Scientific Hypothesis Generation
Structures observations into testable, evidence-bounded hypotheses, rival explanations, and preregistration-ready plans
- Repository
- K-Dense-AI/scientific-agent-skills
- Stars
- 55,780
- Licence
- MIT
- Files
- 15 files
What it does
When installed, the agent helps turn an observation or preliminary finding into a careful research plan. It separates the observation, the research question, candidate hypotheses, rival explanations, and the predictions and measurements that would test them, and it keeps every candidate labeled as a proposal rather than a fact. It also uses bundled templates (search boundary, evidence ledger, preregistration scaffold) and reference notes to record search coverage and avoid causal overclaiming.
Great for
- Researchers or graduate students planning a study from preliminary observations
- Labs preparing a preregistration or analysis plan that needs explicit rival explanations
- Data analysts who want to check whether a pattern supports causal or only associational claims
Skip it if
- Clinical decisions about a specific patient, since the skill explicitly refuses diagnosis, treatment or dosing advice
- Work involving pathogens, toxins or weapons, or sensitive data, where the skill stops at a safety gate and routes to qualified oversight rather than continuing
- Anyone expecting a literature search or automatic ranking of hypotheses; it does not run searches and states it should not score or accept hypotheses
Try asking
- “Turn these preliminary survey results into a set of candidate explanations and rival hypotheses before I design the study.”
- “Help me write a preregistration-ready analysis plan for testing whether this intervention caused the outcome change.”
- “Build a search boundary and evidence ledger documenting what literature I checked for this question.”
- “Review my causal claim and list confounders, selection effects and reverse causation I should rule out.”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 15 original files, the MIT licence and an attribution note.
hypothesis-generation.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip hypothesis-generation.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip hypothesis-generation.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Write down your observation and its measurement details before asking for interpretation, so the skill can freeze it first.
- Ask for rival explanations before choosing tests, and treat the output as a list to challenge rather than conclusions.
- Record the search date and databases yourself; the skill cannot verify literature coverage and says absence in search is not proof of absence.
What's inside
- assets/evidence_ledger_template.csv
- assets/falsification_controls_template.json
- assets/hypothesis_record_template.json
- assets/operationalization_template.json
- assets/prediction_rival_matrix_template.csv
- assets/preregistration_scaffold_template.md
- assets/search_boundary_template.json
- assets/source_ledger.csv
- references/causal_inference_and_claims.md
- references/concepts_and_workflow.md
- references/ethics_safety_and_ai.md
- references/experimental_design_patterns.md
- references/hypothesis_quality_criteria.md
- references/literature_search_strategies.md
- SKILL.md
Redistributed unmodified from K-Dense-AI/scientific-agent-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.