Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →初始化中文毕业论文工作区：检查学校模板与已有材料的放置位置，明确提示当前还缺什么，建立 `codex_md/` / `claude_md/` / `tmp_layout*/` 等中间层目录，并生成材料盘点与初始工作文件。 **Trigger**: 毕业论文初始化, thesis workspace, 中文毕业论文准备, 模板归位, 材料盘点, 初始化论文工程. **Use when**: 你要开始一条毕业论文重构流程，手头已经有学校模板、旧 `tex`、PDF、Overleaf 源稿、bib 或图表材料，需要先把工程和中间层搭起来。
.claude/skills/willoscar-thesis-workspace-init/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 50% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -15% | 0% |
初始化一条中文毕业论文重构流程所需的工作区，而不是只复制通用 workspace 模板。
这个 skill 负责四件事：
建议至少具备以下材料中的一部分：
main.texchapters/*.texpdf/ 下的已发表 / 已投稿论文Overleaf_ref/ 下的源稿或修回稿references/ 下的 bib / 样式文件codex_md/material_index.mdcodex_md/material_readiness.mdcodex_md/missing_info.mdcodex_md/question_list.mdcodex_md/00_thesis_outline.mdclaude_md/review_checklist.mdtmp_layout/tmp_layout2/mermaid/（位于 codex_md/mermaid/）Always read:
references/overview.mdreferences/material-placement.mdMachine-readable contract:
assets/workspace_contract.jsonreferences/material-placement.mdcodex_md/、claude_md/、tmp_layout/、tmp_layout2/codex_md/material_index.mdcodex_md/material_readiness.mdcodex_md/missing_info.mdcodex_md/question_list.mdcodex_md/00_thesis_outline.mduv run python .codex/skills/thesis-workspace-init/scripts/run.py --workspace <workspace>--workspace <dir>workspace-init，再运行这个 skill 补齐毕业论文专用结构。遇到以下情况应停止并先修正：
main.tex 完全缺失且也没有可替代的源稿入口| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 18,011 | 25,995 | +44% | 1 | 1 | 0% | 2,870 | 4,311 | +50% | 0 | 0 | — |
case-02 | fail→fail | 19,502 | 7,155 | -63% | 1 | 1 | 0% | 2,831 | 1,114 | -61% | 0 | 0 | — |
case-03 | fail→fail | 18,650 | 5,959 | -68% | 1 | 1 | 0% | 2,678 | 956 | -64% | 0 | 0 | — |
case-04 | fail→fail | 12,056 | 14,096 | +17% | 1 | 1 | 0% | 1,645 | 2,592 | +58% | 0 | 0 | — |
case-05 | fail→fail | 5,869 | 13,563 | +131% | 1 | 1 | 0% | 905 | 2,898 | +220% | 0 | 0 | — |
case-06 | fail→fail | 9,838 | 18,804 | +91% | 1 | 1 | 0% | 1,536 | 3,587 | +134% | 0 | 0 | — |
case-07 | fail→fail | 18,724 | 8,322 | -56% | 1 | 1 | 0% | 3,074 | 2,033 | -34% | 0 | 0 | — |
case-08 | fail→pass | 13,595 | 13,865 | +2% | 1 | 1 | 0% | 2,037 | 2,735 | +34% | 0 | 0 | — |
case-09 | fail→pass | 14,943 | 3,748 | -75% | 1 | 1 | 0% | 2,318 | 1,350 | -42% | 0 | 0 | — |
case-19 | fail→pass | 11,830 | 2,613 | -78% | 1 | 1 | 0% | 1,975 | 1,203 | -39% | 0 | 0 | — |
case-10 | pass→pass | 9,479 | 6,097 | -36% | 1 | 1 | 0% | 1,391 | 1,746 | +26% | 0 | 0 | — |
case-11 | fail→pass | 10,073 | 3,798 | -62% | 1 | 1 | 0% | 1,495 | 1,269 | -15% | 0 | 0 | — |
case-12 | fail→pass | 15,304 | 2,877 | -81% | 1 | 1 | 0% | 2,013 | 1,167 | -42% | 0 | 0 | — |
case-13 | fail→pass | 10,382 | 3,740 | -64% | 1 | 1 | 0% | 1,549 | 1,294 | -16% | 0 | 0 | — |
case-14 | fail→pass | 11,580 | 2,218 | -81% | 1 | 1 | 0% | 1,822 | 1,071 | -41% | 0 | 0 | — |
case-15 | fail→pass | 11,906 | 2,305 | -81% | 1 | 1 | 0% | 1,575 | 1,164 | -26% | 0 | 0 | — |
case-16 | fail→pass | 12,502 | 3,880 | -69% | 1 | 1 | 0% | 1,888 | 1,286 | -32% | 0 | 0 | — |
case-17 | fail→pass | 8,569 | 2,293 | -73% | 1 | 1 | 0% | 1,338 | 1,100 | -18% | 0 | 0 | — |
case-18 | fail→pass | 21,027 | 11,802 | -44% | 1 | 1 | 0% | 2,988 | 2,657 | -11% | 0 | 0 | — |
case-20 | pass→pass | 14,159 | 5,468 | -61% | 1 | 1 | 0% | 2,075 | 1,611 | -22% | 0 | 0 | — |
case-21 | pass→pass | 12,675 | 3,972 | -69% | 1 | 1 | 0% | 1,709 | 1,324 | -23% | 0 | 0 | — |
case-22 | fail→pass | 16,487 | 2,988 | -82% | 1 | 1 | 0% | 2,319 | 1,223 | -47% | 0 | 0 | — |
case-23 | fail→pass | 11,534 | 1,789 | -84% | 1 | 1 | 0% | 1,804 | 981 | -46% | 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. 23 cases were attempted, and 21 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +61 percentage points is the difference between those two pass rates over the 21 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.