Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Scaffold a research manuscript and project layout from a research brief — sections, config.yaml, src/, scripts/, tests. USE WHEN starting a new paper, new projects/ tree, manuscript from topic description, or aligning with template_code_project exemplar — even without copy-paste prompts.
.claude/skills/docxology-template-manuscript-creation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 51% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 10% | 0% |
projects/templates/template_code_project/; Active Inference multi-track → template_active_inference; prose → template_prose_project; AutoResearch/readiness loop → template_autoresearch_project; meta-template / infrastructure introspection → template_template; search → public template_search_project. Roster: docs/_generated/active_projects.md and projects/AGENTS.md.01_* vs 00_*), not generic skeleton if exemplar differs.config.yaml, preamble.md, references.bib, ordered *.md sections, glossary/references as exemplar does.@fig:, @eq:, @sec:, [@key] per docs/guides/manuscript-semantics.md. Registry: refs/labels.yaml + [[FIG:]] tokens.src/ — typed public APIs, get_logger, fixed seeds, custom exceptions where needed.scripts/ — thin orchestrators only; MPLBACKEND=Agg; print output paths.tests/ — no mocks; ≥90% on projects/<name>/src.docs/guides/new-project-one-shot-prompt.md.projects/<name>/ tree or delta from briefuv run pytest projects/<name>/tests/ --cov=... and uv run python scripts/pipeline/stage_01_test.py --project <name>bashuv sync uv run python scripts/pipeline/stage_01_test.py --project <name> uv run python -m infrastructure.validation.cli markdown projects/<name>/manuscript/
Section checklist and legacy examples: references/structure-checklist.md
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 16,217 | 19,788 | +22% | 1 | 1 | 0% | 3,149 | 4,762 | +51% | 0 | 0 | — |
case-01 | fail→fail | 25,064 | 8,339 | -67% | 1 | 1 | 0% | 6,196 | 1,385 | -78% | 0 | 0 | — |
case-02 | fail→fail | 27,643 | 26,796 | -3% | 1 | 1 | 0% | 6,182 | 6,197 | +0% | 0 | 0 | — |
case-03 | fail→fail | 22,975 | 28,826 | +25% | 1 | 1 | 0% | 5,167 | 6,961 | +35% | 0 | 0 | — |
case-04 | pass→pass | 11,021 | 3,326 | -70% | 1 | 1 | 0% | 1,628 | 1,254 | -23% | 0 | 0 | — |
case-05 | pass→pass | 6,216 | 12,152 | +95% | 1 | 1 | 0% | 1,060 | 2,254 | +113% | 0 | 0 | — |
case-06 | fail→fail | 4,924 | 4,353 | -12% | 1 | 1 | 0% | 241 | 999 | +315% | 0 | 0 | — |
case-08 | fail→pass | 14,300 | 3,065 | -79% | 1 | 1 | 0% | 2,237 | 1,416 | -37% | 0 | 0 | — |
case-09 | fail→pass | 33,637 | 3,526 | -90% | 1 | 1 | 0% | 3,049 | 1,252 | -59% | 0 | 0 | — |
case-10 | fail→fail | 8,007 | 4,259 | -47% | 1 | 1 | 0% | 1,554 | 929 | -40% | 0 | 0 | — |
case-11 | pass→pass | 12,888 | 7,872 | -39% | 1 | 1 | 0% | 2,184 | 2,088 | -4% | 0 | 0 | — |
case-12 | fail→pass | 15,434 | 7,805 | -49% | 1 | 1 | 0% | 2,561 | 2,102 | -18% | 0 | 0 | — |
case-13 | fail→pass | 12,655 | 7,158 | -43% | 1 | 1 | 0% | 1,841 | 2,021 | +10% | 0 | 0 | — |
case-14 | pass→pass | 5,115 | 2,743 | -46% | 1 | 1 | 0% | 874 | 1,144 | +31% | 0 | 0 | — |
case-15 | fail→pass | 9,828 | 1,866 | -81% | 1 | 1 | 0% | 1,414 | 1,018 | -28% | 0 | 0 | — |
case-16 | fail→pass | 10,384 | 4,982 | -52% | 1 | 1 | 0% | 1,896 | 1,518 | -20% | 0 | 0 | — |
case-17 | fail→pass | 12,779 | 5,555 | -57% | 1 | 1 | 0% | 2,199 | 1,565 | -29% | 0 | 0 | — |
case-18 | fail→pass | 6,873 | 1,597 | -77% | 1 | 1 | 0% | 1,002 | 1,014 | +1% | 0 | 0 | — |
case-19 | fail→pass | 4,917 | 1,651 | -66% | 1 | 1 | 0% | 965 | 1,004 | +4% | 0 | 0 | — |
case-20 | fail→pass | 12,755 | 9,794 | -23% | 1 | 1 | 0% | 2,148 | 2,441 | +14% | 0 | 0 | — |
case-21 | fail→pass | 11,566 | 3,116 | -73% | 1 | 1 | 0% | 1,771 | 1,197 | -32% | 0 | 0 | — |
case-22 | fail→pass | 12,360 | 2,268 | -82% | 1 | 1 | 0% | 1,982 | 1,174 | -41% | 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 19 counted toward the lift figure. The other 3 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 +59 percentage points is the difference between those two pass rates over the 19 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.