Perf: Eliminate all heap allocations in mutation result creation and materialization

[rian-be]

Summary

Eliminate per call heap allocations associated with MutationResult<TState> creation and MutationHistoryEntry/MutationAuditEntry materialization by converting output DTOs to readonly record struct, caching singletons for success/empty defaults, and pre-caching supporting objects in benchmark setup.

Changes

• MutationResult<TState> converted from class to readonly record struct zero allocation per Success()/Failure()/PolicyBlocked()
• ValidationResult Success() returns cached singleton
• MutationMetrics cached Empty singleton, shared empty dictionary
• ChangeSet cached Empty singleton
• SideEffect.ResolveContract ConcurrentDictionary cache eliminates repeated reflection
• MutationHistoryEntry / MutationAuditEntry converted from class to readonly record struct zero allocation per materialization
• MutationHistory FirstMutationAt/LastMutationAt fixed for struct FirstOrDefault()

Benchmark Results

Benchmark	Before	After	Δ Time	Δ Alloc
Success_NoSideEffects	88.25 ns / 528 B	18.61 ns / 0 B	−79%	−100%
Success_SingleSideEffect	280.39 ns / 792 B	18.72 ns / 0 B	−93%	−100%
Success_MultipleSideEffects	899.22 ns / 1960 B	18.79 ns / 0 B	−98%	−100%
HistoryEntry_Materialization	688.7 ns / 896 B	23.40 ns / 0 B	−97%	−100%
AuditEntry_Materialization	666.0 ns / 848 B	24.47 ns / 0 B	−96%	−100%

Closes #94