Check 1 (Verbatim Priority): The verbatim quote — 'Jia Beiju sold more than a million COVID tests for nearly $4 million through his California based company to customers across the country' — is the authoritative text for assessment. Check 2 (Extraction Fidelity): The claim_text accurately reflects the verbatim; no extraction-introduced language creates a divergence. Check 7 (ASR Consideration): The name 'Jia Beiju' versus the correct 'Jia Bei Zhu' follows a phonetic pattern consistent with ASR transcription error ('Bei Zhu' → 'Beiju'), and the claim has been flagged for audio verification. Regardless of cause, the name in the verbatim is inconsistent with every available authoritative source — all DOJ-derived reporting and U.S. Attorney's Office materials consistently identify the defendant as 'Jia Bei Zhu.' This constitutes a specific cited-fact error (a proper noun) that triggers the Wrong-Cited-Fact Gate (Check 3): the directional meaning of the claim survives correction entirely, routing to MOSTLY TRUE rather than FALSE. The quantitative figures — more than one million tests sold for nearly $4 million — are confirmed by multiple independent sources citing the U.S. Attorney's Office press release. The description 'California-based company' is geographically accurate but less precise than the DOJ-sourced 'Fresno-based' designation; under the directional understatement standard, this is an immaterial imprecision since Fresno is in California and every stated fact is confirmed. The core assertion — that the named individual, a Chinese national, sold over one million COVID tests for nearly $4 million through his California-based company — is substantially correct, with the name error being the sole identifiable inaccuracy that does not reverse the directional meaning (per MOSTLY TRUE standard: core assertion substantially correct with one identifiable inaccuracy that does not reverse directional meaning). Minimum source requirements for MOSTLY TRUE are met: multiple independent Tier 4 sources confirm the core assertion and identify the specific name inaccuracy.