Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Repo-wide methods orchestration workflow for the Research Project Template. USE WHEN the user asks to add, audit, improve, or validate methods, methodology, method contracts, stage-to-method wiring, artifact/evidence provenance, or orchestration across template projects.
.claude/skills/docxology-template-methods-orchestration/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -74% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -63% | 0% |
infrastructure.methods for one project or --all-public.pipeline.yaml stagecontracts, artifact manifest, evidence registry, and validation commands.
src/, thin scripts, manuscript source, orstage contracts. Do not edit generated output/ as a fix.
then rerun methods plan with --check.
changed.
bashuv run python -m infrastructure.methods plan --project <project> --format markdown uv run python -m infrastructure.methods plan --project <project> --format json --check uv run python -m infrastructure.methods plan --all-public --artifact-mode source --format json uv run python -m infrastructure.methods plan --all-public --artifact-mode rendered --format json uv run python scripts/runner/execute_pipeline.py --project <project> --core-only uv run python -m infrastructure.validation.cli prerender projects/<project>/manuscript --repo-root . uv run pytest tests/infra_tests/methods -q
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 4,778 | 5,224 | +9% | 1 | 1 | 0% | 247 | 727 | +194% | 0 | 0 | — |
case-02 | fail→fail | 3,300 | 4,944 | +50% | 1 | 1 | 0% | 281 | 710 | +153% | 0 | 0 | — |
case-03 | fail→fail | 4,420 | 5,594 | +27% | 1 | 1 | 0% | 134 | 671 | +401% | 0 | 0 | — |
case-04 | pass→pass | 8,609 | 4,784 | -44% | 1 | 1 | 0% | 1,483 | 1,374 | -7% | 0 | 0 | — |
case-05 | fail→pass | 15,564 | 1,698 | -89% | 1 | 1 | 0% | 2,433 | 723 | -70% | 0 | 0 | — |
case-11 | fail→pass | 11,888 | 4,844 | -59% | 1 | 1 | 0% | 2,181 | 1,368 | -37% | 0 | 0 | — |
case-06 | fail→pass | 17,268 | 1,827 | -89% | 1 | 1 | 0% | 2,941 | 777 | -74% | 0 | 0 | — |
case-07 | fail→pass | 13,507 | 2,442 | -82% | 1 | 1 | 0% | 2,355 | 873 | -63% | 0 | 0 | — |
case-08 | fail→pass | 13,539 | 2,448 | -82% | 1 | 1 | 0% | 2,444 | 908 | -63% | 0 | 0 | — |
case-09 | fail→pass | 12,376 | 2,528 | -80% | 1 | 1 | 0% | 2,239 | 879 | -61% | 0 | 0 | — |
case-10 | fail→pass | 5,130 | 1,869 | -64% | 1 | 1 | 0% | 860 | 813 | -5% | 0 | 0 | — |
case-12 | fail→fail | 6,158 | 1,510 | -75% | 1 | 1 | 0% | 1,139 | 690 | -39% | 0 | 0 | — |
case-13 | fail→pass | 13,673 | 4,446 | -67% | 1 | 1 | 0% | 2,245 | 1,120 | -50% | 0 | 0 | — |
case-14 | fail→pass | 27,382 | 1,505 | -95% | 1 | 1 | 0% | 3,054 | 598 | -80% | 0 | 0 | — |
case-15 | pass→pass | 14,783 | 3,593 | -76% | 1 | 1 | 0% | 2,015 | 927 | -54% | 0 | 0 | — |
case-16 | fail→pass | 9,498 | 2,208 | -77% | 1 | 1 | 0% | 1,443 | 819 | -43% | 0 | 0 | — |
case-17 | pass→pass | 10,738 | 4,084 | -62% | 1 | 1 | 0% | 1,578 | 1,033 | -35% | 0 | 0 | — |
case-18 | fail→pass | 6,548 | 2,111 | -68% | 1 | 1 | 0% | 866 | 786 | -9% | 0 | 0 | — |
case-19 | pass→pass | 8,960 | 5,978 | -33% | 1 | 1 | 0% | 1,506 | 1,443 | -4% | 0 | 0 | — |
case-20 | pass→pass | 12,240 | 5,184 | -58% | 1 | 1 | 0% | 2,093 | 1,319 | -37% | 0 | 0 | — |
case-21 | pass→pass | 24,140 | 7,642 | -68% | 1 | 1 | 0% | 1,800 | 1,751 | -3% | 0 | 0 | — |
case-22 | fail→pass | 17,521 | 7,851 | -55% | 1 | 1 | 0% | 3,002 | 1,812 | -40% | 0 | 0 | — |
case-23 | fail→pass | 9,203 | 3,143 | -66% | 1 | 1 | 0% | 1,577 | 907 | -42% | 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 20 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 +57 percentage points is the difference between those two pass rates over the 20 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.