fix: force provider prefix for agent model to prevent incorrect fallback

Dockerfile

@@ -90,13 +90,13 @@ python3 /usr/local/bin/sync.py restore\n\
 # 清理 API Base 地址\n\
 CLEAN_BASE=\$(echo \"\$OPENAI_API_BASE\" | sed \"s|/chat/completions||g\" | sed \"s|/v1/|/v1|g\" | sed \"s|/v1\$|/v1|g\")\n\
 \n\
-# 处理 MODEL 变量：优先使用 FULL_MODEL_ID，保持灵活性\n\
+# 处理 MODEL 变量：确保 FULL_MODEL_ID 始终包含 Provider 前缀，防止 OpenClaw 错误回退
 if [[ \"\$MODEL\" == */* ]]; then\n\
   FULL_MODEL_ID=\"\$MODEL\"\n\
   PROVIDER_NAME=\"\${MODEL%%/*}\"\n\
   PROVIDER_MODEL_ID=\"\${MODEL#*/}\"\n\
 else\n\
-  FULL_MODEL_ID=\"\$MODEL\"\n\
+  FULL_MODEL_ID=\"siliconflow/\$MODEL\"\n\
   PROVIDER_NAME=\"siliconflow\"\n\
   PROVIDER_MODEL_ID=\"\$MODEL\"\n\
 fi\n\
@@ -129,7 +129,8 @@ cat > /root/.openclaw/openclaw.json <<EOF\n\
       \"domain\": \"\${FEISHU_DOMAIN:-feishu}\",\n\
       \"connectionMode\": \"\${FEISHU_CONNECTION_MODE:-websocket}\",\n\
       \"dmPolicy\": \"open\",\n\
-      \"ignoreEvents\": [\"im.message.message_read_v1\"]\n\
+      \"ignoreEvents\": [\"im.message.message_read_v1\"],\n\
+      \"ignoredEvents\": [\"im.message.message_read_v1\"]\n\
     }\n\
   }\n\
 }\n\