Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Draw.io 配图专家：用于生成和导出 `.drawio` 图表。在用户要求 draw.io、diagrams.net、流程图、架构图、时序图、ER 图、状态图、思维导图，或明确提到 `.drawio`、PNG、SVG、PDF 导出、技术文章配图时使用。默认输出专业技术风格，并遵循 floracat-architecture-diagram 的视觉规范。
.claude/skills/snailclimb-drawio-chart/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 44% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 159% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 52% | 0% |
这个 skill 负责把用户的图表需求转成 .drawio 文件，并在需要时导出为 PNG、SVG 或 PDF。
主文件只保留任务路由、工作流和验收规则。详细样式、XML 模板、导出命令和示例按需读取 references/。
在这些场景下优先使用本 skill：
draw.io、diagrams.net 或 .drawio不适合本 skill 的情况：
.drawio按任务需要读取对应参考文件，不要默认把所有参考内容都塞进上下文：
| 需求 | 读取文件 | | --- | --- | | 颜色、字体、节点语义、连线风格、视觉统一规范 | references/style-spec.md | | XML 结构规则、文本标签规则、节点/连线模板、不同图表类型的结构建议 | references/xml-and-layout.md | | 导出命令、文件命名、打开方式、交付规则 | references/export-and-files.md | | 常见提示词、单图/多图文章配图示例、批量页面生成方式 | references/use-cases.md |
.drawio 原生文件，再根据用户要求决定是否导出。.drawio 作为源文件，除非用户明确要求删除。mxCell.value 默认使用纯文本，避免嵌入 HTML 标签。按内容结构选择图表，不要机械套模板：
如果用户指定了图表类型，优先按用户要求执行；只有在用户明显选错图表类型时，才简短提醒并给出替代建议。
先判断当前属于哪一种：
.drawio至少明确这些信息：
如果用户信息不完整，但可以合理推断，就直接推断并在结果里说明；只有缺少关键结构信息时才追问。
references/style-spec.mdreferences/xml-and-layout.mdreferences/export-and-files.mdreferences/use-cases.md生成前先在内部明确：
diagram page不要一边写 XML 一边临时想结构。
.drawio生成时遵循这些顺序：
.drawiodrawio CLI，要明确告知无法自动导出，但 .drawio 已生成当用户给一篇文章并要求“生成多张技术配图”时，按这个规则处理：
diagram page更细的示例和 prompt 写法，读取 references/use-cases.md。
交付前至少检查：
mxCell.value 中是否混入 HTML 标签.drawio，报告失败原因和可手动执行的导出命令| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 22,944 | 23,736 | +3% | 1 | 1 | 0% | 4,600 | 6,690 | +45% | 0 | 0 | — |
case-16 | fail→pass | 28,771 | 29,112 | +1% | 1 | 1 | 0% | 6,175 | 7,606 | +23% | 0 | 0 | — |
case-09 | pass→pass | 27,018 | 31,129 | +15% | 1 | 1 | 0% | 6,182 | 7,613 | +23% | 0 | 0 | — |
case-02 | fail→pass | 27,569 | 27,047 | -2% | 1 | 1 | 0% | 5,292 | 7,630 | +44% | 0 | 0 | — |
case-01 | fail→fail | 25,806 | 27,718 | +7% | 1 | 1 | 0% | 5,258 | 7,634 | +45% | 0 | 0 | — |
case-04 | pass→pass | 23,260 | 24,287 | +4% | 1 | 1 | 0% | 4,921 | 6,606 | +34% | 0 | 0 | — |
case-05 | fail→pass | 28,755 | 31,677 | +10% | 1 | 1 | 0% | 6,170 | 7,601 | +23% | 0 | 0 | — |
case-06 | pass→pass | 23,780 | 28,241 | +19% | 1 | 1 | 0% | 5,095 | 7,627 | +50% | 0 | 0 | — |
case-07 | pass→pass | 28,418 | 22,616 | -20% | 1 | 1 | 0% | 6,042 | 6,533 | +8% | 0 | 0 | — |
case-08 | pass→pass | 17,582 | 28,663 | +63% | 1 | 1 | 0% | 3,050 | 7,607 | +149% | 0 | 0 | — |
case-10 | fail→pass | 18,890 | 28,489 | +51% | 1 | 1 | 0% | 2,920 | 7,556 | +159% | 0 | 0 | — |
case-11 | fail→fail | 11,025 | 17,192 | +56% | 1 | 1 | 0% | 1,666 | 4,677 | +181% | 0 | 0 | — |
case-12 | pass→pass | 11,719 | 29,533 | +152% | 1 | 1 | 0% | 1,772 | 7,614 | +330% | 0 | 0 | — |
case-13 | pass→pass | 27,825 | 27,427 | -1% | 1 | 1 | 0% | 5,702 | 7,610 | +33% | 0 | 0 | — |
case-14 | pass→pass | 13,131 | 16,686 | +27% | 1 | 1 | 0% | 2,724 | 4,869 | +79% | 0 | 0 | — |
case-15 | fail→pass | 23,531 | 24,566 | +4% | 1 | 1 | 0% | 4,216 | 6,408 | +52% | 0 | 0 | — |
case-17 | fail→pass | 4,644 | 8,759 | +89% | 1 | 1 | 0% | 772 | 3,049 | +295% | 0 | 0 | — |
case-18 | pass→pass | 26,366 | 28,699 | +9% | 1 | 1 | 0% | 4,772 | 6,927 | +45% | 0 | 0 | — |
case-19 | pass→pass | 13,815 | 5,505 | -60% | 1 | 1 | 0% | 2,100 | 2,247 | +7% | 0 | 0 | — |
case-20 | pass→pass | 3,329 | 5,613 | +69% | 1 | 1 | 0% | 573 | 2,300 | +301% | 0 | 0 | — |
case-21 | pass→pass | 10,405 | 6,429 | -38% | 1 | 1 | 0% | 1,417 | 2,398 | +69% | 0 | 0 | — |
case-22 | pass→pass | 18,258 | 19,560 | +7% | 1 | 1 | 0% | 2,771 | 4,328 | +56% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +27 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.