agentleFS
Agents4Academia-AIAgents4Academia-AI/UReKA/.claude/skills/alphaxiv/SKILL.md

alphaxiv

Skill9 starsChanged 5 months agoMITView on GitHub

Ingest a paper from AlphaXiv/arXiv (by arXiv ID or title) into a structured source file — fetches metadata, abstract, and the AI overview via the AlphaXiv API

5 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If $ARGUMENTS is empty, ask the user for an arXiv ID or paper title, then continue.

    line 13
  2. Very specific:

    If $ARGUMENTS is an arXiv ID (e.g.

    line 29
  3. Very specific:

    Format: Write the file following .claude/Templates/source.md for the frontmatter contract, with exactly this structure (no need to consult …

    line 43
  4. Distinctive:

    Propose: After writing, ask the user:

    line 83
  5. Very specific:

    Run /collate <title> to synthesise this source into a paper page?

    line 84
  6. Very specific:

    Run /collate <topic> for any new concepts introduced?

    line 85

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Ingest a paper from AlphaXiv (an arXiv front-end with AI overviews and metadata) into sources/alphaxiv_<slug>.md. Unlike /zotero, there's no local PDF — content is fetched from the AlphaXiv API: bibliographic metadata, the abstract, and an AI-generated overview. This is normally invoked by /collate as an ingestion step, but can be run directly.

If $ARGUMENTS is empty, ask the user for an arXiv ID or paper title, then continue.

Layout

Output goes to the personal sources/ by default. When invoked by /curriculum, write instead to that course's per-course library course/<slug>/library/ (the caller passes the destination directory) — fetched course material stays out of the personal base until the learner masters it and /tutor promotes it. Either way the filename is alphaxiv_<slug>.md. Template: .claude/Templates/source.md. The fetch tool is .claude/skills/alphaxiv/scripts/alphaxiv.py (stdlib only; the ingestion commands below need no API key).

Steps

  1. Resolve the paper.
    • If $ARGUMENTS is an arXiv ID (e.g. 2006.11239 or 2006.11239v2), use it.
    • Otherwise treat it as a title/topic and search, then confirm the match with the user before ingesting:
      sh .claude/src/pyrun --need any .claude/skills/alphaxiv/scripts/alphaxiv.py search "$ARGUMENTS" --limit 5
  2. Extract. Fetch the objective content for the resolved arXiv ID <id>:
    sh .claude/src/pyrun --need any .claude/skills/alphaxiv/scripts/alphaxiv.py paper <id>       # title, authors, date, abstract
    sh .claude/src/pyrun --need any .claude/skills/alphaxiv/scripts/alphaxiv.py metadata <id>    # authors, institutions, topics, GitHub
    sh .claude/src/pyrun --need any .claude/skills/alphaxiv/scripts/alphaxiv.py overview <id>    # AI-generated overview
  3. Format. Write the file following .claude/Templates/source.md for the frontmatter contract, with exactly this structure (no need to consult other source files):

    ---
    type: alphaxiv_source
    title: <paper title>
    concepts_mentioned: [<AlphaXiv topics + key concepts from the abstract>]
    source_links:
      - https://arxiv.org/abs/<id>
      - https://www.alphaxiv.org/abs/<id>
    ---
    
    ## Content
    
    ### Abstract
    
    <the paper abstract, verbatim>
    
    ### AI overview (AlphaXiv — AI-generated)
    
    <the AlphaXiv AI overview; this is AI-generated, not the authors' own text>
    
    ### Metadata
    
    - Authors: <...>
    - Institutions: <...>
    - Published: <date>
    - Topics: <...>
    - Code: <GitHub URL, if any>

    source_links holds URLs (a remote source has no local file). This is metadata + abstract + AI overview — not the full paper text; for full text, ingest the PDF via /zotero.

  4. Write. Slugify the title to snake_case and write <dest>/alphaxiv_<slug>.md, where <dest> is sources/ by default or course/<slug>/library/ when /curriculum invoked this skill. If it exists, ask before overwriting.
  5. Propose. After writing, ask the user:
    • Run /collate <title> to synthesise this source into a paper page?
    • Run /collate <topic> for any new concepts introduced?

Notes

  • Vendored, security-reviewed script from danjuan-77/alphaxiv-skill (MIT, see LICENSE in this folder). It calls only https://api.alphaxiv.org.
  • The ingestion commands (search, paper, metadata, overview) need no token. The script also has an ask command that needs ALPHAXIV_TOKEN; it's not used by this skill.

Wrap up

Report: the file written, concepts extracted, anything skipped (one-line reason each).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.