete them separately if unwanted.
FAQ
Which agents does the installer support?
Pass `codex`, `claude`, `cursor`, `copilot`, `gemini`, `opencode` or `openclaw` to `python3 install.py --agent`. The installer copies all seven skills with shared dependencies.
Is Obsidian required?
No. Any folder of Markdown files works. Obsidian helps because of two-way links between papers and concepts.
Does anything run on a schedule by itself?
No. Installation creates no jobs and runs start from a user phrase. Use a system or client scheduler for timed runs.
How do I change research interests?
Edit the personal config created by `python3 skills/_shared/user_config.py --init`: topics and the number of daily picks, ten at most by default.
What does the github-trending skill do?
It compiles daily, weekly and monthly lists of open-source projects and can search by a simulation topic.
Discussion
No questions or reviews yet.
Related
SkillsA skill picks five recent papers and one classic weekly, learns from your choices and writes a bilingual Feishu deep read.
Five skills from one author: arXiv reports, technical diagrams, a report site, a design system library and agent guardrails
The author's personal skills for reading papers and books, dissecting concepts, writing and PNG cards; instructions in Chinese, installed via skills CLI
Seven skills and scripts for the chain Zotero, full text through MinerU, analytical notes in Obsidian and literature search
A set of 16 skills for research work: ideation, experiments, paper writing, review, rebuttal, slides and research memory.
cafe3310's personal skill collection: spec-driven project work, reviews, reports, browser, 3D and video in one repository
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| dailypaper-skills (this card) | no | 1,356 | ||
| paper-reading-skill | no | 151 | turns one arXiv paper into a report; dailypaper-skills also filters a feed of new work and keeps linked notes | |
| arxiv-mcp-server | no | not required | 3,276 | exposes arXiv search and reading as MCP tools; dailypaper-skills adds selection, review and note saving |
README in short
Three stages, fetch, review and notes, run together or separately. The selection mode is configurable: agent scoring, keywords or the external Jev service.