---

name: openfermion description: "Guide OpenFermion workflows for fermionic and qubit operator algebra, Hamiltonian construction, quantum-chemistry data, circuit synthesis, sparse analysis, measurements, and resource-oriented utilities." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

OpenFermion

Use this skill when a task names OpenFermion or asks for Python workflows around fermionic systems, second quantization, qubit Hamiltonians, quantum chemistry, Jordan–Wigner/Bravyi–Kitaev mappings, Trotter/UCC circuits, RDMs, measurement grouping, or sparse operator analysis.

Fast route

1. Install and smoke-check. Install the public openfermion distribution in Python >=3.10, then run python -c "import openfermion; print(openfermion.__version__)". The package uses Cirq core, NumPy, SciPy, SymPy, HDF5, and related runtime dependencies. See troubleshooting for optional extras and import failures.
2. Select the owning route. Use the focused sub-skill that matches the user's artifact, not the source package directory name:
   • operators-and-transforms for symbolic operators, normal ordering, and fermion-to-qubit mappings.
   • hamiltonians-and-chemistry for Hubbard/jellium/model Hamiltonians and molecular metadata or tensors.
   • circuits-and-simulation for Trotter, UCC, Slater/Gaussian preparation, Cirq circuits, and VPE circuits.
   • analysis-and-measurements for sparse matrices, eigensolvers, RDMs, measurement grouping, and resource functionals.
3. Record the contract before computing. Preserve operator family, mode or qubit ordering, coefficient units, tensor shapes, n_qubits, boundary or particle-number choices, optional dependency state, and the expected output.
4. Keep numerical work bounded. Construct tiny fixtures first; validate dimensions and Hermiticity before matrix conversion, dense diagonalization, RDM contraction, or circuit expansion. Do not infer scalability from a smoke test.
5. Handoff across routes explicitly. A model builder supplies an operator contract to the mapping route; a mapping supplies a qubit operator to the circuit or analysis route; analysis returns shapes, convergence signals, and numerical results. Do not silently choose conventions at a boundary.

Public package boundary

OpenFermion is primarily a Python library, not a package-specific command-line application. Install external simulators, electronic-structure packages, or hardware plugins separately. The optional resources extra adds PySCF, JAX, JAXlib, and ASE for selected resource-estimation workflows; it is not required for core operator, transform, Hamiltonian, circuit, or sparse-analysis usage. PubChem lookup is network-dependent even though the base distribution includes its client library. Never claim that MolecularData metadata is a completed quantum-chemistry calculation.

For verified signatures and return-shape notes, read api-index.md. For cross-cutting install, optional package, file, network, and numerical failures, read troubleshooting.md. Run scripts/check_openfermion.py for a safe public import and capability diagnostic; it performs no network access or file writes. Read repo-provenance.md before deciding whether this skill is stale for a particular OpenFermion checkout.

Self-contained smoke helpers

Each route owns a tiny helper that is independent of the original checkout:

• operators-and-transforms/scripts/smoke_transform.py maps a two-mode hopping term and can report a sparse shape.
• hamiltonians-and-chemistry/scripts/build_tiny_hamiltonian.py emits a small Hubbard or molecular-metadata summary without writing files.
• circuits-and-simulation/scripts/circuit_smoke.py reports a bounded Cirq circuit and optional Slater preparation.
• analysis-and-measurements/scripts/sparse_analysis_smoke.py converts a tiny qubit operator and checks a ground-state residual.

Use --help first and keep all input dimensions within the helper's documented bounds. These are structural checks, not scientific benchmarks.