Microsoft's Flint argues agents should emit a chart DSL, not chart code.

microsoft/flint-chart hit 3,012 stars and 264 HN points on August 1. The thesis: asking an agent to configure scales, axes, spacing and layout directly is where agent-generated charts break, so Flint has agents emit a compact semantic spec and its compiler derives the rest from data, semantic types, chart type and encodings. 70+ semantic types across Vega-Lite, ECharts, Chart.js, Plotly and native Excel. No published eval compares Flint specs against direct code generation, which is exactly the number that would settle the argument.