Automated Mock Benchmark Update — Review

This is a clean, automated documentation-only update. No logic changes, no code paths affected. A few observations worth noting for the record:

Observation: Unit change in CombinedWorkflow chart

In CombinedWorkflow.md, the Mermaid chart's y-axis changed from "Time (μs)" to "Time (ns)" and the bar values shifted accordingly:

- bar [2.057, 2.982, 1.714, 406.602, 17.79, 18.321]   // μs
+ bar [1026.9, 1373.1, 882.5, 118764.5, 8453.9, 8383.5] // ns

The new values are self-consistent (table and chart both in ns), so the chart is now more correct than before — the previous run's chart had mixed units that happened to line up visually because the raw numbers were small floats in μs. This looks like a fix in the benchmark-to-docs generator.

Observation: Large performance swings in CombinedWorkflow

The CombinedWorkflow results show some large day-over-day swings (TUnit.Mocks ~50% faster, Moq ~70% faster). This is expected noise in a shared CI environment (Ubuntu runners are not dedicated hardware), but if these numbers are used for external comparisons or marketing, it may be worth considering running benchmarks on a dedicated/pinned machine or taking median over multiple runs.

Summary

The update is correct and the generated output is consistent. The code generation itself is working as intended. ✅