EscapistsReintroduced/docs/evidence/hq-native-checkpoint-2026-10-03

Native mission and HQ checkpoint — 3 October 2026

The normal scripted escape route completes 64 assertions; the corrected HQ fixture completes 172. Both retain 13 engine errors and strict qualification failure, with no script/runtime failure, unchanged inputs and confirmed hidden/muted cleanup. review.json binds all four native archives, including both failed patrol attempts.

The Workbench fixture now leaves prison through canonical admission and prepares two original mobile responders through ordinary finite site preparation. It checks a copied 45 m radius against the actual native getter, queue preemption, injected-contact pause/resume and home-Defend restoration. The separate six-member permanent patrol completes five ordered native moves and at least 8 m displacement. The generic historical completion label is not evidence of same-group physical search/resumption. Radio transit, 110 m area coverage, multiplayer and FPS remain unproven.

All 305 registered regression programs pass on final unchanged inputs. The first 305-program pass, proposal baselines, compiled failing controls, independent reviews and four-file final delta are retained. The normal route preceded the final diagnostic changes; exact differences are listed. Only two prose documents were updated after the final suite, with their final text and hashes retained.

Original 533 referenced files, staged Git content and four sealed candidates remain unchanged. This checkpoint does not rebuild or publish a package. next-work contains a read-only plan/resource metadata for independent static crash aftermath, not an implemented feature; measured collision bounds and replicated wrapper authoring remain prerequisites. No third-party asset bytes are included.