diff --git a/CLAUDE.md b/CLAUDE.md
index 0bd165c..5e1bde3 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -128,16 +128,21 @@ Models are configured via environment variables in `vars.sh`:
 # MODEL_1 - Hermes (always available, default)
 export MODEL_ENDPOINT_1=http://localhost:8080/v1
 export MODEL_API_KEY_1=your-api-key
+export MODEL_NAME_1=model  # Optional: actual model name for the endpoint
 
 # MODEL_2 - Additional model (optional)
 export MODEL_ENDPOINT_2=http://localhost:8081/v1
 export MODEL_API_KEY_2=your-api-key
+export MODEL_NAME_2=gpt-4  # Optional: specify deployment/model name
 
 # MODEL_3 - Qwen3-Coder (recommended for programming)
 export MODEL_ENDPOINT_3=http://localhost:8082/v1
 export MODEL_API_KEY_3=your-api-key
+export MODEL_NAME_3=model  # Optional: defaults to "model" if not specified
 ```
 
+**Note**: `MODEL_NAME_{n}` is optional and defaults to `"model"`. Some endpoints (like Azure OpenAI) require the actual deployment name - set this variable for those cases.
+
 **Example: Programming Activity**
 
 ```yaml
diff --git a/activity_yaml_validator.py b/activity_yaml_validator.py
index d760143..c7c50e5 100644
--- a/activity_yaml_validator.py
+++ b/activity_yaml_validator.py
@@ -687,10 +687,10 @@ class ActivityYAMLValidator:
 
                 for bucket, transition in step["transitions"].items():
                     if "metadata_feedback_filter" in transition:
-                        # Check if step has feedback_tokens_for_ai
-                        if "feedback_tokens_for_ai" not in step:
+                        # Check if step has feedback_tokens_for_ai or feedback_prompts
+                        if "feedback_tokens_for_ai" not in step and "feedback_prompts" not in step:
                             self.warnings.append(
-                                f"Section {section_id}, step {step_id}: metadata_feedback_filter used but no feedback_tokens_for_ai defined"
+                                f"Section {section_id}, step {step_id}: metadata_feedback_filter used but no feedback_tokens_for_ai or feedback_prompts defined"
                             )
 
     def _validate_pre_scripts(self, data: Dict[str, Any]):
diff --git a/research/guarded_ai.py b/research/guarded_ai.py
index 228ea1c..1b5bfca 100644
--- a/research/guarded_ai.py
+++ b/research/guarded_ai.py
@@ -48,14 +48,15 @@ def get_openai_client_and_model(model_name=None):
             model_num = model_name.split("_")[1]
             endpoint_key = f"MODEL_ENDPOINT_{model_num}"
             api_key_key = f"MODEL_API_KEY_{model_num}"
+            model_name_key = f"MODEL_NAME_{model_num}"
 
             endpoint = os.getenv(endpoint_key)
             api_key = os.getenv(api_key_key)
 
             if endpoint and api_key:
                 client = get_client_for_endpoint(endpoint, api_key)
-                # Use a simple default model name for the endpoint
-                actual_model = "model"  # Most endpoints use "model" or ignore this
+                # Get the actual model name from environment, or use sensible default
+                actual_model = os.getenv(model_name_key) or "model"
                 return client, actual_model
         except Exception as e:
             print(f"Warning: Failed to load {model_name}: {e}, falling back to default")