diff --git a/confluence-mdx/bin/confluence_xhtml_to_markdown.py b/confluence-mdx/bin/confluence_xhtml_to_markdown.py
index 075196f94..4cb6277dc 100755
--- a/confluence-mdx/bin/confluence_xhtml_to_markdown.py
+++ b/confluence-mdx/bin/confluence_xhtml_to_markdown.py
@@ -2039,7 +2039,8 @@ def add_import(self, module_name, condition=True):
         else:
             self._imports[module_name] = False
 
-    def load_attachments(self, input_dir: str, output_dir: str, public_dir: str) -> None:
+    def load_attachments(self, input_dir: str, output_dir: str, public_dir: str,
+                         skip_image_copy: bool = False) -> None:
         # Find all ac:image nodes first
         ac_image_nodes = self.soup.find_all('ac:image')
         attachments: List[Attachment] = []
@@ -2049,7 +2050,8 @@ def load_attachments(self, input_dir: str, output_dir: str, public_dir: str) ->
             for node in attachment_nodes:
                 logging.debug(f"add attachment of <ac:image>{node}")
                 attachment = Attachment(node, input_dir, output_dir, public_dir)
-                attachment.copy_to_destination()
+                if not skip_image_copy:
+                    attachment.copy_to_destination()
                 attachments.append(attachment)
 
         logging.debug(f"attachments: {attachments}")
@@ -2154,6 +2156,8 @@ def main():
                         help='/public directory path')
     parser.add_argument('--attachment-dir',
                         help='Directory to save attachments (default: output file directory)')
+    parser.add_argument('--skip-image-copy', action='store_true',
+                        help='이미지 파일 복사를 생략 (경로만 지정대로 생성)')
     parser.add_argument('--log-level',
                         choices=['debug', 'info', 'warning', 'error', 'critical'],
                         default='info',
@@ -2218,7 +2222,8 @@ def main():
         GLOBAL_LINK_MAPPING = build_link_mapping(page_v1)
 
         converter = ConfluenceToMarkdown(html_content)
-        converter.load_attachments(input_dir, output_dir, args.public_dir)
+        converter.load_attachments(input_dir, output_dir, args.public_dir,
+                                   skip_image_copy=args.skip_image_copy)
         markdown_content = converter.as_markdown()
 
         with open(args.output_file, 'w', encoding='utf-8') as f:
diff --git a/confluence-mdx/bin/reverse_sync_cli.py b/confluence-mdx/bin/reverse_sync_cli.py
index 0170c637d..2c52bf02f 100644
--- a/confluence-mdx/bin/reverse_sync_cli.py
+++ b/confluence-mdx/bin/reverse_sync_cli.py
@@ -87,6 +87,15 @@ def _resolve_page_id(ko_mdx_path: str) -> str:
     raise ValueError(f"MDX path '{ko_mdx_path}' not found in var/pages.yaml")
 
 
+def _resolve_attachment_dir(page_id: str) -> str:
+    """page_id에서 pages.yaml의 path를 조회하여 attachment-dir를 반환."""
+    pages = yaml.safe_load(Path('var/pages.yaml').read_text())
+    for page in pages:
+        if page['page_id'] == page_id:
+            return '/' + '/'.join(page['path'])
+    raise ValueError(f"page_id '{page_id}' not found in var/pages.yaml")
+
+
 def _forward_convert(patched_xhtml_path: str, output_mdx_path: str, page_id: str) -> str:
     """patched XHTML 파일을 forward converter로 MDX로 변환한다.
 
@@ -99,11 +108,13 @@ def _forward_convert(patched_xhtml_path: str, output_mdx_path: str, page_id: str
 
     abs_input = Path(patched_xhtml_path).resolve()
     abs_output = Path(output_mdx_path).resolve()
+    attachment_dir = _resolve_attachment_dir(page_id)
     result = subprocess.run(
         [sys.executable, str(converter), '--log-level', 'warning',
          str(abs_input), str(abs_output),
          '--public-dir', str(var_dir.parent),
-         '--attachment-dir', f'/{page_id}/verify'],
+         '--attachment-dir', attachment_dir,
+         '--skip-image-copy'],
         capture_output=True, text=True,
     )
     if result.returncode != 0: