Skip to content
Get packs

Guide · Agent skill

How to use the Semantic Scholar Paper Search agent skill

Updated · written with AI from the project's own documentation and checked against our catalogue data

The semantic-scholar skill lets an AI agent search the Semantic Scholar database for published journal and conference papers, such as IEEE, ACM and Springer venues, rather than only arXiv preprints. This guide explains what an agent skill is, how to install this one, how to trigger it with plain requests, and what it needs to run. By the end you will be able to run a literature search that returns titles, venues, citation counts, DOI links and open-access PDFs when they exist.

Semantic Scholar Paper Searchwanshuiyin/Auto-claude-code-research-in-sleep · MIT

Quick start

  1. 1Download the ZIP. Download the semantic-scholar skill ZIP from the skill page on this site.
  2. 2Unzip into a skills folder. Extract it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder.
  3. 3Or upload in the Claude apps. In the Claude apps, open Settings, then Capabilities, then Skills, and upload the ZIP.
  4. 4Confirm Python 3 and internet access. Make sure python3 runs and that the machine can reach api.semanticscholar.org.
  5. 5Ask in plain language. Request a search such as finding IEEE or journal papers on your topic, or ask for a specific DOI.
  6. 6Add overrides if needed. Append directives like max, type, min-citations or year to narrow the results.
  7. 7Review the table and summaries. Check venue, year, citations and DOI links, and note any papers also on arXiv.

What the semantic-scholar skill makes the agent do

An agent skill is a folder containing instructions (a SKILL.md file) that an AI agent reads when a request matches the skill's description. This skill is the published-venue counterpart to the arXiv skill: arXiv covers latest preprints, while this one covers peer-reviewed journal and conference papers indexed by Semantic Scholar.

For each result the agent shows the title, venue, year, citation count, authors, DOI links, and any free open-access PDF or short summary. If a paper also has an arXiv ID, the skill notes it instead of re-fetching it from arXiv.

  • Searches by topic, or fetches one paper by DOI, Semantic Scholar ID, ArXiv ID or CorpusId
  • Filters by publication type, minimum citations, year range and field of study
  • Runs a bundled Python helper script that calls the Semantic Scholar web API

Install the semantic-scholar skill

You can install the skill in either of two ways. Download the ZIP from the skill page on this site and unzip it into your personal skills folder, ~/.claude/skills/, or into a project's .claude/skills/ folder for use in that project only.

Alternatively, in the Claude apps go to Settings, then Capabilities, then Skills, and upload the ZIP there.

  • Personal install: unzip into ~/.claude/skills/
  • Project install: unzip into the project's .claude/skills/ folder
  • Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Check the requirements before you run it

The skill needs Python 3, because the helper script runs with python3. It also needs internet access to api.semanticscholar.org. Git is optional; the skill uses it only to locate the repository root.

The skill relies on a companion script, semantic_scholar_fetch.py, from the wider ARIS repository. If that file is missing, the skill describes a fallback that calls the API directly with Python's urllib module.

  • Python 3 available as python3
  • Internet access to api.semanticscholar.org
  • Git (optional)
  • The companion file semantic_scholar_fetch.py (fallback available if absent)

Trigger the skill with natural requests

You do not need special syntax. The skill's description lists phrases that should activate it, so you can ask in plain language. Requests that mention published work, venues or citation counts are the best fit.

You can also add overrides to the request. The skill documents directives such as max results, publication type, minimum citations, year range and sort order.

  • "Search Semantic Scholar for federated learning papers"
  • "Find IEEE papers on massive MIMO channel estimation"
  • "Find journal papers on graph neural networks, 2022 onward"
  • "Citation search for transformer-based speech enhancement"
  • Overrides: - max: 20, - type: journal, - min-citations: 50, - year: 2022-, - sort: citations
/semantic-scholar "topic" - type: journal - min-citations: 50

Understand the search modes and filters

By default the skill runs a relevance-ranked search limited to Computer Science and Engineering fields and to journal articles and conference papers. Passing fields: all removes the field-of-study filter. Sorting by citations or date switches to a bulk search.

The skill's notes say --venue needs an exact venue name, such as "IEEE Transactions on Signal Processing", and partial names like "IEEE" will not match. Use publication types and fields of study instead when you are unsure of the exact name.

  • Default: relevance search, max 10 results
  • sort: citations or date uses bulk search
  • DOI lookup example: DOI:10.1109/... as the query
python3 "$S2_FETCHER" search "QUERY" --max 10 --fields-of-study "Computer Science,Engineering" --publication-types JournalArticle,Conference

Review results and check for arXiv duplicates

Results come back as a table with the title, venue, year, citations, authors and type. For the top papers the agent adds a detailed summary including the DOI link, fields of study, TLDR, abstract and open-access status.

The skill checks each result for an arXiv ID. Papers with one are flagged as also on arXiv; venue-only papers are the unique value of this skill. If you keep a research wiki in your project, the skill also records the findings there.

  • Venue-only papers: no arXiv ID, common for IEEE and ACM work
  • Also on arXiv: noted, not re-fetched
  • Open-access PDF shown only when one is listed

Safety notes: scripts and network use

This skill runs a bundled Python script and sends search terms to the external Semantic Scholar API. Its metadata lists both script execution and network use. Do not enter sensitive or confidential research queries, because the search terms leave your machine.

Review the SKILL.md and the helper script before installing if your organisation restricts scripts or outbound network calls. The skill is released under the MIT licence.

  • Search terms are sent to api.semanticscholar.org
  • The skill executes a bundled Python helper script
  • Avoid confidential or unpublished research queries

What people use it for

  • A literature review that needs published IEEE or ACM papers rather than preprints
  • Finding highly cited foundational papers on a topic by sorting with sort: citations
  • Filling gaps left by the arXiv skill with venue-only publications
  • Looking up a specific paper by DOI to get its venue, citations and abstract

Frequently asked questions

How do I install the semantic-scholar skill?

Download the ZIP from the skill page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps you can instead upload the ZIP under Settings, then Capabilities, then Skills.

What does the semantic-scholar skill require?

It requires Python 3, since the helper script runs with python3, and internet access to api.semanticscholar.org. Git is optional and used only to find the repository root. It also expects the companion file semantic_scholar_fetch.py, with a urllib fallback if that file is missing.

Is the semantic-scholar skill free to use?

The skill is released under the MIT licence. This page does not state any pricing for the Semantic Scholar API, so check that service's own terms for usage limits.

How is this different from the arXiv skill?

The arXiv skill covers latest preprints and unrefereed work, while this skill targets published journal and conference papers from IEEE, ACM and Springer venues. The skill notes when a result is also on arXiv so you can avoid duplicates.

Why does my --venue search return nothing?

The skill's notes say --venue needs the exact venue name, such as "IEEE Transactions on Signal Processing". Partial names like "IEEE" will not match, so use publication types and fields of study instead.

Can it find a paper if I only have the DOI?

Yes. Pass the DOI, for example DOI:10.1109/..., and the skill fetches that single paper with its details instead of running a search.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use this skill if you are doing a literature review and need peer-reviewed venue papers with citation counts and DOI links, or if you already use the arXiv skill and want to fill its gaps. Pick the arXiv skill or another source if you need the newest preprints, or avoid this skill for confidential queries because search terms are sent to an external web API.

Download the skill