feat: Add LiteLLM provider with automatic prompt caching support (#3380)

Signed-off-by: HikaruEgashira <account@egahika.dev>

crates/goose/src/providers/openrouter.rs

@@ -199,23 +199,20 @@ fn update_request_for_anthropic(original_payload: &Value) -> Value {
 }
 
 fn create_request_based_on_model(
-    model_config: &ModelConfig,
+    provider: &OpenRouterProvider,
     system: &str,
     messages: &[Message],
     tools: &[Tool],
 ) -> anyhow::Result<Value, Error> {
     let mut payload = create_request(
-        model_config,
+        &provider.model,
         system,
         messages,
         tools,
         &super::utils::ImageFormat::OpenAi,
     )?;
 
-    if model_config
-        .model_name
-        .starts_with(OPENROUTER_MODEL_PREFIX_ANTHROPIC)
-    {
+    if provider.supports_cache_control() {
         payload = update_request_for_anthropic(&payload);
     }
 
@@ -259,7 +256,7 @@ impl Provider for OpenRouterProvider {
         tools: &[Tool],
     ) -> Result<(Message, ProviderUsage), ProviderError> {
         // Create the base payload
-        let payload = create_request_based_on_model(&self.model, system, messages, tools)?;
+        let payload = create_request_based_on_model(self, system, messages, tools)?;
 
         // Make request
         let response = self.post(payload.clone()).await?;
@@ -365,4 +362,10 @@ impl Provider for OpenRouterProvider {
         models.sort();
         Ok(Some(models))
     }
+
+    fn supports_cache_control(&self) -> bool {
+        self.model
+            .model_name
+            .starts_with(OPENROUTER_MODEL_PREFIX_ANTHROPIC)
+    }
 }