Match original Ecaz sample snapshots to their native Occupy and Advanced assassination state. Preserve older absent-power samples and resume the original scenario instead of rejecting fresh saved output. Update stale fixture reactions and module expectations while retaining meaningful Banker guidance and positive legal commitments. Correct the unsupported rules-completion claim.
Validation: typecheck/lint, 173 affected tests across 16 files, bounded independent source review, and one direct Ecaz continuation completed through 178 accepted actions. Runtime fingerprint 95d009d3957650b08bb24826b70aa5f00a74cdd9992b6a6c222f0665d3e3759c; documentation-only proof update a91ab248109bee74a74373fb544027a2e7cc8f178e26d2c9dad475a9618eb64d. Full rules, comprehensive assurance and deployed acceptance remain open. No saved-room or provider changes.
Prototype optional one-spice collection from distinct entered territories along exact legal ground routes. Share route and collection checks between engine, controls and all four AI profiles. Preserve per-pile opt-out, multiple source sectors, physical force/spice custody and unchanged legacy moves. Apply the documented per-movement Hajr scope inference; special relocation, combined modules and ambiguous multiple-pile collection remain unfinished.
Bind Fremen declarations to their actual chosen route even when a shorter route reaches the endpoint. Preserve that commitment across saved responses, collect only on successful movement, and hide its internal proof from every player projection. Independent source, rules and persistence review completed.
Validation: 52 focused tests; typecheck, lint, 4504 offline tests, production build and 43 HTTP tests passed. SQLite restart, exact concurrent settlement, cancellation and legacy continuation passed. Browser exercised genuine assignment, route editing, collection opt-out/reselect, three moved forces, two collected spice, automatic notices and refresh with unchanged private cards. Two genuine four-profile Basic/Advanced samples completed in 196/89 actions with 6/3 JSON restorations and no rejected candidates; neither used Sandmaster collection, so targeted tests establish the effect.
Preserved all 495 opening rooms and original seat/recovery rows (529 rooms after QA/HTTP). SQLite integrity and foreign keys passed. The healthy server was reused; no restart, reset or automation.
Verified source tree: aec472f2dacdd49a13216a92bb4fb715af9d7ee028de6dd007cd496142eb06b1
Quintine