diff --git a/configs/skills/direct_agent.yaml b/configs/skills/direct_agent.yaml
index 4aa6339..ca5de64 100644
--- a/configs/skills/direct_agent.yaml
+++ b/configs/skills/direct_agent.yaml
@@ -24,7 +24,7 @@ prompt:
   instructions: "根据用户需求，直接给出简洁准确的回答。"
 
 llm:
-  model: "openai/gpt-4o-mini"
+  model: "default"
   temperature: 0.3
   max_tokens: 1024
 
diff --git a/configs/skills/plan_exec_agent.yaml b/configs/skills/plan_exec_agent.yaml
index 1306a4d..e1726cd 100644
--- a/configs/skills/plan_exec_agent.yaml
+++ b/configs/skills/plan_exec_agent.yaml
@@ -25,7 +25,7 @@ prompt:
   instructions: "根据用户需求，制定详细的执行计划，逐步执行每个步骤，必要时调整计划。"
 
 llm:
-  model: "anthropic/claude-opus-4-20250514"
+  model: "default"
   temperature: 0.0
   max_tokens: 8192
 
diff --git a/configs/skills/react_agent.yaml b/configs/skills/react_agent.yaml
index 9a5ae7f..c74293e 100644
--- a/configs/skills/react_agent.yaml
+++ b/configs/skills/react_agent.yaml
@@ -25,7 +25,7 @@ prompt:
   instructions: "根据用户需求，动态选择合适的工具和策略完成任务。每一步都要观察中间结果并调整策略。"
 
 llm:
-  model: "anthropic/claude-sonnet-4-20250514"
+  model: "default"
   temperature: 0.1
   max_tokens: 4096
 
diff --git a/configs/skills/reflexion_agent.yaml b/configs/skills/reflexion_agent.yaml
index 30a3080..be7207b 100644
--- a/configs/skills/reflexion_agent.yaml
+++ b/configs/skills/reflexion_agent.yaml
@@ -25,7 +25,7 @@ prompt:
   instructions: "根据用户需求完成任务，完成后自我评估输出质量，如不达标则反思改进并重试。"
 
 llm:
-  model: "anthropic/claude-sonnet-4-20250514"
+  model: "default"
   temperature: 0.0
   max_tokens: 4096
 
diff --git a/configs/skills/rewoo_agent.yaml b/configs/skills/rewoo_agent.yaml
index f081da6..874ffd5 100644
--- a/configs/skills/rewoo_agent.yaml
+++ b/configs/skills/rewoo_agent.yaml
@@ -25,7 +25,7 @@ prompt:
   instructions: "根据用户需求，规划所有需要的数据采集步骤，然后批量执行。"
 
 llm:
-  model: "anthropic/claude-sonnet-4-20250514"
+  model: "default"
   temperature: 0.1
   max_tokens: 4096