Improve jit-regression-test skill with structured description and stop signals

.github/skills/jit-regression-test/SKILL.md

@@ -1,19 +1,28 @@
 ---
 name: jit-regression-test
-description: Extract a standalone JIT regression test case from a given GitHub issue and save it under the JitBlue folder. Use this when asked to create or extract a JIT regression test from an issue.
+description: >
+  Extract a standalone JIT regression test case from a given GitHub issue and
+  save it under the JitBlue folder. USE FOR: creating JIT regression tests,
+  extracting repro code from dotnet/runtime issues, "write a test for this JIT
+  bug", "create a regression test for issue #NNNNN", converting issue repro to
+  xunit test. DO NOT USE FOR: non-JIT tests (use standard test patterns),
+  debugging JIT issues without a known repro, performance benchmarks (use
+  performance-benchmark skill).
 ---
 
 # JIT Regression Test Extraction
 
-When you need to extract a JIT regression test case from a GitHub issue, follow this process to create a properly structured test under `src/tests/JIT/Regression/JitBlue/`.
+> 🚨 **Do NOT create a test when**: the issue has no reproducible code and you cannot compose a minimal repro, the issue is a duplicate of an existing test under `JitBlue/`, or the bug is in libraries/runtime rather than the JIT compiler itself.
+
+Extract a JIT regression test case from a GitHub issue into a properly structured test under `src/tests/JIT/Regression/JitBlue/`.
 
 ## Step 1: Gather Information from the GitHub Issue
 
 From the GitHub issue, extract:
-1. **Issue number** - Used to name the test folder and files (e.g., issue #99391 → `Runtime_99391`)
-2. **Reproduction code** - The C# code that demonstrates the bug. If no code provided, try to compose it yourself.
-3. **Environment variables** - Any DOTNET_* environment variables required to reproduce the bug
-4. **Expected behavior** - What the correct output/behavior should be
+1. **Issue number** → folder/file name (e.g., #99391 → `Runtime_99391`)
+2. **Reproduction code** — if none provided, compose a minimal repro yourself
+3. **Environment variables** — any `DOTNET_*` vars needed to reproduce
+4. **Expected behavior** — correct output/behavior
 
 ## Step 2: Create the Test Directory
 
@@ -52,8 +61,9 @@ public class Runtime_<issue_number>
 
 - **License header**: Always include the standard .NET Foundation license header
 - **Class name**: Match the file name exactly (`Runtime_<issue_number>`)
-- **Test method**: Use `[Fact]` attribute and name the method `TestEntryPoint()` or `Test()`
-- **Assertions**: Use Xunit's helpers or just throw plain exceptions.
+- **Test method**: `[Fact]` attribute, named `TestEntryPoint()` 
+- **Minimize the reproduction**: Strip to the minimal case that triggers the bug
+- **Use `[MethodImpl(MethodImplOptions.NoInlining)]`** when preventing inlining is needed to reproduce
 
 ### Example: Simple Test (from Runtime_99391)
 
@@ -91,6 +101,7 @@ public class Runtime_99391
 A custom `.csproj` file is **only required** when:
 - Environment variables are needed to reproduce the bug (such as `DOTNET_JitStressModeNames`)
 - Special compilation settings are required
+
 Otherwise, register the test file in the existing `src/tests/JIT/Regression/Regression_*.csproj` (`Regression_ro_2.csproj` is a good default) file and skip creating a new .csproj.
 
 If a custom .csproj file is needed, it should be located next to the test source file with the following name: `Runtime_<issue_number>.csproj`. Example:
@@ -111,9 +122,7 @@ If a custom .csproj file is needed, it should be located next to the test source
 </Project>
 ```
 
-## Important Notes
+## Tips
 
-- **No .csproj needed for simple tests**: Most tests only need the `.cs` file in a separate `src/tests/JIT/Regression/JitBlue/Runtime_*/` folder and being registered in `Regression_ro_2.csproj`.
-- **Look at recent tests**: When in doubt, examine recent tests under `src/tests/JIT/Regression/JitBlue/Runtime_*` for the latest conventions.
-- **Use `[MethodImpl(MethodImplOptions.NoInlining)]`**: When you need to prevent inlining to reproduce a JIT bug.
-- **Minimize the reproduction**: Strip down the test code to the minimal case that reproduces the issue.
+- **No .csproj needed for simple tests** — register the `.cs` file in `Regression_ro_2.csproj` instead.
+- **Look at recent tests** under `src/tests/JIT/Regression/JitBlue/Runtime_*` when in doubt about current conventions.