Hand it a CSV, query results or just a description and get one HTML file with KPI cards, Chart.js charts, filters and a sortable table that opens in any browser.
Install the skill or give me an instruction how to install it - build-dashboard from https://github.com/anthropics/knowledge-work-plugins Free Open source Works in Claude, Cursor, Codex & more
Describe the dashboard you need and hand over your data, and you get back a single self-contained HTML file that opens in any browser with no server or install. It lays out 2 to 4 KPI cards with change versus the prior period, 1 to 3 Chart.js charts for trends and breakdowns, filter dropdowns and a sortable detail table.
It works from whatever you have: results from a connected data warehouse, a pasted or uploaded CSV, or nothing but a description, in which case it builds a realistic sample dataset, labels it as sample data and tells you how to swap in the real numbers. The data is embedded as JSON, so the file works offline and you share it by sending it.
Every filter updates all charts and the table at once, numbers are formatted as currency, commas or percentages, and the layout adapts to phones and prints cleanly. For larger data it follows set limits: embed up to 10,000 rows, pre-aggregate beyond that, and paginate tables past 200 rows.
Get an executive dashboard with headline KPI cards, change versus the prior period, and charts for the trends behind them, all in one file.
Paste or upload the data and it parses, cleans and embeds it into an interactive dashboard you can open straight from your desktop.
Turn the results into a self-contained HTML report that anyone can open offline, filter and sort, with nothing to log in to.
Describe the metrics and it builds realistic sample data, marks it as sample, and explains how to swap in the real numbers later.
No. Build a Data Dashboard writes one HTML file with the data embedded as JSON. It opens directly in a browser, works offline, and loads only Chart.js from a CDN, so you share it by sending the file.
Query results from a connected data warehouse, a pasted or uploaded CSV, or just a description. With no data, it creates a realistic sample dataset that matches your schema and labels the dashboard as sample data.
Up to about 10,000 rows embedded directly, pre-aggregating chart data past 1,000. Beyond 10,000 rows it embeds only aggregated data, and above 100,000 it recommends a BI tool. Tables paginate after 100 to 200 rows.
No. Each dashboard is a point-in-time snapshot of the data you gave it. For real-time monitoring the skill suggests a BI tool instead, but you can rebuild the file whenever the numbers change.