Recipes to find ThreadLocals and determine their mutation

[MBoegers]

What's changed?

This PR introduces a comprehensive suite of search recipes to identify ThreadLocal variables that could be migrated to ScopedValue in Java 25+ (JEP 506):

What's your motivation?

Java 25 introduces ScopedValue as a modern alternative to ThreadLocal for immutable thread-local values, offering:

• Better performance (no need for weak references or cleanup)
• Safer scoped value passing (values cannot be mutated)
• Clearer lifecycle management

Unfortunately not all ThreadLocal instances qualify equally.

These recipes help development teams:

1. Assess their codebase's readiness for ScopedValue migration
2. Identify which ThreadLocals can be safely migrated
3. Understand the scope of (manual) refactoring required

Anything in particular you'd like reviewers to focus on?

1. Cross-file analysis implementation - The AbstractFindThreadLocals base class uses scanning to detect mutations across multiple compilation units. Please verify this correctly handles all edge cases.
2. Mutation detection completeness - Review that we're catching all ways a ThreadLocal can be mutated (set, remove, initialValue override, etc.)

Anyone you would like to review specifically?

Have you considered any alternatives or workarounds?

• Dataflow analysis: Considered but opted for AST-based visitor pattern to maintain consistency with OpenRewrite patterns and avoid complexity
• Single-file analysis only: Rejected as it would miss many cross-file mutations, giving false positives
• More granular recipes: Could have created more specific recipes but the current three categories cover the main use cases effectively

Any additional context

Checklist

• I've added unit tests to cover both positive and negative cases
• I've read and applied the recipe conventions and best practices
• I've used the IntelliJ IDEA auto-formatter on affected files

[MBoegers]

needs another round of personal review, but open for comments.