Reject non-JSON-compatible values in canonical_json before hashing

Assisted-by: GitHub Copilot (model: claude-opus, autonomous)

Co-authored-by: nicolehaugen <10600161+nicolehaugen@users.noreply.github.com>

Author: Copilot

src/specify_cli/_identifier.py

@@ -136,6 +136,15 @@ def canonical_json(value: Any) -> bytes:
     are emitted verbatim. This is the byte string that the hook discriminator
     hashes and that the manifest loader uses to detect byte-identical duplicate
     hook entries.
+
+    Only JSON-compatible scalars (``None``, ``str``, ``bool``, ``int``,
+    ``float``), mappings, and lists/tuples are accepted. ``yaml.safe_load()``
+    can produce values outside that set — for example an unquoted date is
+    parsed as ``datetime.date`` — and passing one through here would otherwise
+    surface as a raw, uncaught ``TypeError`` from ``json.dumps()`` deep inside
+    this function. Callers get a clean ``IdentifierComponentError`` instead, so
+    a manifest with such a value can be rejected the same way as any other
+    invalid declared field.
     """
     normalized = _normalize_for_canonical_json(value)
     return json.dumps(
@@ -147,11 +156,16 @@ def canonical_json(value: Any) -> bytes:
 
 
 def _normalize_for_canonical_json(value: Any) -> Any:
+    if value is None or isinstance(value, (str, bool, int, float)):
+        return value
     if isinstance(value, Mapping):
         return {str(k): _normalize_for_canonical_json(v) for k, v in value.items()}
     if isinstance(value, (list, tuple)):
         return [_normalize_for_canonical_json(v) for v in value]
-    return value
+    raise IdentifierComponentError(
+        f"Value of type {type(value).__name__!r} is not JSON-compatible and "
+        "cannot be canonicalized"
+    )
 
 
 def _has_hook_sibling_collision(

tests/test_contribution_ids.py

@@ -173,6 +173,18 @@ def test_preserves_list_order(self):
     def test_utf8_no_ensure_ascii(self):
         assert canonical_json({"k": "café"}).decode("utf-8") == '{"k":"café"}'
 
+    def test_rejects_non_json_scalar(self):
+        with pytest.raises(IdentifierComponentError):
+            canonical_json(datetime.date(2026, 1, 1))
+
+    def test_rejects_non_json_scalar_nested_in_mapping(self):
+        with pytest.raises(IdentifierComponentError):
+            canonical_json({"prompt": datetime.date(2026, 1, 1)})
+
+    def test_rejects_non_json_scalar_nested_in_list(self):
+        with pytest.raises(IdentifierComponentError):
+            canonical_json([datetime.date(2026, 1, 1)])
+
 
 # ---------------------------------------------------------------------------
 # Hook discriminator behaviour