[aarontrowbridge]

Related Issue

None filed yet — draft from the skills-integrity loop (proposal S-fcz-01 item 1, finding F-fcz-13). File an issue before marking ready.

Description

New public skill scqubits: construction of fluxonium/transmon/Kerr subsystems and composite spaces, the Kerr sign map (K_scq = -K/2), the n_operator-is-plain-n offset-charge trap (couple via n - n_g; subclass snippet), dressed-state labelling hazards (DE ordering drops overlap² ≤ 0.5 to None; BE gives no purity; Dumas branch analysis), HilbertSpace.eigenvals/eigensys shadowing after generate_lookup (read hs["evals"][0]), the missing matrixelement_table/hamiltonian() on oscillators, and a truncation-convergence procedure with a per-quantity tolerance.

Evidence

H0 spectrum-calibration observation note (conventions, ladder, label cross-check, API findings), project helpers builders.py/metrics.py/ngcorr.py, D2a/D2b n_g notes, friction note F-fcz-13.

Honesty flags

• Draft, verified on one fork (harmoniqs/scqubits main 6b32be69, 4.3.1+g6b32be69c). Each API fact is marked run (smoke-tested 2026-11-07: Kerr map, plain n_operator, lookup/eigenvals length mismatch 8 vs 6, DE None on a degenerate pair, BE finite labels, hasattr(KerrOscillator,"hamiltonian")==False, matrixelement_table raises) or src (file:line). The card's two code snippets were executed.
• Numbers marked campaign are from the fluxonium-cz geometry and must be re-measured. Whether the physical coupling is to n or n - n_g is stated as a modelling decision, not a ruling.
• Not covered (still proposals): the dressed-basis → optimal-control export recipe and a Modal/batch runner (S-fcz-01 items 2–3).

Verification

cd packages/extension && node scripts/lint-skills.mjs --known scripts/lint-skills-known.txt → lint-skills: 41 skills, 0 error(s), 5 warning(s) — ok (the 5 warnings are pre-existing, none in this card). Also run with the unmerged private-package rule from origin/private-skills-out (--denylist): 0 hits in this card. Tests: skills_integrity, skill_salvage_851, workflow_skills_public 42/42 pass; the full vitest run shows the same 128 failures as clean origin/main in this environment (fleet/sidebar/pack tests; unrelated) plus one flaky fake_serve_reaper case that passes alone. (skill-drift lint, structural: 0 failures.)

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

• X
• Mastodon
• Reddit
• LinkedIn