Skip to content
Back to search

paper-radar

Finds the AI papers 28 tech companies put on arXiv over any date range, and splits lead authorship from bylines. A SKILL.md agent skill — no ML, no state, stdlib only.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/tigerless-labs--paper-radar/badge.svg)](/listings/tigerless-labs--paper-radar)

HTML

<a href="/listings/tigerless-labs--paper-radar"><img src="/api/listings/tigerless-labs--paper-radar/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Hand it to your agent

Paste this into Claude Code, Codex, or anything else with a shell:

Install the paper-radar skill from https://github.com/tigerless-labs/paper-radar

That is the whole install. The skill is one self-contained folder — `skills/paper-radar/`, pure stdlib, nothing to build — so the agent clones the repo and drops that folder wherever it keeps its skills. Say "for this project" if you want it local rather than global.

By hand

If you would rather do it yourself, or your agent has no shell:

git clone https://github.com/tigerless-labs/paper-radar
cp -r paper-radar/skills/paper-radar ~/.claude/skills/     # Claude Code
cp -r paper-radar/skills/paper-radar ~/.codex/skills/      # Codex
cp -r paper-radar/skills/paper-radar ~/.agents/skills/     # generic SKILL.md agents

Per-project instead of global: copy it into `.claude/skills/` inside the repo.

Claude on the web, ChatGPT, and other upload-based agents

Download **[`paper-radar-skill.zip`](../../releases/latest)** from the latest release and upload it as a skill — Claude web takes it under **Settings → Capabilities → Skills**; ChatGPT and other hosted agents take the same folder as project files.

To build the zip yourself:

