cirq
by K-Dense-AIcirq is Google Quantum AI’s open-source Python framework for building, simulating, and running quantum circuits. Use the cirq skill for noise-aware circuit design, quantum characterization experiments, and Google hardware workflows. It is best for low-level circuit work; for IBM hardware use qiskit, for quantum ML with autodiff use pennylane, and for physics simulations use qutip.
This skill scores 78/100, which means it is a solid listing candidate for directory users who want Cirq-specific guidance instead of a generic quantum prompt. The repository provides enough operational detail to trigger the skill correctly, understand its fit for Google Quantum AI / noise-aware circuit work, and begin using it with less guesswork, though it is not yet fully polished for install-time decision making because it lacks supporting files and some workflow depth beyond the main SKILL.md.
- Strong triggerability: the description clearly says when to use Cirq, including Google Quantum AI hardware, noise modeling, and low-level circuit design.
- Good operational clarity: the SKILL.md includes installation commands for cirq and related hardware integrations, plus a basic circuit quick start example.
- Substantial workflow content: the body is large (10k+ chars) with many headings and explicit repo/file references, suggesting more than a placeholder skill.
- No install command in the skill metadata/frontmatter and no support files, so setup/adoption guidance is centered mainly in SKILL.md.
- The repository appears focused on a single markdown skill file, so users should expect limited packaging, tests, or auxiliary references.
Overview of cirq skill
What cirq is for
The cirq skill helps you work with Google Quantum AI’s open-source Python framework for building, simulating, and running quantum circuits. It is most useful when your job is to express circuits clearly, test them on simulators, or prepare code that targets supported quantum hardware and providers.
Best-fit use cases
Use cirq when you need noise-aware circuit design, quantum characterization experiments, or Google Quantum AI workflows. It is a strong fit for low-level circuit work where qubit layout, gate sequence, and measurement strategy matter more than high-level abstractions.
When cirq is the wrong choice
If you need IBM hardware, qiskit is usually the better fit. If your task is quantum machine learning with autodiff, pennylane is a better default. For physics simulation workflows, qutip may be more appropriate than cirq.
How to Use cirq skill
Install cirq in your workspace
Install the skill with:
npx skills add K-Dense-AI/claude-scientific-skills --skill cirq
For runtime dependencies, the core package is usually:
uv pip install cirq
Add provider packages only when needed, such as cirq-google, cirq-ionq, cirq-aqt, cirq-pasqal, or azure-quantum for specific backends.
Give cirq the right input shape
The cirq skill works best when you specify: the circuit goal, the qubit count or topology, the backend or simulator, and whether you need noise, sampling, or hardware execution. A vague request like “build a quantum circuit” leaves too many choices open. A stronger prompt is: “Using cirq, create a 3-qubit GHZ circuit, simulate 1,000 shots, and explain how to adapt it for Google Quantum Engine.”
Read the right files first
Start with SKILL.md to get the workflow, installation, and quick-start patterns. If your local copy includes extra docs, inspect README.md, AGENTS.md, metadata.json, and any rules/, resources/, references/, or scripts/ folders before generating code. For this repository, SKILL.md is the main source of truth.
Practical workflow for better output
Use cirq in two passes: first, define the circuit structure and validation target; second, choose simulator or hardware integration details. If you are generating code, ask for a minimal working example first, then extend it with parameterization, measurement keys, or backend-specific packaging. This reduces hallucinated APIs and makes the cirq usage easier to verify.
cirq skill FAQ
Is cirq beginner-friendly?
Yes, if your goal is simple circuit construction and simulation. It becomes harder when you move into hardware constraints, backend selection, or noise modeling, where exact input details matter more.
What makes cirq different from a generic prompt?
A generic prompt often returns a plausible quantum example. The cirq skill is better when you need code aligned with Cirq’s actual objects, install path, and backend-specific packages, which lowers the chance of wrong imports or mismatched execution assumptions.
When should I not use cirq?
Do not use cirq when the target ecosystem is clearly not Google-aligned or when the task is primarily symbolic math, classical ML, or physics analysis without circuit construction. In those cases, a different quantum or scientific stack will usually produce cleaner results.
How to Improve cirq skill
Specify the execution target
The biggest quality boost comes from naming the target early: simulator, Google Quantum Engine, IonQ, AQT, Pasqal, or Azure Quantum. This determines which cirq install path, imports, and API choices are valid.
Provide circuit constraints up front
State the number of qubits, gate family, measurement format, and whether the circuit must be parameterized or noise-aware. For example, “Build a 2-qubit variational circuit with sympy parameters and explain how to bind values before simulation” is much better than “make it variational.”
Iterate from a testable baseline
Ask for a minimal runnable circuit first, then request extensions like readout error mitigation, backend migration, or experiment logging. If the first result is off, correct the missing constraint rather than asking for a full rewrite; cirq outputs improve fastest when the prompt narrows the hardware, sampling, and circuit-shape assumptions.