From b5312183b8944f56012a0cc446b477c618afcda6 Mon Sep 17 00:00:00 2001
From: lintsinghua <129816813+lintsinghua@users.noreply.github.com>
Date: Thu, 23 Oct 2025 11:05:12 +0800
Subject: [PATCH] Update README_EN.md

---
 README_EN.md | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/README_EN.md b/README_EN.md
index cb1cfa6..1f6daa0 100644
--- a/README_EN.md
+++ b/README_EN.md
@@ -346,6 +346,15 @@ We warmly welcome all forms of contributions! Whether it's submitting issues, cr
 - **Project Link**: [https://github.com/lintsinghua/XCodeReviewer](https://github.com/lintsinghua/XCodeReviewer)
 - **Issue Reports**: [Issues](https://github.com/lintsinghua/XCodeReviewer/issues)
 - **Author Email**: tsinghuaiiilove@gmail.com
+ 
+## 🎯 Future Plans
+
+Currently, XCodeReviewer is positioned in the rapid prototype verification stage, and its functions need to be gradually improved. Based on the subsequent development of the project and everyone's suggestions, the future development plan is as follows (to be implemented as soon as possible):
+
+- **Multi-platform/Local Model Support**: In the future, we will quickly add API calling functions for major mainstream models at home and abroad, such as OpenAI, Claude, Tongyi Qianwen, etc.
+- **Multi-Agent Collaboration**: Consider introducing a multi-agent collaboration architecture, which will implement the `Agent + Human Dialogue` feedback function, including multi-round dialogue process display, human dialogue interruption intervention, etc., to obtain a clearer, more transparent, and supervised auditing process, thereby improving audit quality.
+- **Professional Report File Generation**: Generate professional audit report files in relevant formats according to different needs, supporting customization of file report formats, etc.
+- **Custom Audit Standards**: Different teams have their own coding standards, and different projects have specific security requirements, which is exactly what we want to do next in this project. The current version is still in a "semi-black box mode", where the project guides the analysis direction and defines audit standards through Prompt engineering, and the actual analysis effect is determined by the built-in knowledge of powerful pre-trained AI models. In the future, we will combine methods such as reinforcement learning and supervised learning fine-tuning to develop support for custom rule configuration, define team-specific rules through YAML or JSON, provide best practice templates for common frameworks, etc., to obtain audit results that are more in line with requirements and standards.
 
 ---