Data visualisation uplift

LLMs and default charting libraries produce visualisations with predictable failure modes: grey gridlines on white backgrounds, rainbow colour palettes with no semantic meaning, oversized legends that compete with the data, and placeholder axis labels that communicate nothing. This skill replaces those defaults with spectral data semantics, OLED-optimised dark themes, and annotation-first design that foregrounds the data itself.

The skill is rigid. Every chart produced under its governance must pass the eval criteria at the end of this file. There are no soft recommendations here — each rule exists because its violation is detectable and its correction is measurable.

When to apply

Positive triggers

Activate this skill when the task involves producing or revising data visualisations:

• "chart", "graph", "plot", "visualisation", "visualize", "data display"
• "dashboard data", "metrics", "KPI", "analytics view"
• "Recharts", "Chart.js", "D3", "charting library"
• Any task where the primary output is a chart component, dashboard panel, or data-driven visual
• Metric cards, sparklines, or single-value KPI displays