shushu-novelty-finder

computer science research lineage mapping, novelty discovery, and idea stress-testing for codex. use when the user wants to map extremely similar papers in a research direction, extract each paper's innovation points, generate paper innovation ideas, analyze whether an idea is reasonable, find CS research gaps, analyze a research direction or seed paper, review recent or ten-year literature trends, rank weak/medium/strong novelty ideas, or judge whether an idea is paper-ready. the skill supports two primary outputs: literature-lineage mode for detailed paper-context mapping, and idea-generation mode for concrete novelty candidates. it must ground claims in papers and ask clarifying questions when scope is underspecified instead of blindly searching broadly.