diff --git a/graphify/report.py b/graphify/report.py
index 673c2f2bb..0f1c5e709 100644
--- a/graphify/report.py
+++ b/graphify/report.py
@@ -137,6 +137,21 @@ def generate(
             f"- {detection_result['total_files']} files · ~{detection_result['total_words']:,} words",
             "- Verdict: corpus is large enough that graph structure adds value.",
         ]
+        # #3511: files detect() saw but could not classify (no supported
+        # extension/shebang) were counted nowhere -- a corpus that is mostly
+        # an unsupported language reported the same "well covered" verdict as
+        # one that was actually extracted. Surface the count and its biggest
+        # extensions so a near-total miss (e.g. a Lean/Zig/whatever repo with
+        # no matching extractor) is visible here instead of silent.
+        unclassified = detection_result.get("unclassified") or []
+        if unclassified:
+            from collections import Counter as _Counter
+            ext_counts = _Counter(Path(p).suffix or "(none)" for p in unclassified)
+            top = ", ".join(f"{ext} {n}" for ext, n in ext_counts.most_common(3))
+            lines.append(
+                f"- Unclassified: {len(unclassified)} file(s) not represented in "
+                f"the graph (top: {top})"
+            )
 
     from .analyze import _is_file_node as _ifn
 
diff --git a/tests/test_report.py b/tests/test_report.py
index 03bc67acb..98f256907 100644
--- a/tests/test_report.py
+++ b/tests/test_report.py
@@ -29,6 +29,29 @@ def test_report_contains_corpus_check():
     report = generate(G, communities, cohesion, labels, gods, surprises, detection, tokens, "./project")
     assert "## Corpus Check" in report
 
+
+def test_report_surfaces_unclassified_files():
+    """#3511: detect() already tracks files it saw but could not classify
+    (no supported extension), but nothing surfaced them -- a corpus that is
+    mostly an unsupported language got the same "well covered" verdict as
+    one that was actually extracted."""
+    G, communities, cohesion, labels, gods, surprises, detection, tokens = make_inputs()
+    detection = {
+        **detection,
+        "unclassified": ["Main.lean", "Util.lean", "a.toml", "b.toml", "c.toml", "readme"],
+    }
+    report = generate(G, communities, cohesion, labels, gods, surprises, detection, tokens, "./project")
+    assert "Unclassified: 6 file(s)" in report
+    assert ".lean 2" in report
+    assert ".toml 3" in report
+
+
+def test_report_omits_unclassified_line_when_none():
+    """Backward compatible: no unclassified files, no new line."""
+    G, communities, cohesion, labels, gods, surprises, detection, tokens = make_inputs()
+    report = generate(G, communities, cohesion, labels, gods, surprises, detection, tokens, "./project")
+    assert "Unclassified:" not in report
+
 def test_report_contains_god_nodes():
     G, communities, cohesion, labels, gods, surprises, detection, tokens = make_inputs()
     report = generate(G, communities, cohesion, labels, gods, surprises, detection, tokens, "./project")