seaborn — Claude Code skill: what it does, how to install
A skill that gets your coding agent building seaborn plots from pandas DataFrames: distributions, regressions, pairplots and heatmaps, written against the current API.
npx skills add K-Dense-AI/scientific-agent-skills --skill seaborn -a claude-codeWhat seaborn does
The skill hands the agent a working method for statistical plots in seaborn. It covers the common shapes of the job: distributions, pairwise relationships, category comparisons, regression panels, heatmaps. The agent reads SKILL.md and writes code from those recipes instead of guessing at function signatures.
Two interfaces are documented. The function interface suits quick exploration and single plots. seaborn.objects gives declarative chains for layered figures. Upstream still calls the objects interface experimental in 0.13.2, so the skill leans toward the function interface unless composition genuinely shortens the plot — worth knowing before you put so.Plot into production code.
You need Python 3.8+ and seaborn 0.13.2 with NumPy, pandas and Matplotlib; the tested stack runs on Python 3.12+. scipy and statsmodels cover the heavier regression and clustering work, installed as uv pip install "seaborn[stats]==0.13.2". Version traps get their own section: errorbar replaced ci in lineplot, barplot and pointplot, while regplot and lmplot still take ci — that is why code copied from older tutorials breaks. Passing palette without hue no longer works in categorical functions.
What it helps with
- plot a distribution of one variable split by group
- compare categories with uncertainty bands without aggregating by hand
- build a pairplot across the numeric columns
- draw a correlation heatmap from an existing DataFrame
- move an old
lineplotcall fromcitoerrorbar
Example prompts for the agent
- Plot the distribution of `total_bill` by day with seaborn
- Build a pairplot from the numeric columns of my DataFrame and save it as PNG
- Switch this `lineplot` call from `ci` to `errorbar`
How to install seaborn
1Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill seaborn -a claude-code2Manually: put the files in place
~/.claude/skills/seaborn/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/seaborn/
1Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill seaborn -a codex2Manually: put the files in place
.agents/skills/seaborn/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill seaborn -a cursor2Manually: put the files in place
.cursor/skills/seaborn/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install K-Dense-AI/scientific-agent-skills seaborn2Manually: put the files in place
.github/skills/seaborn/SKILL.md1Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill seaborn -a gemini-cli2Platform CLI
gemini skills install https://github.com/K-Dense-AI/scientific-agent-skills --path skills/seabornManually: put the files in place
.devin/skills/seaborn/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill seaborn -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
From the same set120
The whole set →Similar
Questions about seaborn
How do I install the seaborn skill in Claude Code?
Run npx skills add K-Dense-AI/scientific-agent-skills --skill seaborn -a claude-code. The skill lands at ~/.claude/skills/seaborn/SKILL.md.
What does the seaborn skill do?
It gives the agent a defined way to work with seaborn: which interface to pick, how to shape data into long form, which arguments are current. The output is plotting code rather than general advice about charts.
Where does `sns.load_dataset` pull data from?
It downloads public CSVs from mwaskom/seaborn-data when the set is not cached, and no credentials are involved. For private or offline work the skill suggests loading local files with pandas and passing that DataFrame to seaborn.