Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use this skill for project-scoped literature review built on Sources/Papers, with synthesis landing in Knowledge, writing handoff in Writing, and the default literature canvas under Maps/literature.canvas.
.claude/skills/galaxy-dawn-obsidian-literature-workflow/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 113% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 9% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -12% | 0% |
This skill owns the project literature workflow.
textSources/Papers/ -> Knowledge/ -> Writing/ -> Maps/literature.canvas
Sources/Papers/{paper-slug}.mdKnowledge/Literature Overview.mdKnowledge/Method Taxonomy.mdKnowledge/Research Gaps.mdKnowledge/Claim Map.mdWriting/related-work-draft.md only after promoted claims pass the evidence gateWriting/comparison-matrix.md only after promoted claims pass the evidence gateMaps/literature.canvasSources/Papers/To-Read, but they cannot support Knowledge or Writing conclusionsMaps/literature.canvasreferences/PAPER-NOTE-SCHEMA.mdreferences/LITERATURE-OVERVIEW.mdreferences/CLAIM-EXTRACTION.mdreferences/METHOD-TAXONOMY.mdreferences/RESEARCH-GAPS.mdreferences/LITERATURE-CANVAS.md| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 6,652 | 7,385 | +11% | 1 | 1 | 0% | 773 | 1,648 | +113% | 0 | 0 | — |
case-02 | fail→fail | 4,169 | 2,199 | -47% | 1 | 1 | 0% | 657 | 746 | +14% | 0 | 0 | — |
case-09 | fail→pass | 19,313 | 5,076 | -74% | 1 | 1 | 0% | 3,087 | 1,292 | -58% | 0 | 0 | — |
case-03 | fail→pass | 13,376 | 11,868 | -11% | 1 | 1 | 0% | 2,427 | 2,643 | +9% | 0 | 0 | — |
case-04 | fail→pass | 14,313 | 7,730 | -46% | 1 | 1 | 0% | 2,476 | 1,829 | -26% | 0 | 0 | — |
case-05 | pass→pass | 14,042 | 8,944 | -36% | 1 | 1 | 0% | 2,071 | 1,866 | -10% | 0 | 0 | — |
case-06 | fail→pass | 14,304 | 9,999 | -30% | 1 | 1 | 0% | 2,238 | 1,961 | -12% | 0 | 0 | — |
case-07 | fail→pass | 13,686 | 4,950 | -64% | 1 | 1 | 0% | 2,123 | 1,200 | -43% | 0 | 0 | — |
case-08 | fail→pass | 12,565 | 18,863 | +50% | 1 | 1 | 0% | 1,945 | 1,369 | -30% | 0 | 0 | — |
case-10 | fail→pass | 11,991 | 3,083 | -74% | 1 | 1 | 0% | 2,119 | 980 | -54% | 0 | 0 | — |
case-11 | fail→pass | 9,705 | 3,810 | -61% | 1 | 1 | 0% | 1,567 | 996 | -36% | 0 | 0 | — |
case-12 | fail→pass | 13,673 | 4,850 | -65% | 1 | 1 | 0% | 2,282 | 1,226 | -46% | 0 | 0 | — |
case-13 | fail→pass | 12,361 | 3,314 | -73% | 1 | 1 | 0% | 1,839 | 1,039 | -44% | 0 | 0 | — |
case-14 | fail→pass | 10,174 | 2,292 | -77% | 1 | 1 | 0% | 1,610 | 762 | -53% | 0 | 0 | — |
case-15 | pass→pass | 13,794 | 11,008 | -20% | 1 | 1 | 0% | 2,261 | 2,020 | -11% | 0 | 0 | — |
case-16 | fail→pass | 3,544 | 3,283 | -7% | 1 | 1 | 0% | 623 | 939 | +51% | 0 | 0 | — |
case-17 | fail→pass | 12,760 | 6,181 | -52% | 1 | 1 | 0% | 2,165 | 1,524 | -30% | 0 | 0 | — |
case-18 | fail→pass | 10,996 | 3,513 | -68% | 1 | 1 | 0% | 1,941 | 1,044 | -46% | 0 | 0 | — |
case-19 | fail→pass | 10,447 | 2,763 | -74% | 1 | 1 | 0% | 1,714 | 853 | -50% | 0 | 0 | — |
case-20 | pass→pass | 5,013 | 5,105 | +2% | 1 | 1 | 0% | 1,034 | 1,462 | +41% | 0 | 0 | — |
case-21 | pass→pass | 8,417 | 4,643 | -45% | 1 | 1 | 0% | 1,596 | 1,113 | -30% | 0 | 0 | — |
case-22 | pass→pass | 11,336 | 9,627 | -15% | 1 | 1 | 0% | 1,923 | 2,151 | +12% | 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 +73 percentage points is the difference between those two pass rates over the 22 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.