# Arxiv Py Paper Retrieval

> Incubating Butler wrapper spec for upstream 'arxiv.py'.

- Skill: `ocean326/arxiv-py-paper-retrieval` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add ocean326/arxiv-py-paper-retrieval`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ocean326/arxiv-py-paper-retrieval/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: ocean326 (https://skillmd.com/u/ocean326)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/ocean326/arxiv-py-paper-retrieval

---

# arxiv.py

## Status

- `incubating`
- 这不是生产可执行 skill，而是 Butler 对外部上游的包装草案。
- 默认不得加入 `chat_default` / `codex_default` / `automation_safe`。

## Proposed Butler Shape

- arXiv search, latest-paper watch, author/topic feed, paper metadata hydration

## Upstream

- repo_or_entry: https://github.com/lukasschwab/arxiv.py
- docs: https://lukasschwab.me/arxiv.py/
- upstream_type: open_source_library
- maturity: high

## Why This Matters

- Cleanest path to build an arXiv research-monitor skill for Butler.
- High leverage for ML, CS, math, and physics paper discovery.

## Risks

- Coverage limited to arXiv-indexed papers.
- Needs download and attachment policy if PDFs are fetched.

## Suggested Future Exposure

- chat_default
- automation_safe

## Implementation Checklist

- 明确 Butler 输入输出 contract，不直接复刻上游 API。
- 确认认证、限流、缓存和错误恢复策略。
- 补执行脚本或 tool bridge，而不是只停留在说明文档。
- 补测试与 `skill-pool-verify` / `skill-pool-maintain` 校验。
- 审阅后再决定是否加入 collection registry。