moo.core.tests.test_security_attribute_coverage

Sentinel tests: every public attribute on the sandbox-exposed models must be classified for read-ACL enforcement.

The sandbox read checks in moo/core/code.py derive field and reverse-relation names from Model._meta automatically; methods must be declared by hand in SANDBOX_READ_METHODS (read-checked) or SANDBOX_EXEMPT_METHODS (does its own ACL check). These tests are the CI tripwire: adding a public field or method to Object, Verb, or Property without classifying it fails here, instead of silently skipping the ACL check at runtime.

Functions

test_buckets_are_pairwise_disjoint(kind, ...)
test_declared_method_names_exist_on_models(...)
test_derived_fields_include_expected_accessors()
test_model_public_attributes_are_classified(...)