diff --git a/apps/api/src/import-export/classified-markdown.test.ts b/apps/api/src/import-export/classified-markdown.test.ts
new file mode 100644
index 0000000..6612fbf
--- /dev/null
+++ b/apps/api/src/import-export/classified-markdown.test.ts
@@ -0,0 +1,39 @@
+import { describe, expect, it } from 'vitest';
+
+import { markClassifiedMarkdown, parseClassifiedMarkdown } from './classified-markdown';
+
+const MARKING = 'VS – NUR FÜR DEN DIENSTGEBRAUCH';
+
+describe('classified markdown marking (issue #210)', () => {
+  it('wraps a classified page in frontmatter and top+bottom imprint', () => {
+    const marked = markClassifiedMarkdown('# Title\n\nBody.\n', 'vs_nfd');
+    expect(marked).toBe(
+      `---\nclassification: vs_nfd\n---\n\n${MARKING}\n\n# Title\n\nBody.\n\n${MARKING}\n`,
+    );
+  });
+
+  it('leaves unclassified markdown untouched', () => {
+    expect(markClassifiedMarkdown('# Title\n\nBody.\n', 'unclassified')).toBe('# Title\n\nBody.\n');
+  });
+
+  it('parse is the inverse of mark', () => {
+    const original = '# Title\n\nBody.\n';
+    const { markdown, classification } = parseClassifiedMarkdown(
+      markClassifiedMarkdown(original, 'vs_nfd'),
+    );
+    expect(classification).toBe('vs_nfd');
+    expect(markdown).toBe(original);
+  });
+
+  it('passes documents without our frontmatter through unchanged', () => {
+    for (const raw of [
+      '# Plain\n\nNo frontmatter.\n',
+      '---\ntitle: Foreign frontmatter\ntags: [a]\n---\n\n# Doc\n',
+      `${MARKING}\n\nJust an imprint line without frontmatter.\n`,
+    ]) {
+      const { markdown, classification } = parseClassifiedMarkdown(raw);
+      expect(classification).toBeNull();
+      expect(markdown).toBe(raw);
+    }
+  });
+});
diff --git a/apps/api/src/import-export/classified-markdown.ts b/apps/api/src/import-export/classified-markdown.ts
new file mode 100644
index 0000000..9db6774
--- /dev/null
+++ b/apps/api/src/import-export/classified-markdown.ts
@@ -0,0 +1,43 @@
+import { PageClassification, classificationMarking } from '@dorfteich/shared';
+
+/**
+ * VS-NfD marking of exported Markdown (issue #210, ADR 0022): a classified
+ * page's `.md` carries the level machine-readably in YAML frontmatter AND
+ * human-visibly as the marking line at the top and bottom of the file.
+ * Unclassified pages pass through untouched — no marking, no frontmatter.
+ */
+export function markClassifiedMarkdown(
+  markdown: string,
+  classification: PageClassification,
+): string {
+  const marking = classificationMarking(classification);
+  if (!marking) return markdown;
+  return `---\nclassification: ${classification}\n---\n\n${marking}\n\n${markdown.trimEnd()}\n\n${marking}\n`;
+}
+
+/**
+ * Inverse of {@link markClassifiedMarkdown} for the import side: recognizes
+ * exactly the frontmatter block we generate (a lone `classification:` key)
+ * and the marking lines around the body, so a round-trip re-import yields
+ * the original content — and the page starts at the imported level (content
+ * must not escape its marking by traveling through a ZIP). Anything else —
+ * foreign frontmatter, hand-written documents — passes through unchanged.
