paper-radar skill
A skill that finds AI papers from 28 labs on arXiv over any date range and splits led papers from co-authored ones.
- 224 ★ on the repository
- Last commit
Extended access
Why this level
- The script goes to the network, queries arXiv and fetches paper pages.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/paper-radarDetects 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.
- The script goes to the network, queries arXiv and fetches paper pages.
This is third-party code. Review the repository files before installing.
What it does
paper-radar answers the question of what a given company published on arXiv over a chosen period. arXiv itself barely provides affiliation data: the field is filled for about one percent of papers, Semantic Scholar covers about seven percent and OpenAlex returns nothing for preprints. The skill matches authors by ROR and email domains, reading the author block from each paper HTML. Every count is split into papers the lab led and papers where it merely appears among the authors. The skill is a single `skills/paper-radar` folder with a Python script on the standard library, with no build and no state. Run: `python3 skills/paper-radar/scripts/paper_radar.py --days 14`.
Who it is for. AI analysts and researchers who follow publications of large labs and companies.
Good fit when
- A report is needed on what Google and OpenAI published in the last two weeks.
- Papers a company leads must be told apart from those where it co-authors.
- A digest issue requires per-lab counters over an arbitrary date interval.
- A skill must be installed quickly in Claude Code or Codex or uploaded to a web version.
Not a fit when
- You need full affiliation of all authors outside the listed 28 labs.
- The environment blocks outbound requests to arXiv.
- You want deep quality assessment of papers, while the skill only counts and lists.
Example request
Which papers did DeepMind and Meta publish in the last 14 days? Separate led work from co-authorship.Limitations
The script calls `arxiv.org` over HTTPS, so in sandboxes without outbound network it runs and returns nothing. The list is limited to 28 labs. The method relies on author HTML and matching by ROR and email domains, so accuracy depends on how authors listed their organization. Version v0.1.0 is early, and the repository metadata declares no license although a badge says MIT. The last commit is dated September 2026.
How to disable. Delete the `paper-radar` folder from the agent skills directory. The skill keeps no state, so nothing else needs cleaning.
FAQ
How do I install the skill?
Send the agent an install request from the repository or copy the `skills/paper-radar` folder into the skills directory: `~/.claude/skills/`, `~/.codex/skills/` or `~/.agents/skills/`.
Does it work in the Claude web version?
Yes, the `paper-radar-skill.zip` archive from the latest release is uploaded as a skill. Outbound network is required, otherwise there is no result.
Why is affiliation data missing?
The arXiv affiliation field is filled for about 1% of papers, Semantic Scholar covers 7% and OpenAlex yields zero for preprints.
What counts as a led paper?
A paper where the lab leads the author list. Other cases count as co-authorship presence.
Are dependencies needed?
No. The script uses only the Python standard library and no build is needed.
Discussion
No questions or reviews yet.
Related
SkillsA pipeline that picks remote sensing papers from arXiv daily, reviews each in a GitHub Issue and publishes a daily digest.
Codex skill that picks fresh research papers for a researcher profile and builds a scored summary in Chinese.
An agent skill that turns an arxiv paper into a minimal Python implementation, with code anchored to sections and equations and gaps flagged honestly
A library of agent skills for medical research: literature search and appraisal, study design, data analysis and academic writing
A template for data scientists: AGENTS.md, CLAUDE.md and skills for Python, SQL, Polars and model diagnostics.
Around 200 skills for Claude Code on bioinformatics, chemistry and biostatistics, from RNA-seq analysis to molecular docking
README in short
The README states the task, the matching method, three installation ways and a sample run. The network access limit in sandboxes is stated separately.