fix: address PR #4 review feedback

Author: LeaderOnePro

src/main/java/com/steve/ai/llm/GroqClient.java

@@ -53,8 +53,8 @@ public String sendRequest(String systemPrompt, String userPrompt) {
         messages.add(userMessage);
 
         requestBody.add("messages", messages);
-        requestBody.addProperty("max_tokens", 500); // Keep it short for speed
-        requestBody.addProperty("temperature", 0.7);
+        requestBody.addProperty("max_tokens", SteveConfig.MAX_TOKENS.get());
+        requestBody.addProperty("temperature", SteveConfig.TEMPERATURE.get());
 
         HttpRequest request = HttpRequest.newBuilder()
             .uri(URI.create(GROQ_API_URL))

src/main/java/com/steve/ai/llm/TaskPlanner.java

@@ -70,7 +70,12 @@ public TaskPlanner() {
                 maxTokens, 
                 temperature
             );
-            AsyncLLMClient baseDeepSeek = new AsyncDeepSeekClient();
+            AsyncLLMClient baseDeepSeek = new AsyncDeepSeekClient(
+                SteveConfig.DEEPSEEK_API_KEY.get(),
+                SteveConfig.DEEPSEEK_MODEL.get(),
+                maxTokens,
+                temperature
+            );
 
             // Wrap with resilience patterns (caching, retries, circuit breaker)
             tempAsyncOpenAI = new ResilientLLMClient(baseOpenAI, tempCache, tempFallback);
@@ -171,6 +176,8 @@ public CompletableFuture<ResponseParser.ParsedResponse> planTasksAsync(SteveEnti
             // Build params map with provider-specific model
             String modelForProvider = switch (provider) {
                 case "deepseek" -> SteveConfig.DEEPSEEK_MODEL.get();
+                case "groq" -> SteveConfig.GROQ_MODEL.get();
+                case "gemini" -> SteveConfig.GEMINI_MODEL.get();
                 default -> SteveConfig.OPENAI_MODEL.get();
             };