+ */
+export function parseClassifiedMarkdown(raw: string): {
+  markdown: string;
+  classification: PageClassification | null;
+} {
+  const match = raw.match(/^---\nclassification: (vs_nfd|unclassified)\n---\n\n/);
+  if (!match) return { markdown: raw, classification: null };
+  const classification = match[1] as PageClassification;
+  let body = raw.slice(match[0].length);
+  const marking = classificationMarking(classification);
+  if (marking) {
+    if (body.startsWith(`${marking}\n\n`)) body = body.slice(marking.length + 2);
+    const trimmed = body.trimEnd();
+    if (trimmed.endsWith(`\n\n${marking}`)) {
+      body = `${trimmed.slice(0, -(marking.length + 2)).trimEnd()}\n`;
+    }
+  }
+  return { markdown: body, classification };
+}
diff --git a/apps/api/src/import-export/export.service.db.test.ts b/apps/api/src/import-export/export.service.db.test.ts
index c281d1c..c481bae 100644
--- a/apps/api/src/import-export/export.service.db.test.ts
+++ b/apps/api/src/import-export/export.service.db.test.ts
@@ -193,6 +193,76 @@ describe.skipIf(!hasTestDb)('export (e2e, issue #65)', () => {
     expect(target).toContain(`![dot](media/${image.id}.png)`);
   });
 
+  it('marks classified pages in the pond ZIP with frontmatter+imprint, ships a manifest, and round-trips (#210)', async () => {
+    const marking = 'VS – NUR FÜR DEN DIENSTGEBRAUCH';
+    const classifiedSlug = await seedPage(
+      personalPondId,
+      'Zip Classified',
+      '# Zip Classified\n\nclassified body text',
+    );
+    const openSlug = await seedPage(personalPondId, 'Zip Open', '# Zip Open\n\nopen body text');
+    await prisma.page.updateMany({
+      where: { pondId: personalPondId, slug: classifiedSlug },
+      data: { classification: 'VS_NFD' },
+    });
+
+    const res = await api()
+      .get(`/api/v1/ponds/${personalPondId}/export/markdown`)
+      .set('Cookie', ownerCookie)
+      .buffer(true)
+      .parse((r, cb) => {
+        const chunks: Buffer[] = [];
+        r.on('data', (c: Buffer) => chunks.push(c));
+        r.on('end', () => cb(null, Buffer.concat(chunks)));
+      })
+      .expect(200);
+    const entries = zipEntries(res.body as Buffer);
+
+    // Machine-readable frontmatter AND the visible imprint, top and bottom.
+    const marked = Buffer.from(entries[`${classifiedSlug}.md`]!).toString('utf8');
+    expect(marked.startsWith(`---\nclassification: vs_nfd\n---\n\n${marking}\n\n`)).toBe(true);
+    expect(marked.trimEnd().endsWith(marking)).toBe(true);
+    // Unclassified files are unchanged: no frontmatter, no imprint.
+    const open = Buffer.from(entries[`${openSlug}.md`]!).toString('utf8');
+    expect(open).not.toContain('classification:');
+    expect(open).not.toContain(marking);
+
+    // The manifest lists every file with its level and states the highest once.
+    const manifest = JSON.parse(Buffer.from(entries['manifest.json']!).toString('utf8')) as {
+      classification: string;
+      files: { path: string; classification: string }[];
+    };
+    expect(manifest.classification).toBe('vs_nfd');
+    expect(manifest.files).toContainEqual({
+      path: `${classifiedSlug}.md`,
+      classification: 'vs_nfd',
+    });
+    expect(manifest.files).toContainEqual({
+      path: `${openSlug}.md`,
+      classification: 'unclassified',
+    });
+
+    // Round-trip: re-importing the marked file must not confuse the importer —
+    // the page starts at the imported level, the body carries neither the
+    // frontmatter nor the imprint lines.
+    const imported = await api()
+      .post(`/api/v1/ponds/${personalPondId}/import`)
+      .set('Cookie', ownerCookie)
+      .attach('file', Buffer.from(marked, 'utf8'), 'reimported-classified.md')
+      .expect(201);
+    expect(imported.body.status).toBe('succeeded');
+    const reimported = await prisma.page.findUniqueOrThrow({
+      where: { id: imported.body.resultPageId as string },
+    });
+    expect(reimported.classification).toBe('VS_NFD');
+    const cache = await prisma.pageContentCache.findUniqueOrThrow({
+      where: { pageId: reimported.id },
+    });
+    expect(cache.markdown).toContain('classified body text');
+    expect(cache.markdown).not.toContain(marking);
+    expect(cache.markdown).not.toContain('classification:');
+  });
+
   it('skips an attachment whose bytes are missing on disk instead of crashing', async () => {
     // An attachment row with no file (data drift): upload then remove the bytes.
