Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Literature meta-analysis exemplar — multi-engine retrieval, dedup, full-text, bibliometrics, embeddings, optional knowledge graph. Default term: modafinil.
.claude/skills/docxology-template-literature-meta-analysis/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -65% | 0% |
Project-scoped skill for the in-repo exemplar at projects/templates/template_literature_meta_analysis/. Load this when working inside the project.
template_literature_meta_analysis exemplar — running scripts, editing source,or regenerating outputs.
no-mocks testing) still hold after changes.
bash# From the repository root # Canonical fresh-clone gate: provisions the project's isolated dependencies. uv run python scripts/pipeline/stage_01_test.py --project templates/template_literature_meta_analysis --project-only uv run python scripts/pipeline/stage_02_analysis.py --project templates/template_literature_meta_analysis uv run python scripts/pipeline/stage_03_render.py --project templates/template_literature_meta_analysis uv run python scripts/pipeline/stage_04_validate.py --project templates/template_literature_meta_analysis uv run python scripts/pipeline/stage_05_copy.py --project templates/template_literature_meta_analysis
src/ or sharedinfrastructure/, not in scripts/.
computation.
pytest invocation does not provision this exemplar's scientific dependencies.
output/ — regenerate fromsource and config.
as working directory unless the child AGENTS.md states otherwise.
Use MCP connectors and configured retrieval keys first. For uncovered endpoints, Monid (infrastructure/search/monid/) is infrastructure-only — pricing table: PRICING.md. Offline deep_research replay (scripts/08_deep_research_dispatch.py) remains the exemplar's default long-form synthesis path.
AGENTS.mdREADME.mdTODO.mdprojects/AGENTS.md| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 37,281 | 19,127 | -49% | 1 | 1 | 0% | 3,762 | 1,009 | -73% | 0 | 0 | — |
case-01 | fail→fail | 10,363 | 16,955 | +64% | 1 | 1 | 0% | 780 | 1,068 | +37% | 0 | 0 | — |
case-03 | fail→fail | 19,946 | 18,031 | -10% | 1 | 1 | 0% | 290 | 1,123 | +287% | 0 | 0 | — |
case-04 | fail→pass | 16,850 | 11,389 | -32% | 1 | 1 | 0% | 1,784 | 1,435 | -20% | 0 | 0 | — |
case-05 | pass→pass | 19,745 | 10,168 | -49% | 1 | 1 | 0% | 2,138 | 1,468 | -31% | 0 | 0 | — |
case-06 | fail→pass | 22,779 | 15,183 | -33% | 1 | 1 | 0% | 3,328 | 2,211 | -34% | 0 | 0 | — |
case-07 | pass→pass | 13,665 | 10,416 | -24% | 1 | 1 | 0% | 1,158 | 1,388 | +20% | 0 | 0 | — |
case-08 | fail→pass | 15,472 | 11,806 | -24% | 1 | 1 | 0% | 1,744 | 1,615 | -7% | 0 | 0 | — |
case-09 | fail→pass | 23,864 | 9,784 | -59% | 1 | 1 | 0% | 3,692 | 1,546 | -58% | 0 | 0 | — |
case-10 | fail→pass | 23,600 | 8,085 | -66% | 1 | 1 | 0% | 3,084 | 1,079 | -65% | 0 | 0 | — |
case-11 | fail→pass | 21,644 | 12,033 | -44% | 1 | 1 | 0% | 2,181 | 1,607 | -26% | 0 | 0 | — |
case-12 | fail→pass | 17,035 | 3,325 | -80% | 1 | 1 | 0% | 2,249 | 1,062 | -53% | 0 | 0 | — |
case-13 | fail→pass | 16,818 | 8,895 | -47% | 1 | 1 | 0% | 1,561 | 1,148 | -26% | 0 | 0 | — |
case-14 | fail→pass | 18,141 | 2,696 | -85% | 1 | 1 | 0% | 2,218 | 1,006 | -55% | 0 | 0 | — |
case-15 | fail→pass | 12,618 | 3,444 | -73% | 1 | 1 | 0% | 1,734 | 1,089 | -37% | 0 | 0 | — |
case-16 | fail→pass | 18,268 | 9,296 | -49% | 1 | 1 | 0% | 2,456 | 2,291 | -7% | 0 | 0 | — |
case-17 | fail→pass | 16,847 | 6,655 | -60% | 1 | 1 | 0% | 2,083 | 1,573 | -24% | 0 | 0 | — |
case-18 | fail→pass | 22,575 | 8,850 | -61% | 1 | 1 | 0% | 2,602 | 1,156 | -56% | 0 | 0 | — |
case-19 | fail→pass | 20,667 | 8,454 | -59% | 1 | 1 | 0% | 2,713 | 1,068 | -61% | 0 | 0 | — |
case-20 | pass→pass | 11,877 | 14,651 | +23% | 1 | 1 | 0% | 2,131 | 2,418 | +13% | 0 | 0 | — |
case-21 | pass→pass | 15,541 | 19,052 | +23% | 1 | 1 | 0% | 2,243 | 3,283 | +46% | 0 | 0 | — |
case-22 | pass→pass | 20,847 | 22,124 | +6% | 1 | 1 | 0% | 3,341 | 4,133 | +24% | 0 | 0 | — |
case-23 | fail→pass | 18,950 | 7,647 | -60% | 1 | 1 | 0% | 1,968 | 1,059 | -46% | 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 +65 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 | +55% |
Other measured skills in the registry, with their headline benchmark lift.