This skill should be used when the user wants to visualize data. It intelligently selects the most suitable chart type from 26 available options, extracts parameters based on detailed specifications, and generates a chart image using a JavaScript script.
A workflow for turning data into chart images, handling chart-type selection, parameter extraction, and image generation. It selects from 26 chart types based on the shape of the data, using built-in guidelines: line, area, or dual-axes charts for time series; bar, column, or histogram charts for comparisons; pie or treemap for part-to-whole; scatter, sankey, or venn for relationships and flow; district, pin, or path maps for geography; organization charts and mind maps for hierarchies; and a range of specialized options including radar, funnel, liquid, word cloud, boxplot, violin, network graph, fishbone, flow diagram, and spreadsheet.
Use it whenever the user wants a rendered chart rather than a textual description of their data. Once a chart type is chosen, it reads the matching file in the references directory (for example generate_line_chart.md) to learn the required and optional fields, then maps the user data into the expected args format.
Generation is performed by invoking scripts/generate.js with a JSON payload containing a tool name and an args object of data, title, theme, and style, run as node ./scripts/generate.js followed by the payload JSON. The script outputs the URL of the generated chart image, and the skill returns both that URL and the complete args specification used, keeping the result reproducible. It requires Node.js 18.0.0 or later. The SKILL.md is provided by antvis chart-visualization-skills under the MIT License, and detailed per-chart specifications live in the references directory.
26, spanning time series, comparisons, part-to-whole, relationships and flow, maps, hierarchies, and specialized statistical charts.
By invoking scripts/generate.js with a JSON payload (a tool name plus an args object), run as node ./scripts/generate.js followed by the payload JSON.
Read the corresponding file in the references directory, for example generate_line_chart.md, which lists the required and optional fields.
The URL of the generated chart image, which the skill returns alongside the complete args specification used to produce it.
Node.js version 18.0.0 or later, since chart generation runs through the generate.js script.
Quick Setup:
.claude/skills/Repository
bytedance/deer-flow