perf: Replace exclusive lock with ReaderWriterLockSlim in ModuleScheduler (#1469)

* chore: Add .worktrees/ to .gitignore for parallel development

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* perf: Replace exclusive lock with ReaderWriterLockSlim in ModuleScheduler

Refactor ModuleScheduler locking strategy to reduce contention in the hot
execution path. The previous implementation held an exclusive lock during
read-only operations like querying module states and statistics.

Changes:
- Replace `object _stateLock` with `ReaderWriterLockSlim`
- Use read locks for query operations (GetStatistics, GetPendingModules,
  GetExecutingModules, ShouldExitScheduler snapshot creation)
- Use write locks for state mutations (MarkModuleStarted, MarkModuleCompleted,
  AddModule, FindReadyModules, CancelPendingModules)
- Move LINQ processing and logging operations outside of locks where possible
- Add proper disposal of the ReaderWriterLockSlim

This allows concurrent reads while maintaining exclusive access for writes,
improving performance in scenarios with many parallel module queries.

Fixes #1458

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* fix: Address code review feedback - race conditions and logging bug

1. ShouldExitScheduler: Keep read lock held while evaluating exit
   conditions to prevent state changes between snapshot and decision.
   The isDeadlocked check is now captured under the read lock before
   upgrading to write lock.

2. CancelPendingModules: Fix copy-paste bug where logging statement
   was hardcoded to log ModuleExecutionState.Completed instead of
   the original state before modification. Now captures original state
   before modifying and logs that value.

3. AddModule: Move dependency resolution inside the write lock to
   prevent race conditions where _moduleStates could change between
   resolution and processing. Logging of unregistered dependencies
   is done outside the lock to avoid holding the lock during I/O.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

Author: thomhurst