Integer flag values lack 64-bit support

Flag values above the 32-bit integer range can't be evaluated as integers — they error or fall back to [doubles](https://en.wikipedia.org/wiki/Double-precision_floating-point_format), losing precision. Unblocked once a [Kotlin SDK release](https://github.com/open-feature/kotlin-sdk/releases) ships [64-bit integer support](https://github.com/open-feature/kotlin-sdk/pull/225) (currently unreleased, one commit past `v0.8.0`).

## Acceptance criteria

- Whole-number flags beyond the 32-bit range evaluate as integers, with exact precision.
- Integers within range evaluate as they do today.

Check the Kotlin SDK release and linked 64-bit integer support change first, then locate the provider's integer flag evaluation path. Verify that whole-number values beyond 32-bit range retain exact precision while existing in-range values behave unchanged.

Written by the indexing model from the issue text.