Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Numbered pipeline stage orchestrators (stages 00–12) for the template research framework. Each stage_NN_*.py is a thin orchestrator that wires the repo root onto sys.path and delegates to infrastructure/ modules.
.claude/skills/docxology-template-pipeline/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -35% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -50% | 0% |
Numbered pipeline stage scripts in scripts/pipeline/.
Load this skill when you need to:
| Script | Infrastructure module | |--------|-----------------------| | stage_00_setup.py | infrastructure.core.runtime.setup_checks | | stage_01_test.py | infrastructure.core.pytest_orchestration, infrastructure.core.test_runner, infrastructure.reporting.pipeline_test_runner | | stage_02_analysis.py | infrastructure.core.analysis_pipeline, infrastructure.core.script_discovery | | stage_03_render.py | infrastructure.rendering.pipeline (all enabled formats; historical YAML label PDF Rendering) | | stage_04_validate.py | infrastructure.validation.output.pipeline, infrastructure.validation.publication.rendered_provenance | | stage_05_copy.py | infrastructure.core.files, infrastructure.validation.output.render_formats, infrastructure.validation.output.validator, infrastructure.reporting.output_statistics | | stage_06_llm_review.py | infrastructure.llm.review.pipeline_runner | | stage_07_executive_report.py | infrastructure.reporting.multi_project_reporter, infrastructure.reporting.output_organizer | | stage_08_connector_search.py | infrastructure.search.connectors (opt-in) | | stage_09_provenance_record.py | infrastructure.provenance (opt-in) | | stage_10_research_workflow.py | infrastructure.research.ResearchWorkflow (opt-in) | | stage_11_ebook.py | infrastructure.rendering.ebook_stage | | stage_12_metadata.py | infrastructure.publishing.metadata.metadata_stage |
pythonsys.path.insert(0, str(Path(__file__).resolve().parents[2])) from scripts import ensure_repo_root_on_path ensure_repo_root_on_path()
parents[2] (not parent.parent) — scripts live at depth 3.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-09 | fail→pass | 15,108 | 1,846 | -88% | 1 | 1 | 0% | 2,456 | 906 | -63% | 0 | 0 | — |
case-03 | fail→pass | 12,283 | 4,219 | -66% | 1 | 1 | 0% | 2,048 | 1,322 | -35% | 0 | 0 | — |
case-08 | fail→pass | 12,043 | 2,256 | -81% | 1 | 1 | 0% | 1,966 | 1,002 | -49% | 0 | 0 | — |
case-01 | fail→pass | 43,719 | 36,935 | -16% | 1 | 1 | 0% | 2,382 | 1,936 | -19% | 0 | 0 | — |
case-02 | fail→pass | 17,364 | 4,270 | -75% | 1 | 1 | 0% | 2,773 | 1,390 | -50% | 0 | 0 | — |
case-04 | fail→pass | 9,354 | 22,191 | +137% | 1 | 1 | 0% | 1,519 | 1,336 | -12% | 0 | 0 | — |
case-05 | fail→pass | 12,327 | 5,772 | -53% | 1 | 1 | 0% | 1,846 | 1,615 | -13% | 0 | 0 | — |
case-06 | fail→pass | 19,079 | 7,392 | -61% | 1 | 1 | 0% | 3,465 | 1,895 | -45% | 0 | 0 | — |
case-07 | fail→pass | 9,378 | 1,342 | -86% | 1 | 1 | 0% | 1,434 | 834 | -42% | 0 | 0 | — |
case-10 | fail→pass | 13,338 | 1,293 | -90% | 1 | 1 | 0% | 2,100 | 790 | -62% | 0 | 0 | — |
case-11 | fail→pass | 11,304 | 1,766 | -84% | 1 | 1 | 0% | 1,706 | 870 | -49% | 0 | 0 | — |
case-12 | fail→pass | 8,866 | 2,124 | -76% | 1 | 1 | 0% | 1,376 | 953 | -31% | 0 | 0 | — |
case-13 | fail→pass | 11,420 | 2,330 | -80% | 1 | 1 | 0% | 1,796 | 1,009 | -44% | 0 | 0 | — |
case-14 | fail→pass | 9,896 | 1,243 | -87% | 1 | 1 | 0% | 1,480 | 795 | -46% | 0 | 0 | — |
case-15 | fail→pass | 9,627 | 2,594 | -73% | 1 | 1 | 0% | 1,384 | 1,026 | -26% | 0 | 0 | — |
case-16 | pass→pass | 11,800 | 1,312 | -89% | 1 | 1 | 0% | 1,743 | 786 | -55% | 0 | 0 | — |
case-17 | fail→pass | 14,164 | 1,950 | -86% | 1 | 1 | 0% | 2,041 | 888 | -56% | 0 | 0 | — |
case-18 | fail→pass | 11,856 | 1,316 | -89% | 1 | 1 | 0% | 1,770 | 779 | -56% | 0 | 0 | — |
case-19 | fail→pass | 9,519 | 1,133 | -88% | 1 | 1 | 0% | 1,343 | 799 | -41% | 0 | 0 | — |
case-20 | pass→pass | 8,616 | 5,179 | -40% | 1 | 1 | 0% | 1,463 | 1,482 | +1% | 0 | 0 | — |
case-21 | fail→pass | 13,344 | 2,359 | -82% | 1 | 1 | 0% | 1,820 | 1,024 | -44% | 0 | 0 | — |
case-22 | pass→pass | 10,231 | 4,173 | -59% | 1 | 1 | 0% | 1,617 | 1,269 | -22% | 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 +86 percentage points is the difference between those two pass rates over the 22 comparable cases.
The publisher has shipped newer versions since this run, so these numbers describe v2, not the version currently listed.
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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/4/2026 | +59% |
Other measured skills in the registry, with their headline benchmark lift.