cd skills && zip -r ../paper-radar-skill.zip paper-radar

> **One caveat, stated plainly:** the script fetches from `arxiv.org` over HTTPS. Hosted sandboxes > that block outbound network will run the code and return nothing. This is a property of the > sandbox, not of the skill — check whether yours allows network egress before relying on it there.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 150 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. <h1 align="center">paper-radar</h1>
  2. <p align="center"><strong>What 28 AI labs published on arXiv, over any date range</strong></p>
  3. <p align="center">
  4. <img src="https://img.shields.io/badge/release-v0.1.0-brightgreen.svg" alt="release" />
  5. <img src="https://img.shields.io/badge/agent-Claude%20%7C%20ChatGPT%20%7C%20Codex-blue.svg" alt="agent" />
  6. <img src="https://img.shields.io/badge/format-SKILL.md-lightgrey.svg" alt="format" />
  7. <img src="https://img.shields.io/badge/deps-stdlib%20only-informational.svg" alt="stdlib only" />
  8. <img src="https://img.shields.io/badge/license-MIT-yellow.svg" alt="license MIT" />
  9. </p>
  10. **paper-radar finds the AI papers 28 labs put on arXiv, over any date range you pick.** Every count
  11. splits two ways — papers the lab **led**, and papers it merely **appears on**.
  12. Method: match against [ROR](https://ror.org) and email domains, using the author block read off each
  13. paper's HTML.
  14. You need it because arXiv does not carry affiliation. The `<arxiv:affiliation>` field has a **1%
  15. fill rate**; Semantic Scholar reaches 7%; OpenAlex returns **zero** for preprints. So "what did
  16. Google publish in the last two weeks" has no API that answers it.
  17. ## Install
  18. ### Hand it to your agent
  19. Paste this into Claude Code, Codex, or anything else with a shell:
  20. ```
  21. Install the paper-radar skill from https://github.com/tigerless-labs/paper-radar
  22. ```
  23. That is the whole install. The skill is one self-contained folder — `skills/paper-radar/`, pure
  24. stdlib, nothing to build — so the agent clones the repo and drops that folder wherever it keeps its
  25. skills. Say "for this project" if you want it local rather than global.
  26. ### By hand
  27. If you would rather do it yourself, or your agent has no shell:
  28. ```bash
  29. git clone https://github.com/tigerless-labs/paper-radar
  30. cp -r paper-radar/skills/paper-radar ~/.claude/skills/ # Claude Code
  31. cp -r paper-radar/skills/paper-radar ~/.codex/skills/ # Codex
  32. cp -r paper-radar/skills/paper-radar ~/.agents/skills/ # generic SKILL.md agents
  33. ```
  34. Per-project instead of global: copy it into `.claude/skills/` inside the repo.
  35. ### Claude on the web, ChatGPT, and other upload-based agents
  36. Download **[`paper-radar-skill.zip`](../../releases/latest)** from the latest release and upload it
  37. as a skill — Claude web takes it under **Settings → Capabilities → Skills**; ChatGPT and other
  38. hosted agents take the same folder as project files.
  39. To build the zip yourself:
  40. ```bash
  41. cd skills && zip -r ../paper-radar-skill.zip paper-radar
  42. ```
  43. > **One caveat, stated plainly:** the script fetches from `arxiv.org` over HTTPS. Hosted sandboxes
  44. > that block outbound network will run the code and return nothing. This is a property of the
  45. > sandbox, not of the skill — check whether yours allows network egress before relying on it there.
  46. ## Use it
  47. In Claude Code it is a slash command:
  48. ```
  49. /paper-radar what did big tech publish on arXiv in the last two weeks?
  50. ```
  51. In any agent, just ask — the skill's description is written so it gets picked up on its own:
  52. ```
  53. has Xiaomi published anything on self-evolving agents?
  54. which companies led the most agent-memory papers this month?
  55. ```
  56. Or run it yourself, no agent involved:
  57. ```bash
  58. python3 skills/paper-radar/scripts/paper_radar.py --days 14
  59. ```
  60. A two-week window (~3000 papers) takes about three minutes. Every flag is documented in the skill's
  61. `SKILL.md`.
  62. ## How it works
  63. ```
  64. 0 listing arXiv API: category x submittedDate, deduped by ID
  65. 1 full text arxiv.org/html/{id}, Range-request the first 90KB, cut ltx_authors to ltx_abstract
  66. 2 structure map (author, superscripts) against (superscript, affiliation)
  67. 3 entity email domain, then ROR name variants, then the lab alias table
  68. 4 grading lead = the first author's institution; intern markers surfaced, not judged
  69. 5 supplements Apple RSS, MSR embedded JSON, arXiv team-name queries
  70. 6 report dedupe, write markdown, keep the evidence for every match
  71. ```
  72. The report gives every lab two columns, and the gap between them is the whole point:
  73. ```
  74. | Company | Total hits | Of which lead |
  75. |-----------------|-----------:|--------------:|
  76. | Microsoft | 31 | 18 |
  77. | Meta | 9 | 9 |
  78. | Apple | 8 | 0 |
  79. ```
  80. The second column comes from the superscripts: whoever the first author is attached to led the
  81. paper. In one two-week window Google appeared on 10 papers and **led 4**; Adobe appeared on 5 and
  82. **led 0**.
  83. Every match records what it matched on and the raw text it matched, so you can check any row by eye.
  84. Where it goes wrong, and what to do about it, is in the skill's `SKILL.md`.
  85. ## Coverage
  86. Google · Google DeepMind · Microsoft · Meta · Apple · NVIDIA · Amazon · OpenAI · Anthropic · IBM ·
  87. Adobe · Salesforce · Alibaba · Tencent · ByteDance · Baidu · Huawei · Xiaomi · Ant Group ·
  88. Meituan · Kuaishou · DeepSeek · MiniMax · Moonshot AI · Zhipu AI · StepFun · Shanghai AI Lab · BAAI
  89. 26 companies plus two non-profit research institutes (Shanghai AI Lab, BAAI). Each is anchored to a
  90. real ROR ID, carrying 98 official name variants between them, Chinese included. Adding one is a
  91. single entry in `scripts/data/ror_anchors.json` — but **anchor the ROR ID
  92. by hand**. Fuzzy ROR search lands on the wrong entity often enough to matter: Adobe returns "Adobe
  93. Gastroenterology", Meta returns "Meta House", Google returns the Ireland subsidiary.
  94. ## What it does not do
  95. - **No fuzzy generalization.** A company that isn't in `ror_anchors.json` is invisible, full stop.
  96. This is the direct cost of using lookup instead of a model — in exchange, there are no confident
  97. wrong answers.
  98. - **No heat signal.** Measured: of 231 big tech papers in a two-week window, **zero** were discussed
  99. on Hacker News, while every HN-discussed paper in that window came from academia or a startup.
  100. Paper count is supply, not attention.
  101. - **No PDF parsing.** 12% of submissions ship as PDF only and are simply missing. Reading TeX source
  102. is the planned fix and is not built yet.
  103. - **Per-company miss rates vary wildly.** Apple's authors usually omit their affiliation, so the
  104. author block alone misses ~78% of its papers; that one is patched from Apple's official RSS feed.
  105. Do not apply a single confidence level to the whole report.
  106. ## License
  107. MIT
  108. ---
  109. © Tigerless · [tigerless.ai](https://tigerless.ai) · [tigerless.com](https://www.tigerless.com)

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.