     const image = await files.upload({ id: ownerId } as never, personalPondId, {
diff --git a/apps/api/src/import-export/export.service.ts b/apps/api/src/import-export/export.service.ts
index 129ba6f..94d4041 100644
--- a/apps/api/src/import-export/export.service.ts
+++ b/apps/api/src/import-export/export.service.ts
@@ -9,6 +9,8 @@ import {
   fontSlug,
   PageClassification,
   classificationMarking,
+  classificationRank,
+  highestClassification,
   pondSettingsSchema,
 } from '@dorfteich/shared';
 import { User } from '@prisma/client';
@@ -23,6 +25,7 @@ import { PluginFallbackRenderer } from '../plugins/plugin-fallback-renderer';
 import { PluginsService } from '../plugins/plugins.service';
 import { PrismaService } from '../prisma/prisma.service';
 
+import { markClassifiedMarkdown } from './classified-markdown';
 import { ConversionJobService } from './conversion-job.service';
 import {
   imageExtension,
@@ -133,15 +136,54 @@ export class ExportService {
       attachments.map((a) => [a.id, `${a.id}.${imageExtension(a.mimeType)}`]),
     );
 
+    // Media inherits the highest classification among the readable pages that
+    // reference it (fail-closed, ADR 0022 — a shared image is as classified
+    // as its most classified use).
+    const mediaClassification = new Map<string, PageClassification>();
     for (const page of readablePages) {
-      const markdown = markdownForZip(
-        page.contentCache?.markdown ?? '',
-        readableSlugs,
-        mediaNameById,
+      const level = page.classification.toLowerCase() as PageClassification;
+      for (const id of imageFileIds(page.contentCache?.markdown ?? '')) {
+        const current = mediaClassification.get(id) ?? 'unclassified';
+        if (classificationRank(level) > classificationRank(current)) {
+          mediaClassification.set(id, level);
+        }
+      }
+    }
+
+    const manifestFiles: { path: string; classification: PageClassification }[] = [];
+    for (const page of readablePages) {
+      const level = page.classification.toLowerCase() as PageClassification;
+      // A classified page's file carries the level in YAML frontmatter and
+      // the marking line at top and bottom (#210); unclassified files are
+      // byte-identical to the pre-#210 export.
+      const markdown = markClassifiedMarkdown(
+        markdownForZip(page.contentCache?.markdown ?? '', readableSlugs, mediaNameById),
+        level,
       );
       // Page slugs are unique within a pond, so `<slug>.md` never collides.
       archive.append(markdown, { name: `${prefix}${page.slug}.md` });
+      manifestFiles.push({ path: `${prefix}${page.slug}.md`, classification: level });
     }
+    for (const attachment of attachments) {
+      manifestFiles.push({
+        path: `${prefix}media/${mediaNameById.get(attachment.id)!}`,
+        classification: mediaClassification.get(attachment.id) ?? 'unclassified',
+      });
+    }
+    // The archive-level manifest (#210): every file with its level, and the
+    // highest level contained stated once — the bulk-egress channel stays
+    // machine-checkable even after the ZIP is unpacked and copied onward.
+    archive.append(
+      JSON.stringify(
+        {
+          classification: highestClassification(manifestFiles.map((f) => f.classification)),
+          files: manifestFiles,
+        },
+        null,
+        2,
+      ),
+      { name: `${prefix}manifest.json` },
+    );
     for (const attachment of attachments) {
       const stream = this.storage.createReadStream(pond.id, attachment.id);
       // Defence in depth: a file removed between the existence check and the
diff --git a/apps/api/src/import-export/import.service.ts b/apps/api/src/import-export/import.service.ts
index 48918a2..4b9135a 100644
--- a/apps/api/src/import-export/import.service.ts
+++ b/apps/api/src/import-export/import.service.ts
@@ -24,6 +24,7 @@ import { PagesService } from '../pages/pages.service';
 import { docToState, emptyPageState } from '../pages/yjs-content';
 import { PrismaService } from '../prisma/prisma.service';
 
+import { parseClassifiedMarkdown } from './classified-markdown';
 import { ImportProcessor } from './import.constants';
 import {
   ASSET_PLACEHOLDER_PREFIX,
@@ -544,12 +545,24 @@ export class ImportService implements ImportProcessor {
     // file ids); track what we create so a later failure can be rolled back.
     const storedFileIds: string[] = [];
     try {
-      const markdown = await this.storeEmbeddedImages(rawMarkdown, user, pondId, storedFileIds);
+      // Our own classified export wraps the content in frontmatter + marking
+      // lines (#210) — strip them and carry the level into the new page, so
+      // a round-trip neither duplicates the marking nor loses it.
