fix: Keep booleans and numbers distinct in the in operator

[keelerm84]

Summary

The in clause operator incorrectly matched boolean context values against numeric clause values (and vice versa). Python's bool is a subclass of int, so True == 1 and False == 0 evaluate as true, and _in used a raw context_value == clause_value. As a result:

• A rule checking tier in [1] incorrectly matched a context where tier is True.
• A rule checking beta in [true] incorrectly matched a context where beta is 1.

The in operator now rejects a match when exactly one side is a bool, keeping boolean and numeric JSON values distinct while still matching True == True and False == False. This mirrors the bool-exclusion idea already used by is_number in value_parsing.py.

This is a Python-specific defect. The Ruby SDK (true == 1 is false) and js-core (true === 1 is false via strict equality) are unaffected.

Fixes #432

---

Note

Medium Risk
Changes core flag targeting semantics for in clauses; rules that accidentally relied on bool/number equality will behave differently, but the fix aligns with JSON types and other SDKs.

Overview
Fixes incorrect flag rule matches in the Python SDK where in treated booleans and numbers as equal because True == 1 and False == 0.

_in now returns false when exactly one of the context and clause values is a bool, so JSON booleans and numbers stay distinct while same-type equality still works. Parametrized tests cover bool/bool matches and bool vs int/float non-matches.

^{Reviewed by Cursor Bugbot for commit ea94700. Bugbot is set up for automated code reviews on this repo. Configure here.}