From superpaper
Use when a new dataset arrives and you need to understand it before formal analysis — generates an exploratory analysis spec and a ready-to-copy prompt for a coding agent
How this skill is triggered — by the user, by Claude, or both
Slash command
/superpaper:data-exploreThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
- User says "I have a new dataset" or "I need to understand my data"
SuperPaper does NOT run code itself. This skill generates a spec and a ready-to-copy prompt for a coding agent (Claude Code, Codex, or Cursor). The coding agent does the actual exploration.
Before generating any output, you MUST confirm the following. If the answers are already clear from the user's triggering message, verify with a single confirmation question. If any are unclear, ASK BEFORE PROCEEDING. Do not embed these as "clarifying questions" in the output document — ask them conversationally first.
Mandatory confirmations:
Only proceed to the next steps once these are confirmed.
Ask the user for:
If the user has already provided this context, skip the questions and proceed.
Produce a structured Exploratory Data Analysis Spec covering:
Present in this order:
naniar if R)Optional: Network structure (if dataset includes user-to-user relationships or repost chains):
networkx, R igraphPresent this block for the user to copy to their coding agent:
Project: [repo path or description]
Task: Exploratory data analysis
Dataset: [location, format, key variables]
Questions: [what to explore — paste from Key Questions above]
Expected output:
- Figures saved to references/results/<key>/figures/
- Summary statistics to references/results/<key>/summary-stats.csv
- Data quality report to references/results/<key>/data-quality.md
Notes:
- Do not modify the source data
- Use [Python/R — specify] for all analysis
- Save all figures at 300 DPI
- Include a brief interpretation comment with each figure
Fill in the bracketed fields based on what the user told you. When choosing a <key> slug for the results directory, use: lowercase, hyphenated, format <topic>-<type>[-<year>]. Examples: weibo-eda-2023, polarization-model, sentiment-baseline.
Save this spec to references/results/<key>/spec.md before handing off to the coding agent. This records what was asked, enabling reproducibility if the coding agent is re-run or results change.
After presenting the spec and prompt, tell the user:
Drop figures in
references/results/<key>/figures/and runingestonce the coding agent is done. SuperPaper will organize the results and link them to the relevant chapter sections.
<key> slug in references/results/<key>/ should be consistent with any later analysis run on the same datasetingest to organize them into the projectProvides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
npx claudepluginhub houx15/superpaper --plugin superpaper