Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Registered-report exemplar for preregistration locking, deviation ledgers, replication, and confirmatory claim boundaries.
.claude/skills/docxology-template-registered-report/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -71% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -65% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -48% | 0% |
Load this skill when working inside projects/templates/template_registered_report/.
bashuv run pytest projects/templates/template_registered_report/tests --cov=projects/templates/template_registered_report/src --cov-fail-under=90 uv run python scripts/pipeline/stage_01_test.py --project templates/template_registered_report --project-only
scripts/generate_figures.py; it writesoutput/figures/figure_registry.json only after all four registered-report figures exist, and supports explicit temp destinations for isolated checks.
scripts/generate_review_artifacts.py; thesuite binds the committed output/reports/*.json artifacts and the manuscript's sensitivity row and deviation ledger to a fresh regeneration, so never hand-edit those files or the bound prose.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 2,681 | 6,594 | +146% | 1 | 1 | 0% | 259 | 650 | +151% | 0 | 0 | — |
case-02 | fail→fail | 14,064 | 6,201 | -56% | 1 | 1 | 0% | 2,151 | 593 | -72% | 0 | 0 | — |
case-03 | fail→fail | 5,202 | 7,670 | +47% | 1 | 1 | 0% | 269 | 674 | +151% | 0 | 0 | — |
case-04 | pass→pass | 9,432 | 2,679 | -72% | 1 | 1 | 0% | 1,540 | 597 | -61% | 0 | 0 | — |
case-05 | fail→pass | 13,926 | 2,059 | -85% | 1 | 1 | 0% | 2,157 | 617 | -71% | 0 | 0 | — |
case-06 | fail→fail | 4,776 | 3,960 | -17% | 1 | 1 | 0% | 668 | 882 | +32% | 0 | 0 | — |
case-07 | fail→pass | 9,800 | 2,250 | -77% | 1 | 1 | 0% | 1,509 | 624 | -59% | 0 | 0 | — |
case-08 | fail→fail | 7,360 | 3,018 | -59% | 1 | 1 | 0% | 1,125 | 647 | -42% | 0 | 0 | — |
case-09 | fail→pass | 10,477 | 4,957 | -53% | 1 | 1 | 0% | 1,446 | 1,067 | -26% | 0 | 0 | — |
case-10 | pass→pass | 12,136 | 9,604 | -21% | 1 | 1 | 0% | 1,881 | 1,721 | -9% | 0 | 0 | — |
case-11 | pass→pass | 10,943 | 4,969 | -55% | 1 | 1 | 0% | 1,626 | 952 | -41% | 0 | 0 | — |
case-12 | fail→pass | 13,000 | 2,724 | -79% | 1 | 1 | 0% | 1,910 | 665 | -65% | 0 | 0 | — |
case-13 | pass→pass | 9,090 | 5,345 | -41% | 1 | 1 | 0% | 1,486 | 1,143 | -23% | 0 | 0 | — |
case-14 | fail→pass | 11,407 | 4,084 | -64% | 1 | 1 | 0% | 1,714 | 892 | -48% | 0 | 0 | — |
case-15 | fail→pass | 4,564 | 1,586 | -65% | 1 | 1 | 0% | 733 | 517 | -29% | 0 | 0 | — |
case-16 | fail→pass | 6,579 | 2,115 | -68% | 1 | 1 | 0% | 1,066 | 481 | -55% | 0 | 0 | — |
case-17 | fail→pass | 16,658 | 11,738 | -30% | 1 | 1 | 0% | 2,461 | 1,281 | -48% | 0 | 0 | — |
case-18 | pass→pass | 8,886 | 2,716 | -69% | 1 | 1 | 0% | 1,471 | 645 | -56% | 0 | 0 | — |
case-19 | pass→pass | 10,656 | 3,780 | -65% | 1 | 1 | 0% | 1,779 | 773 | -57% | 0 | 0 | — |
case-20 | pass→pass | 15,061 | 11,332 | -25% | 1 | 1 | 0% | 2,676 | 2,173 | -19% | 0 | 0 | — |
case-21 | pass→pass | 7,077 | 7,015 | -1% | 1 | 1 | 0% | 1,233 | 1,498 | +21% | 0 | 0 | — |
case-22 | pass→pass | 7,484 | 4,299 | -43% | 1 | 1 | 0% | 1,254 | 947 | -24% | 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 20 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 +36 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/4/2026 | +50% |
Other measured skills in the registry, with their headline benchmark lift.