[cyzcz]

Companion executable M1/M2 verification spike for Wildcard proposal #3764.

Summary of Measured Results (spikes/w-linear-must-consume/RESULTS.md)

• H1 (Linear Must-Consume Soundness) — PASS: linear_obligation_verifier.py admits branch-complete linear consumption (01-linear-branch-complete-admitted.hy) and rejects early ? try-return leaks and partial-branch drops (02-linear-early-try-leak-rejected.hy → 2x HYB-E3701).
• H2 (Go Compiler Escape Oracle Agreement) — PASS: go build -gcflags="-m=2" verifies RunLinearStackLoop inlines (cost 79 <= budget 80) with 0 heap escapes, while RunSyncPoolLoop exceeds the inline budget (cost 201 > 80) and forces new(Packet) escapes to heap.
• H3 (Throughput & Zero-Allocation Benchmark) — PASS: BenchmarkLinearStackLoop runs at 2,885 ns/op (0 B/op, 0 allocs/op) vs BenchmarkSyncPoolLoop at 6,127 ns/op (2.12x speedup, go test -race ./... green).
• F1–F2 (Pre-registered Falsifiers) — FIRED: Double-consume (HYB-E3003) and dynamic interface-sink heap escape (HYB-E3702 / FormatterSink.Observe(any)) both fire as predicted.

Verification

• python3 spikes/w-linear-must-consume/linear_obligation_verifier.py --check --live-go
• python3 -m unittest discover -s spikes/w-linear-must-consume/tests -v
• cd spikes/w-linear-must-consume/work && go test -race -v ./... && go test -bench=. -benchmem ./...

[limitlessprojects]

Track-3 (Wildcard) routing — intake acknowledged

Thank you @cyzcz. This executable M1/M2 spike for Wildcard proposal #3764 (linear must-consume obligations & escape-certified stack regions) is routed to the Track-3 process per BOUNTY.md / MOONSHOT.md:

• Status: proposal [wildcard] Linear Must-Consume Obligations & Escape-Certified Stack Regions: Flipping GUARANTEES Rows 10 & 13 without GC Finalizers or sync.Pool #3764 intake triage is scheduled — this spike PR is logged as its companion artifact ("not yet maintainer-verified", same standing as recent spikes spike(w012): harden in-process WASM analyzer M1 #556/feat(spikes): W-019 isolated WASM task cells spike (#462) #566/feat(spikes): complete W-018 static rate task networks spike with runtime FIFO guards and complete results (#378) #572). It stays open until the maintainer verification pass; it is not a main-landing PR and will not be merged through the regular compiler battery.
• Family pointers for triage (proposal + spike will be read against the existing ledger rows): W-001 (linear region capabilities, [wildcard] Kill syntactic alias/loan/freeze tracking: replace path-sensitive value-flow (freeze.rs + closures.rs) with Linear Region Capabilities (cap[R]) and Isolated Sub-Graphs[wildcard] #304), W-002 (ownership-as-capability, [wildcard] Delete the alias graph: make ownership a capability the type carries #279), W-003 (own/share/view + runtime bit, [wildcard] Kill the reconstruction: ownership as a visible capability surface (linear types + one runtime bit) replacing alias/loan/freeze #193), W-016 (A5 Go-escape check-site, [wildcard] Make the A5 Go-escape a check-site with no Go API database: a capture-write at a foreign-call boundary takes managed #327), W-017 (take-and-detach move lowering, [wildcard] Stop erasing move to a no-op: Operational Take-and-Detach Lowering in hy-emit + Deleting the CGo/Rust Staticlib Bridge #200). If [wildcard] Linear Must-Consume Obligations & Escape-Certified Stack Regions: Flipping GUARANTEES Rows 10 & 13 without GC Finalizers or sync.Pool #3764's linear must-consume model lands in the same design space, the review will judge it as a competing/merging family member rather than an independent row, per the ledger's dup-credit practice.
• What the verification pass will check (from the PR's own claims): linear_obligation_verifier.py admit/reject behavior on the branch-complete linear fixtures, the escape-certification H1 PASS evidence, and whether the M2 "learning" is publishable as-is (success or failure alike — a published failure is a valid deliverable).
• Review window: the monthly wildcard review (1st of the month) — next slot 2026-12-01. M1 ($250) on acceptance, M2 ($250) on published learning, per BOUNTY.md.

No silent closes — this PR stays open until the review window resolves it.

[limitlessprojects]

Ledger update: the parent proposal is now intake-acked and ledgered (W-024 in MOONSHOT.md, review slot 2026-12-01). This spike PR is referenced from the row as the companion artifact under maintainer verification — standing unchanged (open, not yet maintainer-verified).