Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Unified math capabilities - computation, solving, and explanation. I route to the right tool.
| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 1 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 191% | 0% |
| case-06 | ✓→✗ | ▼ Worse | 65% | 0% |
One entry point for all computation and explanation. I route to the right tool based on your request.
For formal proofs, use /prove instead.
| You Say | I Use | |---------|-------| | "Solve x² - 4 = 0" | SymPy solve | | "Integrate sin(x) from 0 to π" | SymPy integrate | | "Eigenvalues of 1,2],3,4]]" | SymPy eigenvalues | | "Is x² + 1 > 0 for all x?" | Z3 prove | | "Convert 5 miles to km" | Pint | | "Explain what a functor is" | Category theory skill |
bashuv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/sympy_compute.py" <command> <args>
| Command | Description | Example | |---------|-------------|---------| | solve | Solve equations | solve "x**2 - 4" --var x | | integrate | Definite/indefinite integral | integrate "sin(x)" --var x --lower 0 --upper pi | | diff | Derivative | diff "x**3" --var x | | simplify | Simplify expression | simplify "sin(x)**2 + cos(x)**2" | | limit | Compute limit | limit "sin(x)/x" --var x --point 0 | | series | Taylor expansion | series "exp(x)" --var x --point 0 --n 5 | | dsolve | Solve ODE | dsolve "f''(x) + f(x)" --func f --var x | | laplace | Laplace transform | laplace "sin(t)" --var t |
Matrix Operations: | Command | Description | |---------|-------------| | det | Determinant | | eigenvalues | Eigenvalues | | eigenvectors | Eigenvectors with multiplicities | | inverse | Matrix inverse | | transpose | Transpose | | rref | Row echelon form | | rank | Matrix rank | | nullspace | Null space basis | | linsolve | Linear system Ax=b | | charpoly | Characteristic polynomial |
Number Theory: | Command | Description | |---------|-------------| | factor | Factor polynomial | | factorint | Prime factorization | | isprime | Primality test | | gcd | Greatest common divisor | | lcm | Least common multiple | | modinverse | Modular inverse |
Combinatorics: | Command | Description | |---------|-------------| | binomial | C(n,k) | | factorial | n! | | permutation | P(n,k) | | partition | Integer partitions p(n) | | catalan | Catalan numbers | | bell | Bell numbers |
bashuv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/z3_solve.py" <command> <args>
| Command | Use Case | |---------|----------| | sat | Is this satisfiable? | | prove | Is this always true? | | optimize | Find min/max subject to constraints |
bashuv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/pint_compute.py" convert <value> <from_unit> <to_unit>
Example: convert 5 miles kilometers
bashuv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/math_router.py" route "<natural language request>"
Returns the exact command to run. Use when unsure which script.
When the request is "explain X" or "what is X", I reference these:
| Topic | Skill Location | Key Concepts | |-------|----------------|--------------| | Abstract Algebra | math/abstract-algebra/ | Groups, rings, fields, homomorphisms | | Category Theory | math/category-theory/ | Functors, natural transformations, limits | | Complex Analysis | math/complex-analysis/ | Analytic functions, residues, contour integrals | | Functional Analysis | math/functional-analysis/ | Banach spaces, operators, spectra | | Linear Algebra | math/linear-algebra/ | Matrices, eigenspaces, decompositions | | Mathematical Logic | math/mathematical-logic/ | Propositional, predicate, proof theory | | Measure Theory | math/measure-theory/ | Lebesgue, σ-algebras, integration | | Real Analysis | math/real-analysis/ | Limits, continuity, convergence | | Topology | math/topology/ | Open sets, compactness, connectedness | | ODEs/PDEs | math/odes-pdes/ | Differential equations, boundary problems | | Optimization | math/optimization/ | Convex, LP, gradient methods | | Numerical Methods | math/numerical-methods/ | Approximation, error analysis | | Graph/Number Theory | math/graph-number-theory/ | Graphs, primes, modular arithmetic | | Information Theory | math/information-theory/ | Entropy, coding, channels |
I decide based on your request:
"solve/calculate/compute" → SymPy (exact symbolic)
"is X always true?" → Z3 (constraint proving)
"convert units" → Pint
"explain/what is" → Topic skill for context
"prove formally" → Redirect to /proveUser: Solve x² - 5x + 6 = 0
Claude: uv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/sympy_compute.py" solve "x**2 - 5*x + 6" --var x
Result: x = 2 or x = 3User: Find eigenvalues of [[2, 1], [1, 2]]
Claude: uv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/sympy_compute.py" eigenvalues "[[2,1],[1,2]]"
Result: {1: 1, 3: 1} (eigenvalue 1 with multiplicity 1, eigenvalue 3 with multiplicity 1)User: Is x² + y² ≥ 2xy always true?
