Remove redundant tests from Task SDK

[shahar1]

Human Summary

related: #68502.
This PR removes tests that are either stdlib/3rd-party's concern, trivial attribute round-trip, or duplicated tests.
Summarizing table is in the "AI Summary".

AI Summary

Click Here

Legend — patterns:

• P1 Stdlib / third-party behavior tested directly (trust the library)
• P2 Trivial attribute round-trip (no logic, just obj.field == value_just_passed_in)
• P5 Exact or near-exact duplicate function

Test	File : line	Pattern	Why redundant
test_build_task_group_with_prefix_functionality	test_taskgroup.py:275	P5	Near-exact duplicate of test_build_task_group_with_prefix: same Dag structure, same 7 assertions, only the Dag name string and inline comments differ
test_mask_secret_with_iterable	test_comms.py:68	P5	Literal duplicate of test_mask_secret_with_list: identical setup example_dict = ["test"] and identical assertion — claims to test an iterable but uses the same list
test_mask_secret_with_list	test_comms.py:63	P5	Fully subsumed by the parametrized test_mask_secret_with_objects, which already covers a list input alongside dict and string cases
test_node_not_found_caught_as_key_error	test_exceptions.py:32	P1	Redundant with test_node_not_found_is_subclass_of_key_error; if the class hierarchy is correct the catch is guaranteed by Python — tests except mechanics, not our code
test_node_not_found_caught_as_task_not_found	test_exceptions.py:37	P1	Redundant with test_node_not_found_is_subclass_of_task_not_found; tests Python's exception dispatch, not the exception class definition
test_eq (WaitForAll)	test_wait_policy.py:28	P1	WaitForAll is a plain attrs class with no custom __eq__; WaitForAll() == WaitForAll() tests attrs-generated equality, not project logic
test_hash_consistent (WaitForAll)	test_wait_policy.py:34	P1	hash(WaitForAll()) == hash(WaitForAll()) tests attrs-generated __hash__, not project logic
test_repr (WaitForAll)	test_wait_policy.py:23	P1	repr(WaitForAll()) == "WaitForAll()" tests attrs-generated __repr__ with no custom repr logic
test_stores_n (MinimumCount)	test_wait_policy.py:39	P2	MinimumCount(5).n == 5 — pure attribute round-trip on an attrs field, no computed value or transformation
test_eq_same_n (MinimumCount)	test_wait_policy.py:42	P1	MinimumCount(5) == MinimumCount(5) — attrs-generated value equality, not project logic
test_hash_consistent (MinimumCount)	test_wait_policy.py:51	P1	hash(MinimumCount(5)) == hash(MinimumCount(5)) — attrs-generated __hash__, not project logic
test_repr (MinimumCount)	test_wait_policy.py:48	P1	repr(MinimumCount(5)) == "MinimumCount(n=5)" — attrs-generated repr, no custom logic
test_explicit_values	test_asset_access_control.py:31	P2	Constructs AssetAccessControl(...) and asserts each field equals what was passed in — pure round-trip, no transformation or validation exercised
TestTaskGroupGetItem::test_getitem_missing_is_key_error	test_taskgroup.py	P1	NodeNotFound subclasses KeyError, so the sibling raises(NodeNotFound) already guarantees the raises(KeyError) catch — re-states a fact Python's class hierarchy enforces
TestDagGetItem::test_getitem_missing_is_key_error	test_dag.py	P1	Same as above for Dag.__getitem__
test_override_dag_default_args	test_taskgroup.py	P5	Byte-identical body to test_default_args (same Dag, same TaskGroup default_args, same three owner assertions); only the name and docstring differ

---

Was generative AI tooling used to co-author this PR?

• Yes — Claude Code (Opus 4.8)

Generated-by: Claude Code (Opus 4.8) following the guidelines