hypothesis-generation
Formulates evidence-bounded scientific questions, candidate hypotheses
by k-dense-aiFree
Description
hypothesis-generation is an agent skill that turns an observation into a transparent set of candidate explanations and tests. It formulates evidence-bounded scientific questions, candidate hypotheses, rival explanations, claims that are causal or only associational, and predictions that could tell rival hypotheses apart. It treats a hypothesis as a proposal to be challenged, not a finding, a fact, a diagnosis or a recommendation. The skill has firm boundaries for sensitive material. Before you use unpublished, sensitive, controlled, personal, proprietary, export-controlled or security-relevant material, it asks you to confirm that you are authorised and that the relevant institutional, funder, publisher, data-use, privacy and AI policies allow it. It asks you to keep such material local unless an authorised person approves a named external destination and data scope, to minimise what you send, and to stop at any human, animal, biosafety, dual-use, data-governance or regulatory gate. It must not place sensitive data in web searches or external AI systems without that authorisation.…
Highlights
- Candidate hypotheses and rival explanations
- Causal versus associational claims
- Discriminating predictions to test
- Treats hypotheses as proposals
- Strict rules for sensitive material
Ratings & Reviews
No ratings yet
Be the first to rate hypothesis-generation.
People also visit
discernment-nudge
After you give a substantive answer or draft that the user may act on
hypothesis-generation FAQ
What does the skill produce?
A set of evidence-bounded questions, candidate hypotheses, rival explanations and predictions that could discriminate between them.
Is a hypothesis a finding?
No. The skill treats a hypothesis as a proposal to be challenged, not a finding, fact, diagnosis or recommendation.
What about sensitive data?
The skill asks you to confirm authorisation and the relevant policies, keep material local unless approved, and minimise what is shared.
When does it stop?
At human, animal, biosafety, dual-use, data-governance or regulatory gates.
How do I install it?
Run npx skills add https://github.com/k-dense-ai/scientific-agent-skills --skill hypothesis-generation.
Is it checked for security?
Its security audits show passes from Gen Agent Trust Hub, Socket and Snyk.
What are the best hypothesis-generation alternatives?
Popular alternatives to hypothesis-generation include council, gget and discernment-nudge. See all hypothesis-generation alternatives