fix: prevent hybrid mode from leaking OpenRouter model to local llamacpp client

When backend=hybrid with LLM_BACKEND=llamacpp, the user-selected model
(an OpenRouter id like "google/gemini-3-flash-preview") was being applied
to the local LlamaCppClient's primary_model and vision_model. This caused
describe_image to send the OpenRouter model name to llama-swap, which
returned 400 because it has no such slot.

Guard the local-client model override with !is_hybrid so it only applies
in local-only mode (where the user is selecting a different local model).
Bump to v1.2.0.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

src/ai/llamacpp.rs

@@ -409,10 +409,7 @@ impl LlmClient for LlamaCppClient {
             tools.len()
         );
         let mut body = serde_json::Map::new();
-        body.insert(
-            "model".into(),
-            Value::String(self.primary_model.clone()),
-        );
+        body.insert("model".into(), Value::String(self.primary_model.clone()));
         body.insert(
             "messages".into(),
             Value::Array(Self::messages_to_openai(&messages)),
@@ -1071,6 +1068,28 @@ mod tests {
         assert_eq!(local.embedding_model, "embed");
     }
 
+    #[test]
+    fn hybrid_mode_local_client_preserves_vision_model() {
+        // In hybrid mode, overrides.model is an OpenRouter model id
+        // (e.g. "google/gemini-3-flash-preview"). The local llamacpp
+        // client must NOT adopt that as its vision_model — it should
+        // keep the configured LLAMA_SWAP_VISION_MODEL so describe_image
+        // hits the correct local slot instead of sending an unknown
+        // model name to llama-swap.
+        let mut base = LlamaCppClient::new(None, Some("chat".into()));
+        base.set_vision_model("vision".into());
+        base.set_embedding_model("embed".into());
+
+        // Simulate what resolve_backend SHOULD do for hybrid mode:
+        // clone but do NOT override primary_model / vision_model.
+        let local = base.clone();
+
+        // The local client keeps its configured slots.
+        assert_eq!(local.primary_model, "chat");
+        assert_eq!(local.vision_model, "vision");
+        assert_eq!(local.embedding_model, "embed");
+    }
+
     #[test]
     fn assistant_tool_calls_emit_null_content() {
         let msg = ChatMessage {
@@ -1086,7 +1105,10 @@ mod tests {
             images: None,
         };
         let wire = LlamaCppClient::messages_to_openai(&[msg]);
-        assert!(wire[0]["content"].is_null(), "empty content + tool_calls should emit null");
+        assert!(
+            wire[0]["content"].is_null(),
+            "empty content + tool_calls should emit null"
+        );
     }
 
     #[test]