Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add SuMayaBee/DataViz-MCP --skill dataviz-plot-selectiongit clone --depth 1 https://github.com/SuMayaBee/DataViz-MCPWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/sumayabee/dataviz-mcp/dataviz-plot-selection)<a href="https://agentmods.dev/skills/sumayabee/dataviz-mcp/dataviz-plot-selection"><img src="https://agentmods.dev/badge/skills/sumayabee/dataviz-mcp/dataviz-plot-selection/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/sumayabee/dataviz-mcp/dataviz-plot-selection"><img src="https://agentmods.dev/badge/skills/sumayabee/dataviz-mcp/dataviz-plot-selection.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00059 | $0.00837 |
| Opus 5.5 | $0.00024 | $0.00335 |
| Sonnet 5.5 | $0.00012 | $0.00167 |
| Haiku 4.5 | $0.00006 | $0.00084 |
Grade A, and why
dataviz-plot-selection scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.
DataViz Plot Selection
Choose the plot from the user's question and the data's structure. Do not select a chart only because a library supports it. Preserve any chart type, columns, aggregation, or style the user explicitly requests unless it would be misleading or cannot render.
Workflow
- Identify the analytical intent: trend, comparison, distribution, relationship, composition, geography, hierarchy, flow, or multidimensional exploration.
- Inspect the relevant columns, types, missing values, cardinality, row count, units, and time
grain or labeled dimensions. When the user only asks about the data, call
load_data(..., visualize=False)and explain the returned analysis and recommendations without rendering. When the user asks for automatic plotting, callload_data(..., visualize=True). Useshowfor a custom chart or dashboard. - Select the chart and encodings from references/use-cases.md.
- Prefer hvPlot for ordinary DataFrame plots, HoloViews for composable or multidimensional views, Datashader for dense large data, and Panel for widgets or multi-view dashboards.
- Label axes and units, use a descriptive title, keep legends readable, and state material filtering or aggregation in the description.
- Render, then use
screenshotto check clipping, empty output, unreadable labels, misleading scales, and excessive overlap before presenting the visualization.
Data-scale routing
- Treat
localandremoteas source locations, not data shapes. Classify either source as normal, multidimensional, or large after metadata inspection. - NetCDF data: use
load_data; it inspects Xarray dimensions and variables, chooses a numeric data variable, and creates a bounded tidy view. For a time-latitude-longitude cube, prefer one time slice with longitude and latitude as position and the selected variable as color. - Data with at least 100,000 rows or cells, 100 MB on disk, about 500 MB estimated in memory, or
500 columns: use the Datashader strategy. Prefer
rasterize=Truefor numeric point density so axes, hover, and color mapping remain interactive. - Data with at least three numeric dimensions or labeled array dimensions: use HoloViews and encode the additional dimensions through color, size, facets, or linked views. Do not force every dimension into one overloaded plot.
- Smaller normal data: use hvPlot unless the requested interaction needs HoloViews or Panel.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 7d ago First seen · 63 lines · 59 tokens per session scan A 695688fddb3d
dataviz-plot-selection is a skill published in the GitHub repository SuMayaBee/DataViz-MCP (3 stars, last pushed 7d ago), licensed BSD-3-Clause. It adds 59 tokens to every session and 837 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-10.
Other skills, from other repositories
panel-custom-components
Build custom Panel components using JSComponent (vanilla JS, web components), ReactComponent (React/JSX), AnyWidgetComponent (AnyWidget spec for cross-platform), or MaterialUIComponent (Material UI themed). Use when wrapping JS libraries, creating interactive widgets, or building themed components. Includes decision…
panel
Best practices for developing tools, dashboards and interactive data apps with HoloViz Panel. Create reactive, component-based UIs with widgets, layouts, templates, and real-time updates. Use when developing interactive data exploration tools, dashboards, data apps, or any interactive Python web application. Supports…
holoviews
Best practices for developing advanced, interactive, and publication-quality data visualizations using HoloViz HoloViews.
panel-holoviews
Best practices for integrating HoloViews and hvPlot visualizations into Panel applications. Use when embedding HoloViews/hvPlot plots in Panel panes, preserving zoom/pan state across data refreshes with DynamicMap, composing DynamicMap overlays without type errors, using HoloViews streams (Selection1D, RangeXY, Tap…
param
Use when building Python classes with validated, typed parameters using the Param library. Triggers include creating configuration classes, building reusable components with state, implementing reactive dependencies between parameters, adding type-safe attributes with bounds/constraints, creating testable…
panel-material-ui
Best practices for developing modern looking tools, dashboards and data apps using HoloViz Panel and Panel Material UI components.