[https://nvbugs/6206179][test] Waive failing TestQwen3_5_35B_A3B::test_bf16_mtp[mtp_on]

The test crashes during autotuner warmup with
'NoneType' object has no attribute 'gather_ids' at
modeling_speculative.py:1748 when MTP eagle one-model is combined
with Qwen3.5-35B-A3B. Pre-existing regression on main introduced by
the Qwen3.5 VL MoE landing (96a4a09); unrelated to this PR's
changes. Tracked in https://nvbugs/6206179.

Signed-off-by: Brian Nguyen <brnguyen@nvidia.com>

Author: brnguyen2

tests/integration/test_lists/waives.txt

@@ -142,6 +142,7 @@ accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp1-cutlass]
 accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4_adp_on-trtllm] SKIP (https://nvbugs/6094068)
 accuracy/test_llm_api_pytorch.py::TestQwen3_235B_A22B::test_fp8[latency] SKIP (https://nvbugs/6177390)
 accuracy/test_llm_api_pytorch.py::TestQwen3_235B_A22B::test_fp8[throughput_latency] SKIP (https://nvbugs/6177390)
+accuracy/test_llm_api_pytorch.py::TestQwen3_5_35B_A3B::test_bf16_mtp[mtp_on] SKIP (https://nvbugs/6206179)
 accuracy/test_llm_api_pytorch.py::TestQwen3_8B::test_fp8_block_scales_early_first_token_response SKIP (https://nvbugs/6200128)
 accuracy/test_llm_api_pytorch_multimodal.py::TestKimiK25::test_nvfp4[dep8] SKIP (https://nvbugs/6182617)
 accuracy/test_llm_api_pytorch_multimodal.py::TestMistralLarge3_675B::test_nvfp4_4gpus[latency_moe_trtllm] SKIP (https://nvbugs/6181383)