补齐 LlmRunRequest 字面量的 responses_input 字段

api-server LLM 代理与 platform-agent 适配器补 None
避免 cargo run api-server 因缺字段编不过

server-rs/crates/api-server/src/llm/mod.rs

@@ -114,6 +114,7 @@ pub async fn proxy_llm_chat_completions(
             .into_iter()
             .map(map_chat_message)
             .collect::<Vec<_>>(),
+        responses_input: None,
         max_output_tokens: None,
         enable_web_search: false,
         request_timeout_ms: None,

server-rs/crates/platform-agent/src/apimart_gpt5_adapter.rs

@@ -48,6 +48,7 @@ pub fn build_gpt5_multimodal_request(
             LlmMessage::new(LlmMessageRole::System, system_prompt.into()),
             LlmMessage::multimodal(LlmMessageRole::User, user_parts),
         ],
+        responses_input: None,
         max_output_tokens: None,
         request_timeout_ms: None,
         enable_web_search: false,