diff --git a/CHANGELOG.md b/CHANGELOG.md
index c82baccf9..3ea02861f 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -2,6 +2,10 @@
 
 Full release notes with details on each version: [GitHub Releases](https://github.com/safishamsi/graphify/releases)
 
+## Unreleased
+
+- Fix: `graphify query` now surfaces the work-memory lesson annotations. The renderer has carried `learning=<status>[:stale]` NODE-line support since the overlay shipped, and both `graphify explain` and the MCP server attach the `.graphify_learning.json` sidecar at load time — but the CLI `query` loader never did, so the same question answered over MCP showed the lessons while the CLI answer silently dropped them (the docs promise `query` surfaces a lesson hint). The CLI query path now attaches the overlay exactly like `serve._load_graph` (display-only, fail-safe to empty), so annotations — including the `:stale` "code changed since — re-verify" marker — render identically on both surfaces; without a sidecar the output is byte-identical to before.
+
 ## 0.9.50 (2026-08-25)
 
 - Fix: Ruby methods whose names end in `!`, `?`, or `=` now keep distinct node ids, so `save` and `save!` (or `foo` and `foo=`) no longer collide into one node; the label keeps the raw spelling and member-call resolution still matches (#3077, thanks @hopstreax).
diff --git a/graphify/cli.py b/graphify/cli.py
index bd2956815..02620a508 100644
--- a/graphify/cli.py
+++ b/graphify/cli.py
@@ -1162,6 +1162,17 @@ def dispatch_command(cmd: str) -> None:
         except Exception as exc:
             print(f"error: could not load graph: {exc}", file=sys.stderr)
             sys.exit(1)
+        # Work-memory overlay: attach the .graphify_learning.json sidecar the
+        # same way the MCP loader does (serve._load_graph), so the renderer's
+        # learning=<status>[:stale] NODE annotations appear in CLI `query`
+        # output too — `explain` and MCP already surface them, and the docs
+        # promise `query` does as well. Empty overlay on any error, leaving
+        # un-annotated output byte-identical.
+        try:
+            from graphify.reflect import load_learning_overlay as _llo
+            G.graph["_learning_overlay"] = _llo(gp)
+        except Exception:
+            G.graph["_learning_overlay"] = {}
         import time as _time
         _t0 = _time.perf_counter()
         _mode = "dfs" if use_dfs else "bfs"
diff --git a/tests/test_query_cli.py b/tests/test_query_cli.py
index 0db4e6fa8..5f429673a 100644
--- a/tests/test_query_cli.py
+++ b/tests/test_query_cli.py
@@ -123,3 +123,68 @@ def test_query_cli_rejects_oversized_graph(monkeypatch, tmp_path, capsys):
     err = capsys.readouterr().err
     assert "exceeds" in err
     assert "byte cap" in err
+
+
+# --- work-memory overlay: lesson annotations in CLI query output ---------------
+
+
+def _write_sidecar(tmp_path, nodes: dict) -> None:
+    """A .graphify_learning.json next to graph.json, in the writer's shape."""
+    (tmp_path / ".graphify_learning.json").write_text(
+        json.dumps({"version": 1, "generated_at": "2026-01-01T00:00:00+00:00",
+                    "nodes": nodes}),
+        encoding="utf-8",
+    )
+
+
+def test_query_cli_annotates_nodes_with_learning_status(monkeypatch, tmp_path, capsys):
+    """`graphify query` must merge the learning sidecar into NODE lines with the
+    same learning=<status> form the MCP server renders — the docs promise
+    lessons surface in `query` output, not only in `explain`/MCP."""
+    graph_path = _write_graph(tmp_path)
+    _write_sidecar(tmp_path, {"n1": {"status": "preferred"}})
+    monkeypatch.setattr(mainmod, "_check_skill_version", lambda _: None)
+    monkeypatch.setattr(
+        mainmod.sys,
+        "argv",
+        ["graphify", "query", "extract", "--graph", str(graph_path)],
+    )
+    mainmod.main()
+    out = capsys.readouterr().out
+    assert "learning=preferred]" in out
+    # Only the annotated node carries the suffix.
+    assert out.count("learning=") == 1
+
+
+def test_query_cli_marks_stale_learning_entries(monkeypatch, tmp_path, capsys):
+    """An entry whose cited code changed (or vanished) since the verdict must
+    carry the :stale marker, matching the MCP annotation form."""
+    graph_path = _write_graph(tmp_path)
+    _write_sidecar(
+        tmp_path,
+        {"n1": {"status": "tentative", "source_file": "gone.py",
+                "code_fingerprint": "0123abcd"}},
+    )
+    monkeypatch.setattr(mainmod, "_check_skill_version", lambda _: None)
+    monkeypatch.setattr(
+        mainmod.sys,
+        "argv",
+        ["graphify", "query", "extract", "--graph", str(graph_path)],
+    )
+    mainmod.main()
+    out = capsys.readouterr().out
+    assert "learning=tentative:stale]" in out
+
+
+def test_query_cli_without_sidecar_stays_unannotated(monkeypatch, tmp_path, capsys):
+    """No sidecar -> no learning= anywhere; un-annotated output is unchanged."""
+    graph_path = _write_graph(tmp_path)
+    monkeypatch.setattr(mainmod, "_check_skill_version", lambda _: None)
+    monkeypatch.setattr(
+        mainmod.sys,
+        "argv",
+        ["graphify", "query", "extract", "--graph", str(graph_path)],
+    )
+    mainmod.main()
+    out = capsys.readouterr().out
+    assert "learning=" not in out