Instant Apply -> respect proxy's /models

[24anisha]

Enables dynamic model selection for instant apply operations through the ProxyModelsService, gated behind an experimental configuration flag.

[Copilot]

Pull request overview

This PR enables dynamic model selection for instant apply operations by integrating the ProxyModelsService, which fetches model configurations from a remote endpoint. The change is gated behind the experimental configuration flag UseProxyModelsServiceForInstantApply to allow gradual rollout and testing. When enabled, both instant apply endpoints will use models fetched from the proxy models service instead of the statically configured model names.

Key Changes

• Added IProxyModelsService dependency injection to both Proxy4oEndpoint and ProxyInstantApplyShortEndpoint constructors
• Implemented conditional model selection logic that checks the experimental flag and falls back to existing configuration-based model selection when disabled or when no models are available
• Introduced new configuration key UseProxyModelsServiceForInstantApply in the TeamInternal namespace to control the feature rollout

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
src/platform/configuration/common/configurationService.ts	Added experimental configuration flag UseProxyModelsServiceForInstantApply to gate the new dynamic model selection behavior
src/platform/endpoint/node/proxyInstantApplyShortEndpoint.ts	Integrated ProxyModelsService with conditional model selection logic, falling back to InstantApplyShortModelName config or SHORT_INSTANT_APPLY constant
src/platform/endpoint/node/proxy4oEndpoint.ts	Integrated ProxyModelsService with conditional model selection logic, falling back to InstantApplyModelName config or GPT4OPROXY constant