Claude: uv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/z3_solve.py" prove "x**2 + y**2 >= 2*x*y"
Result: PROVED (equivalent to (x-y)² ≥ 0)User: How many kilometers in 26.2 miles?
Claude: uv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/pint_compute.py" convert 26.2 miles kilometers
Result: 42.16 kmUse /prove when you need:
/math is for computation. /prove is for verification.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-18 | fail→fail | 7,842 | 3,859 | -51% | 1 | 1 | 0% | 1,264 | 2,422 | +92% | 0 | 0 | — |
case-01 | fail→fail | 11,493 | 14,437 | +26% | 1 | 1 | 0% | 2,718 | 5,075 | +87% | 0 | 0 | — |
case-02 | fail→fail | 15,690 | 15,561 | -1% | 1 | 1 | 0% | 2,966 | 4,582 | +54% | 0 | 0 | — |
case-03 | fail→pass | 18,289 | 12,629 | -31% | 1 | 1 | 0% | 3,369 | 4,080 | +21% | 0 | 0 | — |
case-04 | fail→pass | 13,538 | 1,635 | -88% | 1 | 1 | 0% | 3,223 | 2,114 | -34% | 0 | 0 | — |
case-05 | fail→fail | 5,153 | 4,793 | -7% | 1 | 1 | 0% | 1,184 | 2,736 | +131% | 0 | 0 | — |
case-06 | pass→fail | 5,505 | 4,452 | -19% | 1 | 1 | 0% | 1,230 | 2,027 | +65% | 0 | 0 | — |
case-07 | fail→pass | 12,699 | 1,474 | -88% | 1 | 1 | 0% | 2,869 | 1,997 | -30% | 0 | 0 | — |
case-08 | fail→fail | 7,203 | 12,754 | +77% | 1 | 1 | 0% | 1,721 | 3,290 | +91% | 0 | 0 | — |
case-09 | pass→pass | 16,292 | 2,408 | -85% | 1 | 1 | 0% | 3,650 | 2,262 | -38% | 0 | 0 | — |
case-10 | fail→fail | 11,302 | 5,517 | -51% | 1 | 1 | 0% | 2,319 | 2,143 | -8% | 0 | 0 | — |
case-11 | fail→fail | 5,026 | 5,668 | +13% | 1 | 1 | 0% | 1,012 | 2,142 | +112% | 0 | 0 | — |
case-12 | fail→pass | 4,304 | 4,222 | -2% | 1 | 1 | 0% | 857 | 2,494 | +191% | 0 | 0 | — |
case-13 | fail→fail | 13,691 | 12,654 | -8% | 1 | 1 | 0% | 2,289 | 4,175 | +82% | 0 | 0 | — |
case-14 | fail→fail | 15,674 | 15,302 | -2% | 1 | 1 | 0% | 2,774 | 4,220 | +52% | 0 | 0 | — |
case-15 | fail→fail | 16,749 | 11,597 | -31% | 1 | 1 | 0% | 3,226 | 4,268 | +32% | 0 | 0 | — |
case-16 | fail→fail | 20,089 | 11,510 | -43% | 1 | 1 | 0% | 2,656 | 3,960 | +49% | 0 | 0 | — |
case-17 | fail→fail | 13,679 | 9,280 | -32% | 1 | 1 | 0% | 2,519 | 3,477 | +38% | 0 | 0 | — |
case-19 | pass→pass | 7,372 | 8,069 | +9% | 1 | 1 | 0% | 1,735 | 3,336 | +92% | 0 | 0 | — |
case-20 | pass→pass | 13,547 | 11,267 | -17% | 1 | 1 | 0% | 2,172 | 3,836 | +77% | 0 | 0 | — |
case-21 | pass→pass | 2,798 | 2,248 | -20% | 1 | 1 | 0% | 417 | 2,096 | +403% | 0 | 0 | — |
case-22 | fail→fail | 21,674 | 17,571 | -19% | 1 | 1 | 0% | 4,365 | 5,277 | +21% | 0 | 0 | — |
case-23 | fail→fail | 7,450 | 6,618 | -11% | 1 | 1 | 0% | 1,233 | 2,045 | +66% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 19 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +13 percentage points is the difference between those two pass rates over the 19 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/29/2026 | +30% |
Other measured skills in the registry, with their headline benchmark lift.