Generate and test research hypotheses from tabular datasets and linked literature.
npx skills add https://github.com/dralkh/seerai --skill hypogenic-dralkh
Please help me install this Agent Skill. Skill: hypogenic Source: https://github.com/dralkh/seerai/tree/main/skills/hypogenic Command: npx skills add https://github.com/dralkh/seerai --skill hypogenic-dralkh
It helps researchers move from raw observations and papers to testable hypotheses, reducing the time and guesswork involved in exploring patterns, comparing competing explanations, and validating ideas.
Use the hypogenic skill to configure your dataset paths and prompt templates, then generate and test a set of research hypotheses for your target task.
High-intent search queries and answers about installing and using this skill.
To generate testable hypotheses from tabular data, you configure task-specific JSON datasets and YAML prompt templates to discover candidate explanations from empirical patterns. This data-driven discovery method supports classification-style studies like deception or AI-generated text detection.
Literature-informed hypothesis refinement combines empirical evidence with linked literature to improve hypothesis quality. It uses union methods that ground data-driven discovery in paper-based theory to validate competing explanations for observational datasets.
Yes, systematic review workflows support mental health indicators and deception detection research. The method uses custom label extraction and repeated inference to systematically compare competing explanations across observational datasets.
PDF preprocessing with GROBID is optional but recommended for scalable literature-grounded workflows. It parses linked papers to extract text for literature-informed refinement, with Redis caching available to support large-scale repeated inference runs.
A key limitation of LLM-based hypothesis generation is the strict requirement for task-specific JSON datasets and YAML prompt templates. It is designed for observational data classification rather than controlled experiments, and scalable runs require additional Redis caching setup.
gil0mendesConverts Markdown product requirement documents into executable JSON for AI agents.
braselogGenerate 3-5 testable mechanistic hypotheses and initial experiment designs.
RamanEbrahimiGenerate and test hypotheses about empirical data patterns with literature insights.
zm2231Generate PRISMA-compliant systematic literature reviews with thematic analyses and citation networks.
