AML.T0043.003

Manual Modification

Adversaries may manually modify the input data to craft adversarial data. They may use their knowledge of the target model to modify parts of the data they suspect helps the model in performing its task. The adversary may use trial and error until they are able to verify they have a working adversarial input.

22 CVEs mapped View on MITRE ATLAS →

Severity	CVE	Headline	Package	CVSS
CRITICAL	CVE-2026-22778	vllm: security flaw enables exploitation	vllm	9.8
HIGH	CVE-2021-29607	TensorFlow: heap OOB write in SparseAdd op	tensorflow	7.8
HIGH	CVE-2021-29537	TensorFlow: heap overflow in QuantizedResizeBilinear op	tensorflow	7.8
HIGH	CVE-2021-29529	TensorFlow: heap buffer overflow in quantized image resize	tensorflow	7.8
HIGH	CVE-2021-29514	TensorFlow: heap buffer overflow in RaggedBincount op	tensorflow	7.8
HIGH	CVE-2022-36001	TensorFlow: DoS via type confusion in DrawBoundingBoxes	tensorflow	7.5
HIGH	CVE-2020-5215	TensorFlow: type confusion DoS crashes eager mode inference	tensorflow	7.5
HIGH	CVE-2026-44549	open-webui: XSS via XLSX preview enables session hijack	open-webui	7.3
MEDIUM	CVE-2026-44222	vLLM: token injection DoS via multimodal placeholders	vllm	6.5
MEDIUM	CVE-2022-23584	TensorFlow: use-after-free in PNG decode causes DoS	tensorflow	6.5
MEDIUM	CVE-2022-23583	TensorFlow: SavedModel type confusion triggers DoS crash	tensorflow	6.5
MEDIUM	CVE-2022-23557	TensorFlow TFLite: DoS via divide-by-zero in BiasAndClamp	tensorflow	6.5
MEDIUM	CVE-2026-34760	vLLM: audio downmix mismatch enables adversarial input		5.9
MEDIUM	CVE-2026-40190	langsmith: prototype pollution enables auth bypass, RCE	langsmith	5.6
MEDIUM	CVE-2022-29211	TensorFlow: NaN input crashes histogram op (CPU DoS)	tensorflow	5.5
MEDIUM	CVE-2021-29519	TensorFlow SparseCross: type confusion DoS	tensorflow	5.5
MEDIUM	CVE-2022-29192	TensorFlow: DoS via QuantizeAndDequantize input validation	tensorflow	5.5
MEDIUM	GHSA-26jh-r8g2-6fpr	Gradio: Dropdown validation bypass enables arbitrary input	gradio	5.3
MEDIUM	CVE-2025-46148	PyTorch: PairwiseDistance silent miscalculation, integrity risk	pytorch	5.3
LOW	CVE-2025-25183	vLLM: hash collision enables prefix cache poisoning	vllm	2.6
LOW	CVE-2026-7845	Langchain-Chatchat: weak image hash allows integrity bypass	langchain-chatchat	2.6
MEDIUM	CVE-2024-52524	Giskard: ReDoS in text perturbation causes DoS		—