Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →用于生成中国软件著作权申请材料的完整工具包。支持从项目代码、文档等自动提取信息，生成软件著作权登记申请表、源代码文档（前后各30页）、用户手册和设计说明书，并自动转换为PDF文件。适用于微信小程序、Web应用、移动App、桌面应用等各类软件项目。当用户需要申请中国软件著作权时使用此skill。
.claude/skills/thomasmoreai-chinese-copyright-application/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-19 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 45% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 32% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 59% | 0% |
从以下位置收集项目信息：
微信小程序项目：
app.json - 获取软件名称（navigationBarTitleText）project.config.json - 获取appid、libVersionpackage.json - 获取版本号、描述、作者README.md - 获取详细描述、功能特性Web/Node.js项目：
package.json - 获取名称、版本、描述、作者README.md - 获取详细描述、功能特性其他项目：
pom.xml, build.gradle, Cargo.toml 等）使用 application-form-template.md 模板生成申请表。
格式要求：
必填字段：
要求：
提取策略：
代码文件优先级：
app.js, main.js, index.js）utils/, helpers/）pages/, components/）config/）使用 user-manual-template.md 模板。
内容结构：
信息来源：
使用 design-doc-template.md 模板。
内容结构：
信息来源：
所有文档以 Markdown 格式生成，并自动转换为 PDF 文件。所有输出文件将放置在专门的文件夹中。
输出文件夹：
copyright-application-materials/输出文件：
copyright-application-materials/软件著作权登记申请表.mdcopyright-application-materials/软件著作权登记申请表.pdfcopyright-application-materials/源代码文档.mdcopyright-application-materials/源代码文档.pdfcopyright-application-materials/用户手册.mdcopyright-application-materials/用户手册.pdfcopyright-application-materials/设计说明书.mdcopyright-application-materials/设计说明书.pdf| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 28,145 | 1,865 | -93% | 1 | 1 | 0% | 5,079 | 1,592 | -69% | 0 | 0 | — |
case-19 | fail→pass | 15,483 | 10,508 | -32% | 1 | 1 | 0% | 2,480 | 3,026 | +22% | 0 | 0 | — |
case-02 | fail→fail | 15,601 | 2,864 | -82% | 1 | 1 | 0% | 2,680 | 1,713 | -36% | 0 | 0 | — |
case-03 | fail→fail | 29,583 | 2,745 | -91% | 1 | 1 | 0% | 5,471 | 1,709 | -69% | 0 | 0 | — |
case-04 | pass→pass | 10,519 | 8,935 | -15% | 1 | 1 | 0% | 1,756 | 2,796 | +59% | 0 | 0 | — |
case-05 | pass→pass | 10,624 | 9,150 | -14% | 1 | 1 | 0% | 1,767 | 2,897 | +64% | 0 | 0 | — |
case-06 | pass→pass | 14,898 | 11,562 | -22% | 1 | 1 | 0% | 2,180 | 2,985 | +37% | 0 | 0 | — |
case-07 | pass→pass | 15,653 | 12,246 | -22% | 1 | 1 | 0% | 2,593 | 3,363 | +30% | 0 | 0 | — |
case-08 | fail→pass | 14,429 | 12,559 | -13% | 1 | 1 | 0% | 2,328 | 3,382 | +45% | 0 | 0 | — |
case-09 | pass→pass | 12,772 | 7,540 | -41% | 1 | 1 | 0% | 2,155 | 2,784 | +29% | 0 | 0 | — |
case-10 | pass→pass | 11,600 | 8,171 | -30% | 1 | 1 | 0% | 2,023 | 2,739 | +35% | 0 | 0 | — |
case-11 | pass→pass | 14,235 | 11,147 | -22% | 1 | 1 | 0% | 2,558 | 3,119 | +22% | 0 | 0 | — |
case-12 | pass→pass | 13,898 | 7,306 | -47% | 1 | 1 | 0% | 2,168 | 2,456 | +13% | 0 | 0 | — |
case-13 | pass→pass | 12,493 | 9,976 | -20% | 1 | 1 | 0% | 2,051 | 3,063 | +49% | 0 | 0 | — |
case-14 | pass→pass | 15,401 | 13,322 | -13% | 1 | 1 | 0% | 2,611 | 3,605 | +38% | 0 | 0 | — |
case-15 | pass→pass | 19,440 | 12,384 | -36% | 1 | 1 | 0% | 2,361 | 3,514 | +49% | 0 | 0 | — |
case-16 | pass→pass | 13,830 | 6,629 | -52% | 1 | 1 | 0% | 1,598 | 2,289 | +43% | 0 | 0 | — |
case-17 | fail→pass | 10,299 | 3,200 | -69% | 1 | 1 | 0% | 1,707 | 1,841 | +8% | 0 | 0 | — |
case-18 | fail→pass | 18,353 | 15,935 | -13% | 1 | 1 | 0% | 3,008 | 3,959 | +32% | 0 | 0 | — |
case-20 | pass→pass | 19,817 | 17,994 | -9% | 1 | 1 | 0% | 3,525 | 4,977 | +41% | 0 | 0 | — |
case-21 | pass→pass | 17,389 | 21,091 | +21% | 1 | 1 | 0% | 2,864 | 4,605 | +61% | 0 | 0 | — |
case-22 | pass→pass | 31,522 | 24,995 | -21% | 1 | 1 | 0% | 3,694 | 5,355 | +45% | 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 +18 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.