Open Science Skills plugin
43 skills for Claude Code along the research cycle: surveys, experiments, topic models, reporting and manuscript review
- 70 ★ on the repository
- Last commit
Minimal access
Why this level
- The skills are text instructions, but some of them start subagents and external models
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/open-science-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills are text instructions, but some of them start subagents and external models
This is third-party code. Review the repository files before installing.
What it does
A library for computational social science and digital humanities. The `oss` plugin installs 43 skills in Claude Code, and the parallel Codex set holds 42. The skills follow the research stages: survey and conjoint experiment design, list experiments, survey data cleaning and audit, topic modeling, LLM text classification, OCR through vision models, citation and fact checks. Another group covers writing and reporting: the manuscript, tables and figures, referee responses, replication packages and pre-registration. There is multi-model orchestration, an independent model committee and a pre-submission review. Every skill rests on published methods sources listed in `SOURCES.md`. A skill loads into context only when it runs, while just its description stays resident.
Who it is for. Researchers in political science, sociology and digital humanities who run analysis and manuscripts with Claude Code or Codex.
Good fit when
- A conjoint experiment is being prepared and the design check, cleaning and diagnostics are needed
- A survey goes through an audit before analysis and the data comes from Qualtrics
- A text corpus is classified by a language model with logprob calibration
- A manuscript gets a review, citation check and fact check before journal submission
- A replication package and a pre-registration are due
Not a fit when
- The work lies outside social science: the skills are tuned to its methods and reporting standards
- You need one narrow technique: the plugin installs all 43 skills
- You do not want to spend tokens on calls to other models: some skills start subagents
Example request
Run /oss:conjoint-design and help me plan a conjoint experiment on attitudes to immigrationLimitations
The skills are tuned to the APSA, JARS, DA-RT, TOP and FAIR standards. Orchestration, the model committee and the `--codex` mode start subagents or external models and spend tokens, so they run only on an explicit request. Codex has no plugin and uses the skills library instead, which leaves out the `presubmit` skill. The author calls the set a personal working toolkit that grows as sources are added, so the skill list changes between versions.
How to disable. Run `claude plugin uninstall oss@open-science-skills`. If you copied skills by hand, delete their folders from `.claude/skills/`.
FAQ
How do I install Open Science Skills in Claude Code?
Run `claude plugin marketplace add scdenney/open-science-skills` and `claude plugin install oss@open-science-skills`. Add `--scope project` for one project only.
How do I call a skill?
Name it with the prefix, for example `/oss:conjoint-design`, or describe the task in plain words. In Codex the call looks like `$skill-name`.
Does the set work in Codex?
Yes, Codex has a parallel library of 42 skills in the `codex/` folder. It installs without a plugin.
Where do the skill methods come from?
Each rests on published methods papers, and the sources are listed in `SOURCES.md` and in the repository knowledge base.
Can I install only some skills?
Yes, `plugin/scripts/install.sh` takes `--skill` with names or an interactive pick. The target folder is set with `--target`.
Discussion
No questions or reviews yet.
Related
PluginsAround 200 skills for Claude Code on bioinformatics, chemistry and biostatistics, from RNA-seq analysis to molecular docking
26 skills and four reviewer agents for science in Claude Code: figure and statistics review, literature search, work handoff.
StatsClaw assembles nine Claude Code agents to build statistical packages from a paper or formulas with isolated pipelines.
Claude Code skills for empirical economics: code and replication audit, bibliography checks, PDF reading and slide decks.
31 skills for Claude Code and Codex for the analyst workflow: data checks, cohorts, funnels, A/B tests, dashboards, summaries
16 Claude Code skills for data engineers: source profiling, failure investigation, change review and documentation as web pages.
README in short
The README explains the principle: each research stage has a skill linked to methods sources. Next come a platform table with skill counts, a quick start, trigger rules and several install paths for Claude Code and Codex. Sections on the knowledge base, contributing and the license close the page.