Retarget exercise 3 to net10.0 and verify with dotnet test

Verified with SDK 10.0.4xx (mise): builds clean, solution test green
deterministically, pristine test flaky as designed (failed 2/3 runs).
net7.0 runtime is out of support and unavailable, so the exercise now
targets net10.0; MSTest bumped 2.1.1 -> 3.6.4 to silence the NU1903
transitive-dependency warning. .gitignore for bin/obj.

ideas/split-loop-with-lambda/exercise3/README.md

@@ -5,7 +5,8 @@ A third, minimal example in the same kata theme: the sellers exist
 `Buyers` computes a buyer's shopping-list total, and `PriceList` stands for an
 external pricing service (it adds random surge pricing, so the test is flaky).
 
-Written in **C#** (MSTest, .NET 7 — same setup as the kata's `csharp/` folder).
+Written in **C#** (MSTest, .NET 10; verified with `dotnet test` on SDK
+10.0.4xx — pristine test is flaky by design, solution is deterministic).
 Run with `dotnet test test/BuyersTests.csproj`.
 
 Contents:
@@ -33,8 +34,9 @@ Next-time arc: marketing adds surge discounts, `PriceList` grows
 `DiscountFor(...)` — two methods to stub → extract a role interface
 (`IPricing`) instead of two lambdas. See the demo script's closing note.
 
-## Not verified by a build
+## Verified
 
-This exercise was authored without a local .NET SDK (structurally mirrored from
-the kata's `csharp/` project files). Run `dotnet test test/BuyersTests.csproj`
-once on a machine with the SDK before facilitating.
+Built and run with `dotnet test` (SDK 10.0.4xx): solution green deterministically,
+pristine test flaky (failed 2 of 3 runs, passed 1 — both surges rolled 0).
+MSTest bumped to 3.6.4 (kata ships 2.1.1, whose transitive Newtonsoft.Json trips
+a NU1903 vulnerability warning).