Round 2 verification audit for C065

Model: gpt-5

Claim: When drive (A) is applied, the blue-sideband tone causes an AC Stark shift of Delta ~= 2 pi x 25 kHz on |down>.

Source alignment: source/supp_content.tex:125-125 (Supp. Mat. S4)

Prior official verdict: not_verified with failure_reason data_unavailable.

Executable evidence: run.py. Sandbox rerun logs: run.log.

Independent audit: I checked the paper inventory and source context. This claim depends on measured/raw experimental data or device-specific calibration values, and inventory.json records no paper-provided dataset/code artifact. The correct action is not to synthesize data; the data_unavailable verdict is preserved.

Decision: Round 2 verdict is not_verified with failure_reason data_unavailable and limitations ['data_unavailable']. Notes: Paper: drive-(A) blue-sideband tone causes AC Stark shift Delta ~= 2pi x 25 kHz on |down>. MEASURED/device value; no raw data. The 2-level formula Delta = Omega^2/(4*det) is NOT closeable from the claim's stated constants (the relevant tone Rabi frequency Omega on |down> is not given; only det ~2pi x 0.9 MHz). Cross-check: required Omega ~ 2pi x 300 kHz is a plausible carrier Rabi frequency, and main.tex independently quotes a ~2pi x 17 kHz drive-(A) AC Stark shift, confirming the 10s-of-kHz scale, but the specific 25 kHz cannot be reconstructed -> data_unavailable. No data synthesized.