Sorting LiteLLM GitHub issues by difficulty to implement. Higher = harder.

{
The left issue spans cache-control injection, prompt-management integration, and model registration behavior, requiring broader tracing across shared proxy and provider flows. The right issue is a localized parity fix with focused filtering and regression tests.
}
https://github.com/berriai/litellm/issues/31887 3:1 https://github.com/berriai/litellm/issues/30371