What problem does it solve? When someone cites a paper, benchmark, or repo README to support a stronger claim than the source actually makes, it is hard to respond precisely without straw-manning or vague disagreement. This Skill isolates what the evidence actually supports and drafts concise replies that correct the overreach. ## Core Features & Use Cases - Claim-vs-source comparison: Extracts the specific proposition being asserted and splits findings into supported, not established, and unproven extrapolations. - Evidence ladder labeling: Ranks claims from direct empirical findings down to repo philosophy, so the inferential gap is explicit. - Forum-ready reply drafting: Produces Reddit/X/forum responses that lead with the misunderstanding, state what the source shows and does not show, and tie back to the concrete artifact under debate. - Use Case: Someone argues a persona-benchmarking paper proves long narrative persona files are better prompt engineering. Use this Skill to show the paper supports persona sensitivity but not verbosity, then draft a postable correction. ## Quick Start Compare this person's claim about the cited paper to what the paper actually shows and draft a reply that addresses the misunderstanding of the research.
