echarts-ai-skill（echartsAI 技能）

ECharts Chart Skill

Use this skill when the user wants chart output from a short description or from table-like data.

Workflow

1. Translate the user's request into a ChartRequest JSON object.
2. If the chart type is unclear, run the recommendation command first.
3. Run the generation command to produce a stable ECharts option.
4. Run the render command when the user wants an embeddable html or svg.

Files

• Core types: src/types/chart.ts
• Chart recommendation: src/core/recommend.ts
• Option generation: src/core/spec-to-option.ts
• Rendering: src/core/render.ts
• Sample input: examples/study-progress.request.json
• Universal metadata: manifest.yaml, agents/openai.yaml

Setup

npm install

Output rules

• --out writes to an exact file path.
• --out-dir writes the default file into a directory you choose.
• desktop and home are valid aliases for --out-dir when the user explicitly asks for those locations.
• ~ is expanded to the current user's home directory.
• If no output path is provided, files default to the current working directory.

Commands

node dist/cli/recommend-chart.js --input examples\study-progress.request.json
node dist/cli/generate-chart.js --input examples\study-progress.request.json
node dist/cli/render-chart.js --input option.json --format html
node dist/cli/render-chart.js --input option.json --format svg --out D:\
eports\study-chart.svg

Default output filenames:

• Recommendation: spec.json
• Option: option.json
• HTML preview: preview.html
• SVG preview: preview.svg

ChartRequest shape

{
  "title": "Monthly study completion",
  "dataset": [
    { "day": "2026-03-01", "completionRate": 62, "targetRate": 75 },
    { "day": "2026-03-02", "completionRate": 68, "targetRate": 75 }
  ],
  "goal": "trend",
  "xField": "day",
  "yField": "completionRate",
  "series": [
    { "name": "Completion", "field": "completionRate" },
    { "name": "Target", "field": "targetRate" }
  ]
}

Guidance

• Prefer deterministic field mapping over free-form inference when the user has already named fields.
• For pie charts, keep one category field and one metric field.
• For scatter charts, require numeric xField and yField.
• If the user only gave natural language, construct the smallest valid ChartRequest before calling scripts.
• If the user needs a report artifact, render html first and svg second.