diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index aec64b0..fae8954 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -121,6 +121,9 @@ jobs:
             "lighthouse@${LIGHTHOUSE_VERSION}"
           npx --no-install playwright install --with-deps chromium
 
+      - name: Validate embed source pages
+        run: python scripts/check_embed_sources.py
+
       - name: Build site and capture Hugo duration
         id: hugo-build
         run: |
diff --git a/scripts/check_embed_sources.py b/scripts/check_embed_sources.py
new file mode 100644
index 0000000..6343d3a
--- /dev/null
+++ b/scripts/check_embed_sources.py
@@ -0,0 +1,75 @@
+#!/usr/bin/env python3
+"""Validate that every embed page's source_page resolves to an existing data page.
+
+A `content/embeds/*` page renders a chart from another page referenced by its
+`source_page` front matter. If that target page is missing, the Hugo build aborts
+(`layouts/embeds/single.html` and `layouts/shortcodes/observatory-chart.html`
+call `errorf`). This check catches a dangling reference before the build, so the
+failure surfaces in CI rather than in the production deploy (see issue #627).
+"""
+
+from __future__ import annotations
+
+import argparse
+import re
+import sys
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parent.parent
+
+SOURCE_PAGE_PATTERN = re.compile(
+    r"""^\s*source_page\s*[:=]\s*["']?(?P<value>[^"'\n]+?)["']?\s*$""",
+    re.MULTILINE,
+)
+
+
+def _source_page(document: str) -> str | None:
+    match = SOURCE_PAGE_PATTERN.search(document)
+    return match.group("value").strip() if match else None
+
+
+def _resolves(content_root: Path, source_page: str) -> bool:
+    relative = source_page.strip("/")
+    # NFR-012: an embed's source_page must be a data page under content/data/.
+    if relative != "data" and not relative.startswith("data/"):
+        return False
+    target = content_root / relative
+    return (target / "index.md").is_file() or (target / "_index.md").is_file()
+
+
+def check_embed_sources(root: Path = ROOT) -> list[str]:
+    """Return a list of human-readable problems, empty when all embeds resolve."""
+    content_root = root / "content"
+    problems: list[str] = []
+
+    for embed in sorted(content_root.glob("embeds/*/index.md")):
+        source_page = _source_page(embed.read_text(encoding="utf-8"))
+        relative_embed = embed.relative_to(root).as_posix()
+        if not source_page:
+            problems.append(f"{relative_embed}: missing source_page")
+            continue
+        if not _resolves(content_root, source_page):
+            problems.append(
+                f"{relative_embed}: source_page does not resolve to a content/data page: "
+                f"{source_page}"
+            )
+
+    return problems
+
+
+def main(argv: list[str] | None = None) -> int:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--root", default=ROOT, type=Path, help="Repository root")
+    args = parser.parse_args(argv)
+
+    problems = check_embed_sources(Path(args.root))
+    for problem in problems:
+        print(f"::error::{problem}", file=sys.stderr)
+    if problems:
+        return 1
+    print("All embed source_page references resolve.")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
diff --git a/tests/test_embed_sources.py b/tests/test_embed_sources.py
new file mode 100644
index 0000000..95fa0ea
--- /dev/null
+++ b/tests/test_embed_sources.py
@@ -0,0 +1,60 @@
+from __future__ import annotations
+
+import tempfile
+from pathlib import Path
+
+from scripts.check_embed_sources import check_embed_sources
+
+ROOT = Path(__file__).resolve().parent.parent
+
+
+def test_repository_embed_sources_all_resolve() -> None:
+    assert check_embed_sources(ROOT) == []
+
+
+def _write(path: Path, text: str) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    path.write_text(text, encoding="utf-8")
+
+
+def test_missing_source_page_is_reported() -> None:
+    with tempfile.TemporaryDirectory() as tmp:
+        root = Path(tmp)
+        _write(root / "content/embeds/chart/index.md", '+++\ntitle = "Chart"\n+++\n')
+        problems = check_embed_sources(root)
+        assert any("missing source_page" in problem for problem in problems)
+
+
+def test_dangling_source_page_is_reported() -> None:
+    with tempfile.TemporaryDirectory() as tmp:
+        root = Path(tmp)
+        _write(
+            root / "content/embeds/chart/index.md",
+            '+++\nsource_page = "/data/does-not-exist/"\n+++\n',
+        )
+        problems = check_embed_sources(root)
+        assert any("does not resolve" in problem for problem in problems)
+
+
+def test_resolving_source_page_passes() -> None:
+    with tempfile.TemporaryDirectory() as tmp:
+        root = Path(tmp)
+        _write(
+            root / "content/embeds/chart/index.md",
+            '+++\nsource_page = "/data/live/"\n+++\n',
+        )
+        _write(root / "content/data/live/index.md", '+++\ntitle = "Live"\n+++\n')
+        assert check_embed_sources(root) == []
+
+
+def test_non_data_source_page_is_reported() -> None:
+    with tempfile.TemporaryDirectory() as tmp:
+        root = Path(tmp)
+        _write(
+            root / "content/embeds/chart/index.md",
+            '+++\nsource_page = "/weekly/2026/w22/"\n+++\n',
+        )
+        # The referenced page exists, but it is not a content/data page.
+        _write(root / "content/weekly/2026/w22/index.md", '+++\ntitle = "W22"\n+++\n')
+        problems = check_embed_sources(root)
+        assert any("does not resolve to a content/data page" in problem for problem in problems)