Math

**Judgment:** This Math skill appears well-designed for broad math assistance and education, with unusually strong attention to pedagogy, user level adaptation, and epistemic safety. It is likely a good default math helper for most users on the platform. **Strengths:** - Explicit, detailed strategies for working with children, students, experts, and teachers. - Encourages step-by-step reasoning, checking arithmetic, and sanity-checking results. - Focus on learning over answer-giving (especially for homework and proofs). - Clear safety norms: acknowledges limits, distinguishes proven vs conjectural results, avoids pretending to solve open problems. **Risks / Limitations:** - Actual mathematical correctness and depth can’t be fully assessed from the description alone; complex or niche topics may still yield errors. - Verification and error-detection are presented as goals; how consistently they’re achieved in practice is uncertain. - Performance characteristics (latency, behavior on very large or symbolic problems) are not evident from the metadata. **Recommended Scenarios:** - General-purpose math help for users from elementary level to advanced university level. - Homework support where maintaining student learning and integrity is important. - Teacher workflows: generating problem sets, scaffolds, and multi-modal explanations. - Expert or research-adjacent discussions that need clear separation between known results, conjectures, and open problems. **Use with Caution For:** - High-stakes, error-intolerant computations (e.g., production engineering calculations) without human verification. - Cutting-edge research areas post-dating the model’s training cutoff, where the skill explicitly may not know the latest results.