gaia-formalize-coarse

To convert an academic paper into a probabilistic reasoning package, the skill applies a four-phase extraction workflow to reconstruct reasoning chains and emits a standalone Gaia DSL directory that compiles and runs Bayesian belief updates.

Formalizing scientific claims from a single document involves extracting a cross-conclusion logic graph, auditing weak points, calibrating leaf priors, and generating a reviewable Gaia knowledge package for inference.

You cannot use this to formalize a review or survey paper. A suitability gate detects perspectives, reviews, or corrupted text and writes a concise skip report instead of inventing contributions.

Yes, you need python-dsl and gaia-cli. They are required dependencies for the skill to emit staged Gaia DSL artifacts and for the caller to compile and run the resulting probabilistic belief updates.

LKM cross-grounding audits claims by reverse-provenance checking the paper's extracted arguments against LKM's existing knowledge graph. This process runs best-effort and non-blocking to validate alignment.