hypogenic

SUMMARY

What it does

This Agent Skill provides procedural guidance for planning and auditing the use of the ChicagoHAI HypoGeniC/HypoRefine software for LLM-assisted hypothesis generation from labeled text datasets. It covers the `hypogenic` package, its task configs, hypothesis banks, and HypoBench datasets. The skill emphasizes scientific boundaries, safe configuration, dataset safety, cost planning, and local evaluation without model calls. It includes bundled local audit tools and references for configuration, upstream details, datasets, evaluation, security, and sources. The skill is part of Scientific Agent Skills by K-Dense and requires Python 3.10+, uv, and specific dependencies for local tools. Actual HypoGeniC runs may require separate LLM provider credentials, Redis, local model resources, and network access.