Verification record

The six automated suites report 599 passing tests for runtime 0.4.3. Native Chromium reports 110 checks across four browser journeys, and eight separate launcher checks exercise real-process persistence. These are recorded runs; this page does not claim that every check is continuously rerun.

Both existing save snapshots opened without migration at Patch 03 activation. Startup checks preserved recorded state and passed ledger audits. Mechanical content and story definitions remained byte-identical to Patch 02.

The quick actual-engine laboratory passes its 11 selected local gates. It includes 192 combat profiles, 512 encounter-condition cases, 1,296 profession cases and four synthetic economies. Those economies executed 2,283 commands and 94 trades, with exact replay and conserved inventory. The 88-command material journey and 81-command encounter-variation journey also replayed exactly.

The laboratory explicitly retains a HUMAN_REVIEW_REQUIRED release decision. Structural checks, synthetic producers and short earned journeys cannot establish a complete player career, human enjoyment, a stable large economy, or hundreds of hours of varied active play.

The current portal and agent deployment adds its own browser, security-boundary and real-execution checks. Those results appear in the deployment report and live agent panel; they do not replace game regression evidence.