Ads paper search

Skill xnchu/ads-paper-search/skills/ads-paper-search

Use when searching for scientific papers in astronomy, astrophysics, space physics, heliophysics, or planetary science — finding literature by topic/author/year, looking up citations or references of a paper, exporting BibTeX, resolving a bibcode/DOI/arXiv ID, or computing citation metrics. Connects to the NASA ADS (Astrophysics Data System, ui.adsabs.harvard.edu) API.From its SKILL.md

Install

npx -y skills add xnchu/ads-paper-search --skill ads-paper-search

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

5.5 KB, ~1.4k tokens by cl100k_base, as published. Nobody here has run it

NASA ADS Paper Search

Overview

Search the NASA Astrophysics Data System (ADS / SciX) — the standard literature database for astronomy, astrophysics, space physics, and planetary science (~15M records). This skill wraps the ADS API (https://api.adsabs.harvard.edu/v1) via a bundled stdlib-only Python script.

Setup (one-time, per user)

A personal ADS API token is required (free). The script looks for it in:

ADS_API_TOKEN environment variable (use this on claude.ai or in CI)
~/.ads/token — written by the setup subcommand below

If a command exits with "No ADS API token found", relay its setup steps to the user and ask for their token:

Create a free account at https://ui.adsabs.harvard.edu/
Copy the token from https://ui.adsabs.harvard.edu/user/settings/token
Run python3 scripts/ads.py setup THEIR_TOKEN — this verifies the token against the API and saves it to ~/.ads/token (mode 600), so setup never needs to happen again.

Never hardcode a token into this skill or share one token between users — ADS tokens are personal and rate limits (5000 requests/day) are per-account.

Quick reference

Run all commands with python3 scripts/ads.py (paths relative to this skill's directory). Add --json for raw JSON, -h for full options.

Task	Command
Search papers	`search 'QUERY' --rows 10 --sort "citation_count desc"`
Search with abstracts	`search 'QUERY' --abstract`
Look up bibcode(s)	`get 2016SSRv..199....5B [more...]`
BibTeX export	`export bibtex 2016SSRv..199....5B [more...]`
Papers citing X	`citations BIBCODE --rows 25`
Papers cited by X	`references BIBCODE --rows 25`
Just a citation count	`citations BIBCODE --rows 1` — total is in the `# N citations` header
Citation metrics / h-index	`metrics BIBCODE [more...]`

Sort options: score desc (relevance, default), citation_count desc, date desc (newest first), read_count desc.

Query syntax essentials

Unfielded terms search title+abstract+keywords. Combine with AND (implicit), OR, NOT, parentheses. Quote phrases.

Field	Example
Author	`author:"Smith, J"` — first author: `author:"^Smith, J"`
Year / range	`year:2024`, `year:2019-2024`
Title / abstract	`title:"magnetic reconnection"`, `abs:"radiation belt"`
Full text	`full:"chorus waves"`
Journal	`bibstem:ApJ`, `bibstem:JGRA`, `bibstem:GeoRL`
Object (SIMBAD)	`object:"Sgr A*"`
Identifiers	`bibcode:2022ApJ...930L..12E`, `doi:10.1126/science.aaf2939`, `arXiv:2101.01234` (as unfielded term or `identifier:`)
Refereed only	add `property:refereed`
Open access	`property:openaccess`
Collection	`collection:astronomy` / `physics` / `heliophysics` / `planetary_science`
Citation graph	`citations(bibcode:X)`, `references(bibcode:X)`, `reviews(topic)`, `useful(...)`, `similar(...)`, `trending(...)`

Example — highly-cited recent first-author papers on radiation belts:

python3 scripts/ads.py search 'author:"^Li, W" abs:"radiation belt" year:2020-2026 property:refereed' \
  --sort "citation_count desc" --rows 10

Full field list and operators: see references/search-syntax.md.

Workflow guidance

Literature search: start broad, check the reported total count, then narrow with fields (year:, bibstem:, property:refereed) or paginate with --start. Fetch abstracts (--abstract) only for the shortlist.
"Find THE paper": prefer --sort "citation_count desc" for foundational work, date desc for latest results.
Review articles: reviews(TOPIC) or add doctype:review / property:refereed.
Follow the citation graph: use citations/references subcommands on a key paper to expand a literature review in both directions.
Paper URL: any result links as https://ui.adsabs.harvard.edu/abs/<BIBCODE>/abstract.

Common mistakes

Author names need quotes and "Last, First" order: author:"Chu, X", not author:Chu X. Use ^ inside the quotes for first-author.
Bibcodes are case-sensitive and exactly 19 characters; don't reconstruct them by hand — get them from search results.
Escape & in shell: bibcodes like 1979A&A....75..228L must be quoted in bash.
collection: filters silently drop relevant records (e.g. collection:physics excludes mission/instrument papers indexed only under astronomy). Search without a collection filter first; add one only if the unfiltered results are noisy.
Rate limits: 5000 requests/day for search. The script warns when <100 remain. Batch bibcodes into one get/export/metrics call instead of looping.
Large downloads: --rows max is 2000 per request; paginate with --start.

Direct API (no script)

If the script is unavailable, call the API directly:

curl -H "Authorization: Bearer $ADS_API_TOKEN" \
  "https://api.adsabs.harvard.edu/v1/search/query?q=exoplanet+atmospheres&fl=bibcode,title,year&rows=5"

Full API docs: https://ui.adsabs.harvard.edu/help/api/api-docs.html

What ships with it: 2 files

13.9 KB alongside SKILL.md, 1 of them executable

references/

search-syntax.md4.7 KB

scripts/

ads.pyruns9.2 KB

Gives 0 of the 12 instructions most marketing audience skills give in ~1.4k tokens

Counted across 690 of the 894 authors here whose files we hold, read 2026-08-07

Apply Poppins font to headingsin 41 of 690, across 6 files
Apply Lora font to body textin 41 of 690, across 6 files
Use Arial fallback for headingsin 39 of 690, across 4 files
Use Georgia fallback for body textin 39 of 690, across 4 files
Maintain text hierarchy and formattingin 39 of 690, across 4 files
Use accent colors for non-text shapesin 38 of 690, across 3 files
Use RGB values for precise color matchingin 38 of 690, across 3 files
Use brand colors for primary text and backgroundsin 36 of 690, across 1 file
Read product marketing context file before asking questions, starting, or auditingin 35 of 690, across 23 files
Use active voice instead of passive voicein 26 of 690, across 10 files
Implement or generate appropriate JSON-LD structured datain 24 of 690, across 17 files
Prioritize clarity over clevernessin 22 of 690, across 8 files

Said here and by no other author read

use ads api token from environment or config file
run python3 scripts/ads.py with subcommands for searches
use setup subcommand to save api token
format author names as last name first in quotes
get bibcodes from search results never by hand
search without collection filters first

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.