What problem does it solve? Researchers risk investing weeks into ideas that are not novel, would fail peer review, or are strategically unsound. This Skill screens candidate research ideas across three evaluation modules and produces ranked, scored recommendations before any execution effort is committed. ## Core Features & Use Cases - Novelty Assessment: Verifies each idea's core claims against recent literature via multi-source web search, cross-referencing with prior literature survey data, and cross-model verification through an external LLM. - Venue Reviewer Simulation: Simulates a 3-reviewer plus meta-review committee calibrated to a target venue (ICML, VLDB, NeurIPS, or all) to predict whether the executed idea would be accepted. - Strategic Fit Assessment: Scores each idea on five dimensions (longevity, roadmap viability, application grounding, execution uniqueness, iteration readiness) and computes a weighted composite score with PROCEED, PROCEED WITH CAUTION, or ABANDON recommendations. - Use Case: After generating candidate ideas with an idea-generation step, run this Skill to rank them by composite score and receive a detailed screening report plus a ranked summary ready for downstream refinement. ## Quick Start Screen the research ideas in outputs/IDEAS_FILTERED.md against the ICML venue and produce a ranked screening report.
