fix(telemetry): the human half of the pull ledger recorded one kind in three

Opening a snippet in the UI recorded rest_snippet. Opening a note or a task
recorded nothing — so the most direct evidence the product has that anyone
cares about a record existed for one kind out of three, and the other two sat
at zero pulls looking like dead weight beside a kind that merely had a counter.

Not an open question about intent: models/note_usage.py already documented
'rest_note' as a source value. Nothing wrote it. The design named it and the
implementation stopped at snippets.

Adds rest_note and rest_task. Tagged by SURFACE rather than by the record's
kind, matching rest_snippet — the kind is a join away, but which surface asked
is not recoverable after the fact. The mcp_/rest_ split stays load-bearing:
"is this dead weight" is served by any pull, "was that injected line useful" by
agent pulls alone, and a human clicking a link would inflate exactly the number
#1038 and #2085 gate on.

The vocabulary comment in the model was itself the stale-enumeration shape this
survey keeps finding — it named a source nothing wrote while omitting sources
that existed. Replaced with the naming CONVENTION plus a pointer to grep, which
cannot drift, rather than a longer list that would go stale the same way.

Guard extended to the REST surface, same derivation as the MCP half: a route
registered at exactly /<int:x> for GET is a detail view, and one reaching a
note-backed loader must record. Handler source is expanded one level through
module-private helpers, without which get_snippet_route — the route that
already got this right — would drop out of the check by loading via
_load_snippet. Verified the guard fires when a call is removed.

Renamed test_mcp_pull_telemetry.py -> test_pull_telemetry.py; it is no longer
only about MCP.

Closes #2476

tests/test_pull_telemetry.py

@@ -1,4 +1,8 @@
-"""Every getter that opens ONE note-backed record must record the pull.
+"""Every surface that opens ONE note-backed record must record the pull.
+
+Covers both halves: the MCP getters an agent calls, and the REST detail views a
+human opens. They are one ledger with two prefixes (`mcp_*` / `rest_*`), and a
+gap on either side makes the same records look untouched.
 
 WHY THIS EXISTS
 
@@ -14,7 +18,9 @@ That has now happened twice:
   #2476  `get_process` recorded nothing — and the auto-inject menu header names
          `get_process` as the way to open that kind. Processes were embedded
          when #2245 was fixed; the fix enumerated the kinds someone thought of
-         rather than the kinds that exist.
+         rather than the kinds that exist. The same issue found the REST half:
+         `rest_snippet` was recorded, note and task detail were not, and the
+         model's own comment named a `'rest_note'` source nothing wrote.
 
 A missing call is the shape no behavioural test catches: it changes no return
 value (#2278, shape 4). Source inspection is the only thing that sees it.
@@ -55,7 +61,9 @@ SINGLE_NOTE_LOADERS = (
     "get_snippet",
 )
 
-TOOLS_DIR = pathlib.Path(__file__).resolve().parents[1] / "src" / "scribe" / "mcp" / "tools"
+_SRC = pathlib.Path(__file__).resolve().parents[1] / "src" / "scribe"
+TOOLS_DIR = _SRC / "mcp" / "tools"
+ROUTES_DIR = _SRC / "routes"
 
 
 def _getters():
@@ -88,6 +96,88 @@ def test_every_single_record_getter_records_a_pull():
     )
 
 
+def _detail_routes():
+    """(module, handler, expanded source) for every bare-id GET route.
+
+    A route registered at exactly `/<int:x>` for GET is the DETAIL view of one
+    record — that shape is what distinguishes it from a list, a sub-resource
+    (`/<int:x>/versions`) or a write. Nothing else about the handler has to be
+    guessed.
+
+    The source is expanded one level through module-private helpers, because
+    `get_snippet_route` loads via `_load_snippet` rather than calling the loader
+    itself. Without the expansion the snippet route — the one that already got
+    this right — would drop out of the check.
+    """
+    import re
+    bare_id = re.compile(r"^/<int:\w+>$")
+    for path in sorted(ROUTES_DIR.glob("*.py")):
+        source = path.read_text()
+        tree = ast.parse(source)
+        helpers = {
+            node.name: ast.get_source_segment(source, node) or ""
+            for node in tree.body
+            if isinstance(node, (ast.AsyncFunctionDef, ast.FunctionDef))
+            and node.name.startswith("_")
+        }
+        for node in tree.body:
+            if not isinstance(node, (ast.AsyncFunctionDef, ast.FunctionDef)):
+                continue
+            for dec in node.decorator_list:
+                if not isinstance(dec, ast.Call):
+                    continue
+                route = next((a.value for a in dec.args
+                              if isinstance(a, ast.Constant)), None)
+                if not isinstance(route, str) or not bare_id.match(route):
+                    continue
+                methods = [
+                    e.value for kw in dec.keywords if kw.arg == "methods"
+                    and isinstance(kw.value, ast.List)
+                    for e in kw.value.elts if isinstance(e, ast.Constant)
+                ]
+                if "GET" not in methods:
+                    continue
+                body = ast.get_source_segment(source, node) or ""
+                expanded = body + "".join(
+                    src for name, src in helpers.items() if name in body
+                )
+                yield path.name, node.name, expanded
+
+
+def test_every_rest_detail_view_records_a_pull():
+    """The human half of the same ledger.
+
+    `rest_snippet` was recorded; note and task detail recorded nothing, so the
+    UI's most direct evidence of interest — someone opened the record — existed
+    for one kind out of three. The model's own comment listed `'rest_note'` as
+    a source, which means the design intended it and the implementation stopped
+    at snippets.
+
+    Same derivation as the MCP test above, over the other surface: the routes
+    are discovered, and the ones that must record are identified by the loader
+    they reach. A `/<int:x>` GET added for a fourth note-backed kind is covered
+    the day it is written.
+    """
+    missing = []
+    checked = []
+    for module, handler, body in _detail_routes():
+        if not any(loader in body for loader in SINGLE_NOTE_LOADERS):
+            continue  # not note-backed — groups and projects land here
+        checked.append(f"{module}:{handler}")
+        if "record_pulled" not in body:
+            missing.append(f"{module}:{handler}")
+
+    assert checked, (
+        "found no note-backed detail routes; the route shape or the loader "
+        "names must have moved"
+    )
+    assert not missing, (
+        f"these detail views open a record without recording the pull: "
+        f"{missing}. Add record_pulled(user_id=…, note_id=…, source='rest_<kind>') "
+        f"before returning — see routes/snippets.py:get_snippet_route."
+    )
+
+
 def test_every_named_loader_still_exists():
     """Pins the hand-written list against a rename.