Disciples II original receiver continuation, 2026-11-20

Run: scratch/runs/20261120-disciples2-native-receiver-onward/. Original gameplay remains unproved. The prior 17 real two-CPUWorker cases off/on, ten focused groups and locked build were reused unchanged from 20261120-disciples2-native-receiver-gameplay-20261120; no qualification campaign was repeated. All four released observer/fixture paths match those prior tested hashes. No receiver or readiness repair is established here.

Before the fixed runtime budget, a fresh immutable holder compiled Git fd430d03e9d6037ed7054872fb5e0174526b6a6e plus only the released passive observer. Module SHA-256 is exactly cc4fb8225e07c1136c7e200484fb321827f20a756b71b599748237786fdc692a. The complete graph, actual Worker/process inheritance, 128 original media hashes and five loaded DLLs passed preflight. Manual capture/choice templates were prepared before any capture. No foreign RPC, Hexen or GLTron WIP was adopted into this graph.

The fresh original started at 03:33:20.837Z and its actual wait was signal 15 at 03:33:21.161Z, before an endpoint request or a frame was obtained. Its independent watchdog completed with deadline-or-disk, not an unknown external-signal receipt. The deadline was 04:33:40.802Z, so the watchdog's only earlier terminating branch was its actual /home free-space check below 1,500,000,000 bytes. This is an inference from its recorded wait and exact source, not a retained numeric watchdog sample. Its 50 ms sampler did not retain that value; the outer 500 ms sampler missed the transient and recorded a minimum of 1,500,045,312 bytes. A later direct sample was below the floor. Do not replace these distinct observations with a fabricated first-breach timestamp/value, whole-task capacity claim, or guest Quit claim.

The runtime is permanently UNKNOWN/closed. There were zero endpoint requests, inputs, tickets, captures and original consumer observations. Hover, New Quest, map consequence and ordinary Quit are unproved. No screenshot was borrowed from the old scope. No closed endpoint was replayed. A query would still not prove POINT consumption or its visible consequence.

The new concrete boundary is the independent watchdog's pre-frame disk gate. Its combined reason omits its own numeric floor sample, so the exact transient size remains unknown. There is no evidence supporting a receiver-code change. Existing observer/fixture WIP remains released and uncommitted. Git, exact patch references, graph/media pins, unique driver inputs, actual waits and offline cause checks are retained; no source/media bundle was created.

All runtime parent waits and both pipe EOFs completed; native-supervise reports leaked:false. Browser/display were not created. Box ID null, state NOT_CREATED, expiry null, stop receipt null. No agents, Codex workers, Python, Wine or deployment were used.