+      const { markdown: unwrapped, classification } = parseClassifiedMarkdown(rawMarkdown);
+      const markdown = await this.storeEmbeddedImages(unwrapped, user, pondId, storedFileIds);
       const json = markdownToDoc(markdown).toJSON() as unknown as PmNode;
       const { title, doc } = this.splitTitle(json, sourceName);
       const state = docToState(Node.fromJSON(editorSchema, doc));
 
-      const page = await this.pages.createWithState(user, pondId, title, state);
+      const page = await this.pages.createWithState(
+        user,
+        pondId,
+        title,
+        state,
+        null,
+        undefined,
+        classification,
+      );
       await this.files.linkAttachmentsToPage(storedFileIds, page.id);
       return page;
     } catch (error) {
diff --git a/apps/api/src/pages/pages.service.ts b/apps/api/src/pages/pages.service.ts
index 6020346..2e77f6e 100644
--- a/apps/api/src/pages/pages.service.ts
+++ b/apps/api/src/pages/pages.service.ts
@@ -274,8 +274,9 @@ export class PagesService {
     state: Uint8Array<ArrayBuffer>,
     parentId: string | null = null,
     presetSlug?: string,
+    atLeastClassification: PageClassification | null = null,
   ): Promise<Page> {
-    return this.insertPage(user, pondId, title, state, parentId, presetSlug);
+    return this.insertPage(user, pondId, title, state, parentId, presetSlug, atLeastClassification);
   }
 
   /**
@@ -300,6 +301,7 @@ export class PagesService {
     state: Uint8Array<ArrayBuffer>,
     parentId: string | null = null,
     presetSlug?: string,
+    atLeastClassification: PageClassification | null = null,
   ): Promise<Page> {
     const pond = await this.prisma.pond.findFirst({ where: { id: pondId, deletedAt: null } });
     if (!pond) throw new NotFoundException();
@@ -316,11 +318,19 @@ export class PagesService {
     const sortKey = generateKeyBetween(last?.sortKey ?? null, null);
     const content = deriveContent(state);
     // A new page starts at the instance-wide default level (ADR 0022, #204),
-    // raised to its parent's level when that is higher (#205): a subpage of
-    // classified content must never begin unmarked.
+    // raised to its parent's level when that is higher (#205) — and to an
+    // imported document's own level (#210): content must not escape its
+    // marking by traveling through an export/import. A subpage of classified
+    // content must never begin unmarked.
     let classification: PageClassification = await this.settings.get(
       'classification.newPageDefault',
     );
+    if (
+      atLeastClassification &&
+      classificationRank(atLeastClassification) > classificationRank(classification)
+    ) {
+      classification = atLeastClassification;
+    }
     if (parentId) {
       const parent = await this.prisma.page.findUniqueOrThrow({
         where: { id: parentId },
diff --git a/docs/vs-nfd/20-massnahmenplan.md b/docs/vs-nfd/20-massnahmenplan.md
index e67bbbb..9b36183 100644
--- a/docs/vs-nfd/20-massnahmenplan.md
+++ b/docs/vs-nfd/20-massnahmenplan.md
@@ -56,7 +56,7 @@ _Meilenstein: `M26 — VS-NfD: classification metadata`_
   - [x] **Print-CSS** (`@media print`, Kopf/Fuß je Seite) — fehlt komplett · 1 AT · #207
   - [x] PDF via gotenberg (`pdf-html.ts` Header/Footer-Template) · 1 AT · #208
   - [x] DOCX/ODT via pandoc (Reference-Doc mit Kopf-/Fußzeile) · 2–3 AT · #209
-  - [ ] Markdown-ZIP (Frontmatter + Aufdruck) · 1 AT · #210
+  - [x] Markdown-ZIP (Frontmatter + Aufdruck) · 1 AT · #210
   - [ ] Atom-Feeds, Public-API, Suchergebnisse, No-JS-Shell · 2–3 AT · #211
   - [ ] Attachment-Download (Dateiname-Präfix + Begleitdatei) · 1–2 AT · #212
 - [ ] Warnung/Sperre beim Anhängen an eingestufte Seiten · 1 AT · #213