This is a comprehensive guide to Claude's math computation stack, not a tool itself. It maps five layers of mathematical capability: SymPy for symbolic algebra, Z3 for constraint solving and proofs, a scratchpad for verifying reasoning chains, a tutor mode with progressive hints (1-5 levels from conceptual to full solution), and Lean 4 for formal proofs. The quick reference table is genuinely useful for picking between sympy_compute.py, numpy_compute.py, z3_solve.py, and the rest. Think of it as internal documentation that helps you choose the right hammer when you need exact algebraic solutions versus numerical optimization versus theorem proving. The workflows section shows common patterns like solve-then-verify. It's meta-tooling that prevents you from reaching for scipy when you actually need symbolic computation.
npx skills add https://github.com/parcadei/continuous-claude-v3 --skill math-help