arxiv | MediaUse
Install This Skill<br>npx skills add mediause/agent-skills/arxivCopy
This skill is synced from github.com/mediause/agent-skills and automatically maps to the matching site plugin data by plugin name. You only need to follow the CLI steps in the skill guide.<br>skill.md<br>arxiv<br>Discover arXiv research with a guest-only workflow for paper search, category recents, author lookups, and detailed paper retrieval.
Discover arXiv research with a guest-only workflow for paper search, category recents, author lookups, and detailed paper retrieval.
Scope
Use this skill when the task targets arXiv operations such as:
Account: plugin health
Search: papers by keyword
Recent: latest submissions in a category
Paper: full details for a specific paper ID
Author: papers by a named author
arXiv is a public read-only repository. No login required. All operations are read-only.
1. Install MediaUse CLI (Windows Only)
Use the official install script for Windows:
https://release.mediause.dev/install.ps1
Run:
powershell -C "iwr https://release.mediause.dev/install.ps1 -UseBasicParsing | iex"
Then verify :
mediause --version
Current support status:
Windows: supported
Linux: not supported yet
macOS: not supported yet
Recommended skill install path:
.mediause/skills/arxiv/SKILL.md
2. Get and Configure MediaUse Key
2.1 Apply for key
Open https://mediause.dev/
Sign in to your account.
Open Project.
Create or copy your API key.
2.2 Configure key in CLI
mediause manage key --json
3. Core Flow (Mandatory Order)
Always follow this order:
Discover site and commands.
Bind account context with use account (guest is the only mode for arXiv).
Execute dynamic site actions.
Verify with trace/task.
arXiv is a fully public API — no login required. Always use arxiv:guest as the account context. Skip auth health for guest mode.
3.1 Discover and plugin setup
mediause plugin list --json<br>mediause plugin add arxiv --json<br>mediause arxiv -h<br>mediause arxiv search -h<br>mediause arxiv get -h<br>mediause arxiv user -h<br>mediause arxiv account -h
3.2 Bind guest context
arXiv does not require login. Use guest mode.
mediause use account arxiv:guest --show --json
Guest mode rules:
All arXiv operations are read-only (account health, search papers, get paper, get recent, user author).
No write operations exist for arXiv.
If page shows unusual traffic or captcha, repeat with --show to manually resolve.
3.3 Auth health
Not required for guest mode. Skip this step for arXiv.
4. arXiv Dynamic Command Map (v1)
public arXiv API plugin with guest default account and read-only commands.
4.1 account.health
Check plugin/runtime health for current arXiv context.
mediause arxiv account health --json
4.2 search.papers
Search papers by keyword across all fields (title, abstract, authors, etc.).
mediause arxiv search papers --query "" [--limit ] --json
--query (required): keyword or phrase, e.g. "attention is all you need"
--limit: max results, default 10, max 25
Columns returned: id, title, authors, published, primary_category, url
Example:
mediause arxiv search papers --query "transformer language model" --limit 10 --json<br>mediause arxiv search papers --query "diffusion model image generation" --limit 5 --json
4.3 get.recent
List recent submissions in a specific category, sorted by submission date descending.
mediause arxiv get recent --category [--limit ] --json
--category (required): arXiv category code, e.g. cs.CL, cs.LG, math.PR, q-bio.NC
--limit: max results, default 10, max 50
Columns returned: id, title, authors, published, primary_category, url
Common categories:
CategoryDescriptioncs.CLComputation and Language (NLP)cs.LGMachine Learningcs.CVComputer Visioncs.AIArtificial Intelligencecs.CRCryptography and Securitymath.STStatistics Theoryq-bio.NCNeurons and Cognitionphysics.comp-phComputational Physics
Example:
mediause arxiv get recent --category cs.CL --limit 20 --json<br>mediause arxiv get recent --category cs.LG --limit 10 --json
4.4 get.paper
Get full details for a specific paper by arXiv ID.
mediause arxiv get paper --id --json
--id (required): arXiv paper ID, e.g. 1706.03762 or 2303.08774
Columns returned: id, title, authors, published, updated, primary_category, categories, abstract, comment, pdf, url
Example:
mediause arxiv get paper --id 1706.03762 --json<br>mediause arxiv get paper --id 2303.08774 --json
4.5 user.author
List papers by a named author, newest first. Author name matching is fuzzy — try alternate spellings if no results.
mediause arxiv user author --author "" [--limit ] --json
--author (required): author full name or initials, e.g. "Yoshua Bengio" or "Y Bengio"
--limit: max results, default 20, max 50
Columns returned: id, title, authors, published, primary_category, url
Example:
mediause arxiv user author --author "Yoshua Bengio" --limit 20 --json<br>mediause arxiv user author --author "Andrej Karpathy" --limit 10 --json<br>mediause arxiv user author --author "Y LeCun"...