qutip

Overview of qutip skill

What qutip is for

The qutip skill helps you use QuTiP, the Quantum Toolbox in Python, for open quantum systems, dissipation, time evolution, and quantum optics workflows. It is a good fit when you need a qutip guide for master equations, Lindblad dynamics, decoherence models, cavity QED, or state/operator simulation in research code.

Who should install it

Install this qutip skill if you are a scientist, engineer, or student working in quantum physics simulation and you want faster, more reliable outputs than a generic prompt. It is especially useful for Scientific users who need working Python examples, solver selection, or help translating physics notation into QuTiP objects.

When it is a strong fit

This skill is strongest for modeling closed and open systems, checking expectation values, plotting dynamics, and exploring advanced features like Floquet methods or Bloch-sphere visualization. It gives you a practical path from theory to executable code instead of a broad repo skim.

When not to use it

Do not use qutip for circuit-based quantum computing, hardware execution, or algorithm benchmarking. If your task is quantum algorithms or device workflows, qiskit, cirq, or pennylane are a better match than qutip.

How to Use qutip skill

Install qutip in your skill workflow

Use the qutip install command in the skills manager, then confirm the skill files are available before you ask for code or analysis. A typical install looks like:

npx skills add K-Dense-AI/claude-scientific-skills --skill qutip

If your environment already uses uv, QuTiP itself installs with uv pip install qutip.

Give qutip the right input shape

The best qutip usage starts with a physics statement, not a vague request. Include:

• system type: qubit, cavity, oscillator, spin chain, etc.
• closed vs open dynamics
• Hamiltonian, collapse operators, and initial state if known
• solver target: sesolve, mesolve, mcsolve, or frequency-domain methods
• what you want back: time traces, steady state, Bloch sphere, Wigner function, or plots

A strong prompt looks like: “Use qutip to simulate a driven two-level system with decay, compute ⟨σz⟩ over time, and explain how to set c_ops.”

Read these files first

Start with SKILL.md, then inspect the support references that match your task:

• references/core_concepts.md for Qobj, states, and operators
• references/time_evolution.md for solver choice and dynamics setup
• references/analysis.md for expectation values and entropy
• references/visualization.md for Bloch and phase-space plots
• references/advanced.md for Floquet and other specialized methods

Use a workflow that avoids rework

For best qutip usage, ask for one layer at a time: define the system, select the solver, run the evolution, then add analysis or visualization. This reduces errors from mixing Hamiltonian setup, solver syntax, and post-processing in one oversized request. If you already have code, ask the skill to adapt it to QuTiP conventions rather than rewrite everything from scratch.

qutip skill FAQ

Is qutip only for open quantum systems?

No. Open systems are a major strength, but qutip also handles closed-system unitary evolution, operator algebra, and state preparation. The deciding question is whether you need physics-oriented simulation rather than quantum-circuit execution.

Do I need to know QuTiP before using the qutip skill?

No. The qutip skill is suitable for beginners if you can describe the physical system and the quantity you want to compute. You get better results when you name the model ingredients clearly, but you do not need to know every API call in advance.

How is qutip different from a normal prompt?

A normal prompt may produce plausible code, but the qutip skill is organized around QuTiP’s actual workflow: quantum objects, solver selection, expectations, and visualization. That reduces guesswork when choosing between sesolve and mesolve, or when converting equations into Python objects.

When should I choose something else?

Choose another tool if your task is about gate-level circuits, noise models for devices, or algorithmic quantum computing. qutip is best when the question is “How does this quantum system evolve?” rather than “How do I compile or run a circuit?”

How to Improve qutip skill

State the model before asking for code

The biggest quality gain comes from specifying the system clearly: Hilbert-space size, basis, drive terms, dissipation channels, and measurement targets. For example, “two-level atom with spontaneous emission and drive” is much better than “simulate a qubit.”

Tell qutip what output you need

If you want better results from qutip, say whether you need runnable Python, derivation help, parameter sweeps, or plotting code. A request like “return a mesolve example plus a plot of population decay and a note on choosing c_ops” is more actionable than “use qutip for this problem.”

Watch the common failure modes

The most common problems are choosing the wrong solver, forgetting tensor dimensions, and under-specifying collapse operators or initial states. If the first answer looks too generic, add the missing physics rather than asking for a broader explanation.

Iterate with one correction at a time

Improve qutip outputs by fixing one layer per follow-up: first the model, then the solver, then diagnostics, then visualization. If your result is close but not usable, ask for an update that preserves the existing code and changes only the specific part that is wrong.