Install any skill in seconds. Free to start, no credit card required.
Get Started Free →This skill should be used when the user asks to "brainstorm research ideas", "use 5W1H framework", "identify research gaps", "conduct gap analysis", "start research project", "conduct literature review", "define research question", "select research method", "plan research", or mentions research project initiation phase. Provides comprehensive guidance for research startup workflow from idea generation to planning.
.claude/skills/galaxy-dawn-research-ideation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 29% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 21% | 0% |
Supports the complete workflow for the research project initiation phase, from vague idea to a traceable research contract: research question, hypothesis, evidence needs, falsification criteria, method selection, and next action.
Systematically brainstorm research ideas using the 5W1H framework:
Integration with superpowers:brainstorming: Can invoke the superpowers:brainstorming skill for interactive brainstorming to help rapidly generate and evaluate research ideas.
Systematically search, analyze, and synthesize related literature:
Systematically identify and evaluate research gaps:
Analysis Dimensions:
Formulate specific research questions based on literature analysis:
Research Question Card that states the hypothesis, current evidence, missing evidence, support criteria, falsification criteria, and minimal next actionSelect appropriate research methods:
Develop detailed research plans:
Use the research-ideation skill in the following situations:
Research interest → Idea brainstorming (5W1H) → Research Question Card → Evidence Records → Gap analysis → Proposal readiness gate → Select method or experiment → DecisionConditional Outputs:
research-question-card.md - Always generate this for research startup or intake.literature-review.md - Generate only when evidence-labeled sources are sufficient for synthesis.research-proposal.md - Generate only when a selected Research Question Card passes the evidence gate.references.bib - Generate only when reliable citation metadata exists.research-ideation (Research initiation)
↓
Experiment execution (completed by user)
↓
results-analysis (Results analysis)
↓
ml-paper-writing (Paper writing)Through the Zotero MCP server, the research-ideation workflow automates literature management:
zotero_add_items_by_identifier to prefer paper/preprint imports before any webpage fallbackzotero_create_collection creates topic-based collections with standard sub-collections (Core Papers, Methods, Applications, Baselines, To-Read)zotero_add_items_by_identifier(..., attach_pdf=true) runs a PDF cascade (landing-page PDF hints → direct PDF → Unpaywall), and zotero_find_and_attach_pdfs can be used as a follow-up sweep for remaining itemszotero_get_item_fulltext reads indexed PDF content for analysis and note-takingzotero_search_items and zotero_get_collection_items browse existing papers to avoid duplicatesResearch-{topic}-{YYYY} formatDetailed methodology guides, loaded on demand:
references/5w1h-framework.md - 5W1H Framework Guidereferences/literature-search-strategies.md - Literature Search Strategiesreferences/zotero-integration-guide.md - Zotero MCP Integration Guidereferences/gap-analysis-guide.md - Gap Analysis Guidereferences/research-question-formulation.md - Research Question Formulationreferences/research-contract.md - Research Contractreferences/method-selection-guide.md - Method Selection Guidereferences/research-planning.md - Research PlanningComplete working examples:
examples/example-literature-review.md - Literature Review Exampleexamples/example-research-proposal.md - Research Proposal Example| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 32,511 | 6,333 | -81% | 1 | 1 | 0% | 5,420 | 2,342 | -57% | 0 | 0 | — |
case-01 | fail→fail | 25,184 | 10,747 | -57% | 1 | 1 | 0% | 4,032 | 2,504 | -38% | 0 | 0 | — |
case-02 | fail→fail | 36,482 | 10,885 | -70% | 1 | 1 | 0% | 6,218 | 2,959 | -52% | 0 | 0 | — |
case-04 | fail→pass | 18,145 | 14,633 | -19% | 1 | 1 | 0% | 3,445 | 4,274 | +24% | 0 | 0 | — |
case-05 | fail→pass | 31,530 | 12,634 | -60% | 1 | 1 | 0% | 5,173 | 3,951 | -24% | 0 | 0 | — |
case-06 | pass→pass | 16,204 | 21,874 | +35% | 1 | 1 | 0% | 2,148 | 5,888 | +174% | 0 | 0 | — |
case-07 | fail→fail | 15,167 | 4,760 | -69% | 1 | 1 | 0% | 807 | 2,801 | +247% | 0 | 0 | — |
case-08 | fail→pass | 15,817 | 8,519 | -46% | 1 | 1 | 0% | 2,688 | 3,457 | +29% | 0 | 0 | — |
case-09 | pass→pass | 12,494 | 7,882 | -37% | 1 | 1 | 0% | 2,002 | 3,139 | +57% | 0 | 0 | — |
case-10 | fail→pass | 18,141 | 11,184 | -38% | 1 | 1 | 0% | 2,897 | 3,898 | +35% | 0 | 0 | — |
case-11 | pass→pass | 53,990 | 13,538 | -75% | 1 | 1 | 0% | 2,789 | 4,468 | +60% | 0 | 0 | — |
case-12 | fail→pass | 14,772 | 6,391 | -57% | 1 | 1 | 0% | 2,473 | 2,981 | +21% | 0 | 0 | — |
case-13 | fail→pass | 19,591 | 13,433 | -31% | 1 | 1 | 0% | 3,291 | 4,075 | +24% | 0 | 0 | — |
case-14 | fail→pass | 12,012 | 8,140 | -32% | 1 | 1 | 0% | 1,971 | 3,258 | +65% | 0 | 0 | — |
case-15 | fail→pass | 13,179 | 9,169 | -30% | 1 | 1 | 0% | 2,281 | 3,453 | +51% | 0 | 0 | — |
case-16 | pass→pass | 10,195 | 5,711 | -44% | 1 | 1 | 0% | 1,813 | 3,039 | +68% | 0 | 0 | — |
case-17 | pass→pass | 13,103 | 13,067 | -0% | 1 | 1 | 0% | 2,116 | 4,311 | +104% | 0 | 0 | — |
case-18 | pass→pass | 11,996 | 7,455 | -38% | 1 | 1 | 0% | 2,034 | 3,420 | +68% | 0 | 0 | — |
case-19 | fail→fail | 11,682 | 2,604 | -78% | 1 | 1 | 0% | 1,925 | 2,369 | +23% | 0 | 0 | — |
case-20 | fail→fail | 10,171 | 15,220 | +50% | 1 | 1 | 0% | 2,228 | 5,209 | +134% | 0 | 0 | — |
case-21 | fail→fail | 50,262 | 28,743 | -43% | 1 | 1 | 0% | 4,686 | 8,145 | +74% | 0 | 0 | — |
case-22 | pass→pass | 21,741 | 26,883 | +24% | 1 | 1 | 0% | 3,674 | 7,428 | +102% | 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 19 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 +36 percentage points is the difference between those two pass rates over the 19 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.