trace-claude-code
Automatically trace Claude Code conversations to Braintrust for observability. Captures sessions, conversation turns, and tool calls as hierarchical traces.
Unified math capabilities - computation, solving, and explanation. I route to the right tool.
$ npx -y skills add parcadei/Continuous-Claude-v3 --skill math-unified --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/math-unifiedContext preview
The summary Claude sees to decide when to auto-load this skill.
Unified math capabilities - computation, solving, and explanation. I route to the right tool.
name: math description: Unified math capabilities - computation, solving, and explanation. I route to the right tool. triggers: ["calculate", "compute", "solve", "integrate", "derivative", "eigenvalue", "matrix", "simplify", "factor", "limit", "series", "differential equation", "unit convert", "explain", "what is", "how does"] allowed-tools: [Bash, Read, Write] priority: high
**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 |
---
uv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/sympy_compute.py" <command> <args>
| Command | Description | Example | |---------|-------------|---------| | `solve` | Solve equations | `solve "x**2 - 4" --var x` | | `integrate` | Definite/indefinite integral | `integrate "sin(x)" --var x --lower 0 --upper pi` | | `diff` | Derivative | `diff "x**3" --var x` | | `simplify` | Simplify expression | `simplify "sin(x)**2 + cos(x)**2"` | | `limit` | Compute limit | `limit "sin(x)/x" --var x --point 0` | | `series` | Taylor expansion | `series "exp(x)" --var x --point 0 --n 5` | | `dsolve` | Solve ODE | `dsolve "f''(x) + f(x)" --func f --var x` | | `laplace` | Laplace transform | `laplace "sin(t)" --var t` |
**Matrix Operations:** | Command | Description | |---------|-------------| | `det` | Determinant | | `eigenvalues` | Eigenvalues | | `eigenvectors` | Eigenvectors with multiplicities | | `inverse` | Matrix inverse | | `transpose` | Transpose | | `rref` | Row echelon form | | `rank` | Matrix rank | | `nullspace` | Null space basis | | `linsolve` | Linear system Ax=b | | `charpoly` | Characteristic polynomial |
**Number Theory:** | Command | Description | |---------|-------------| | `factor` | Factor polynomial | | `factorint` | Prime factorization | | `isprime` | Primality test | | `gcd` | Greatest common divisor | | `lcm` | Least common multiple | | `modinverse` | Modular inverse |
**Combinatorics:** | Command | Description | |---------|-------------| | `binomial` | C(n,k) | | `factorial` | n! | | `permutation` | P(n,k) | | `partition` | Integer partitions p(n) | | `catalan` | Catalan numbers | | `bell` | Bell numbers |
---
uv run python "$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 |
---
uv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/pint_compute.py" convert <value> <from_unit> <to_unit>
Example: `convert 5 miles kilometers`
---
uv run python "$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 /prove
---
User: Solve x² - 5x + 6 = 0 Claude: uv run python "$CLAUDE_PROJECT_DIR/.claude/scripts/cc_math/sympy_compute.py" solve "x**2 - 5*x + 6" --var x Result: x = 2 or x = 3
User: Find eigenvalues of [[2, 1], [1, 2]]
Claude: uv run python "$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 km
---
Use `/prov
A persistent, learning, multi-agent development environment built on Claude Code Continuous Claude transforms Claude Code into a continuously learning system that maintains context across sessions, orchestrates specialized agents, and eliminates wasting
Repo: parcadei/Continuous-Claude-v3
Automatically trace Claude Code conversations to Braintrust for observability. Captures sessions, conversation turns, and tool calls as hierarchical traces.
Guide for integrating Agentica SDK with Claude Code CLI proxy
Reference guide for Agentica multi-agent infrastructure APIs