Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Fills all [FILL:] placeholders across the template to initialize a new research project. Use when setting up a freshly cloned project.
.claude/skills/brycewang-stanford-init/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -47% | 0% |
Fill in all [FILL:] placeholders across the template to set up a new research project.
[FILL: patterns and group them by file:bash grep -rn "\[FILL:" --include="*.md" --include="*.qmd" --include="*.toml" --include="*.tex" --include="*.yml" . Report how many placeholders exist and in which files.
index.qmd, README.md, CLAUDE.md, pyproject.toml)index.qmd)index.qmd)index.qmd)CLAUDE.md: Idea / Data collection / Analysis / Writing / Revision)CLAUDE.md and README.md)README.md)pyproject.toml name field, e.g., regional-gdp-study)index.qmd — title, subtitle, authors (full YAML array with affiliations/ORCID/email), abstract, keywordsREADME.md — project title, description paragraph, data section, repository URL, Quick Start clone URLCLAUDE.md — Project Context table: title, authors, stage, data sourcepyproject.toml — name, description, authors_quarto.yml — no [FILL:] placeholders by default, but verifytemplates/chadManuscript/manuscript.tex — author names and keywords[FILL:] placeholders:bash grep -rn "\[FILL:" --include="*.md" --include="*.qmd" --include="*.toml" --include="*.tex" --include="*.yml" . Report how many remain and where (some are expected in section bodies like "Describe the data...").
bash scripts/render.sh to regenerate latex/index.tex with the real content.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | fail→pass | 10,583 | 6,535 | -38% | 1 | 1 | 0% | 1,847 | 1,842 | -0% | 0 | 0 | — |
case-01 | fail→fail | 4,359 | 2,099 | -52% | 1 | 1 | 0% | 242 | 855 | +253% | 0 | 0 | — |
case-02 | fail→fail | 17,737 | 3,188 | -82% | 1 | 1 | 0% | 2,912 | 1,031 | -65% | 0 | 0 | — |
case-03 | fail→fail | 8,034 | 2,194 | -73% | 1 | 1 | 0% | 1,354 | 837 | -38% | 0 | 0 | — |
case-04 | fail→pass | 8,468 | 3,321 | -61% | 1 | 1 | 0% | 1,580 | 1,275 | -19% | 0 | 0 | — |
case-05 | fail→pass | 8,801 | 2,798 | -68% | 1 | 1 | 0% | 1,788 | 1,112 | -38% | 0 | 0 | — |
case-06 | fail→pass | 20,698 | 1,903 | -91% | 1 | 1 | 0% | 1,823 | 1,004 | -45% | 0 | 0 | — |
case-07 | pass→pass | 11,436 | 4,183 | -63% | 1 | 1 | 0% | 1,910 | 1,402 | -27% | 0 | 0 | — |
case-08 | pass→pass | 14,581 | 7,652 | -48% | 1 | 1 | 0% | 2,895 | 2,106 | -27% | 0 | 0 | — |
case-09 | pass→pass | 12,317 | 4,105 | -67% | 1 | 1 | 0% | 2,298 | 1,364 | -41% | 0 | 0 | — |
case-11 | fail→pass | 12,720 | 3,295 | -74% | 1 | 1 | 0% | 2,212 | 1,162 | -47% | 0 | 0 | — |
case-12 | fail→pass | 9,577 | 2,469 | -74% | 1 | 1 | 0% | 1,675 | 1,026 | -39% | 0 | 0 | — |
case-13 | fail→pass | 8,466 | 2,105 | -75% | 1 | 1 | 0% | 1,537 | 1,015 | -34% | 0 | 0 | — |
case-14 | fail→fail | 6,992 | 1,618 | -77% | 1 | 1 | 0% | 1,226 | 839 | -32% | 0 | 0 | — |
case-15 | fail→pass | 10,812 | 5,494 | -49% | 1 | 1 | 0% | 2,113 | 1,540 | -27% | 0 | 0 | — |
case-16 | fail→pass | 10,322 | 3,147 | -70% | 1 | 1 | 0% | 2,345 | 1,206 | -49% | 0 | 0 | — |
case-17 | pass→pass | 10,993 | 4,522 | -59% | 1 | 1 | 0% | 2,332 | 1,635 | -30% | 0 | 0 | — |
case-18 | pass→pass | 6,782 | 1,720 | -75% | 1 | 1 | 0% | 1,376 | 954 | -31% | 0 | 0 | — |
case-19 | fail→pass | 10,476 | 4,003 | -62% | 1 | 1 | 0% | 1,739 | 1,479 | -15% | 0 | 0 | — |
case-20 | pass→pass | 12,069 | 11,299 | -6% | 1 | 1 | 0% | 2,418 | 2,820 | +17% | 0 | 0 | — |
case-21 | pass→pass | 13,989 | 9,183 | -34% | 1 | 1 | 0% | 2,729 | 2,437 | -11% | 0 | 0 | — |
case-22 | pass→pass | 9,030 | 5,035 | -44% | 1 | 1 | 0% | 1,886 | 1,595 | -15% | 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, and 21 counted toward the lift figure. The other 1 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 +45 percentage points is the difference between those two pass rates over the 21 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.