菜单
Generate downloadable SVG charts from structured data, including radar, bar, line, pie, doughnut, scatter, and score distribution charts.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/zgiai/zgi --skill chart-generator命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Navigate the visible console to a safe internal page when the user asks to open, switch to, or go to another module.
Generate downloadable files from provided content, including basic exports, SVG files, styled DOCX documents, HTML-rendered PDFs, and static editable PPTX decks.
Call workflows explicitly bound to the current Agent as structured process tools.
基于 SOC 职业分类
正在显示 SKILL.md
| name | chart-generator |
| description | Generate downloadable SVG charts from structured data, including radar, bar, line, pie, doughnut, scatter, and score distribution charts. |
| when_to_use | Use this skill when the user asks to create, export, or generate a chart, graph, radar chart, spider chart, bar chart, line chart, pie chart, doughnut chart, scatter chart, score distribution chart, score chart, comparison chart, or data visualization from provided data. For casual, vague, incomplete, or non-structured chart and data visualization requests, first route through prompt-professionalizer to optimize the visualization prompt and extract chart requirements, then call this skill. |
| provider_type | builtin |
| provider_id | chart_generator |
| runtime_type | tool |
| tools | ["generate_chart"] |
| supported_callers | ["aichat","agent"] |
| max_calls_per_turn | 5 |
| timeout_seconds | 5 |
| tool_governance | {"generate_chart":{"tool_id":"chart.generate","skill_id":"chart-generator","domain":"files","effect":"create","asset_type":"file","risk_level":"medium","requires_asset_resolution":false,"reversible":true,"bulk_sensitive":false,"external_side_effect":false,"permission_scopes":["file:create"],"default_approval_policy":"never_ask","allowed_permission_tiers":["basic","advanced","full"],"audit_required":true,"idempotency_required":false}} |
| display | {"icon":"chart-no-axes-combined","category":"data_analysis","scenarios":["data_insights","content_creation","office_collaboration"],"label":{"en_US":"Chart Generator","zh_Hans":"图表生成器"},"description":{"en_US":"Designed for reports and data analysis that need visual evidence; turns structured data into SVG radar, bar, line, pie, doughnut, scatter, or distribution charts.","zh_Hans":"适用于汇报和数据分析中的可视化展示，可将结构化数据生成雷达图、柱状图、折线图、饼图、环形图、散点图或分布图。"},"when_to_use":{"en_US":"Use when the answer should include a generated chart artifact.","zh_Hans":"当回答需要生成图表文件时使用。"},"tags":{"en_US":["Chart","Visualization","Data"],"zh_Hans":["图表","可视化","数据"]}} |
Use this skill to generate downloadable SVG chart artifacts from structured data. Supported chart types are radar, bar, line, pie, doughnut, scatter, and score_distribution; the structure is designed so future chart types can be added with a reference document and builtin renderer.
radar: radar/spider charts for multi-dimensional score profiles and personal-vs-average comparisons.bar: bar/column charts for category comparison, grouped comparison, and score comparison.line: line charts for trends, ordered stages, time series, and score changes.pie: pie charts for proportions, composition, and share-of-whole views.doughnut: doughnut/ring charts for proportions with a visible center total.scatter: scatter plots for two-variable relationships such as rank vs score.score_distribution: score band distribution charts from raw scores or precomputed band counts.chart_type: radar, bar, line, pie, doughnut, scatter, or score_distribution.prompt-professionalizer to produce an optimized data visualization prompt and chart requirements.request_user_input before calling generate_chart.generate_chart.call_skill_tool with tool_name set to generate_chart.file-manager/save_file_to_managementtool_file_idfile_idWhen any required decision is missing or ambiguous, call request_user_input instead of writing a plain clarification message. This ensures the UI renders the clarification as a structured confirmation card with optional quick replies.
Generic chart requests are incomplete even when the data can be parsed. Do not infer bar just because the data is ranking or category scores, do not infer line just because the data is ordered, do not infer pie or doughnut just because values can be totaled, do not infer scatter just because values have order, and do not infer radar just because the data contains multiple values. If the user did not explicitly name the chart type, ask.
For generic chart requests, chart type, chart title, and rendering style are required decisions. Ask for them before reading a chart reference or calling generate_chart.
Use a brief message explaining what needs confirmation, then provide 1-4 focused questions. Include options only when each option is a concrete answer that can be used directly. Omit options for open-ended questions such as the chart title, because the user can type freely.
After calling request_user_input, stop the turn and wait for the user's answer. Do not call generate_chart in the same turn.
Ask about:
radar, bar, line, pie, doughnut, scatter, or score_distribution.simple, business, teaching, comparison.score_distribution but did not specify the band rules.Example generic request that must trigger request_user_input:
1张三98 2孙八98 3李四88 4吴十82 5钱七80 6周九78 7王五74 8赵六67 生成图表
Do not answer "I will use a bar chart" for this kind of request. Ask the user to confirm chart type, title, and style first.
Example request_user_input payload:
{
"message": "I can generate the chart, but need to confirm a few details first.",
"questions": [
{
"id": "chart_type",
"question": "Which chart type should I generate: bar, line, radar, pie, doughnut, scatter, or score_distribution?"
},
{
"id": "title",
"question": "What chart title should be shown?"
},
{
"id": "style",
"question": "Which rendering style should I use?",
"options": [
{ "label": "simple" },
{ "label": "business" },
{ "label": "teaching" },
{ "label": "comparison" }
]
},
{
"id": "show_values",
"question": "Should values be displayed on the chart?",
"options": [
{ "label": "show values" },
{ "label": "hide values" }
]
}
]
}
Read exactly one reference after choosing the chart type:
| Requested chart | Read reference |
|---|---|
radar, spider, score profile, subject ability chart, personal-vs-average comparison | chart-radar.md |
bar, column, category comparison, grouped comparison | chart-bar.md |
line, trend, time series, score change, progress over attempts | chart-line.md |
pie, proportion, composition, share of whole | chart-pie.md |
doughnut, donut, ring chart, share of whole with center total | chart-doughnut.md |
scatter, scatter plot, two-variable relationship | chart-scatter.md |
score_distribution, score band distribution, grade range count | chart-score-distribution.md |
If the user requests a chart type that is not listed, say it is not supported yet and offer to structure the data for a future chart type.
generate_chart accepts:
chart_type: required. Supported values: radar, bar, line, pie, doughnut, scatter, score_distribution.title: optional chart title.output_filename: optional ASCII filename without extension. Defaults to chart.data: required chart-specific data object.options: optional rendering options. Common keys: width, height, style, show_values, legend, grid.options.show_labels controls point labels.lifecycle: optional file lifecycle, persistent or temporary. Defaults to persistent.generate_chart, use prompt-professionalizer when the user's request is casual, vague, incomplete, or not already structured for chart generation. Direct tool calls are allowed only when the chart type, data mapping, title or purpose, and key rendering requirements are already complete.generate_chart until the selected chart reference has been read.request_user_input.