ZeroHour

CVE-2026-83551

large

Cleartext HMAC key leak in Amazon SageMaker Python SDK enables pipeline code execution

CVSS 4.0
8.5 high
EPSS
<1%p31
Published
()
Modified
AI analysis

The @step and @remote decorator pipeline component of the Amazon SageMaker Python SDK stores the HMAC signing key used for payload integrity verification in cleartext within SageMaker DescribePipeline API responses (CWE-312, cleartext storage of sensitive information). An authenticated remote user with high privileges in the same AWS account can call DescribePipeline, extract the signing key, and forge valid integrity signatures for specially crafted function payloads. Successful abuse lets the attacker achieve code execution in another user's pipeline execution context within the same AWS account. Users of the SageMaker Python SDK before v3.11.0 (v3 line) and before v2.256.0 (v2 line) who use the @step/@remote pipeline feature are affected. No public proof-of-concept is known, the flaw is not in CISA KEV, and EPSS estimates only a 0.4% probability of exploitation within 30 days.

What to do: Upgrade the SageMaker Python SDK to v3.11.0 or later (for the 3.x line) or v2.256.0 or later (for the 2.x line). In the interim, audit IAM policies for which principals in the account can call sagemaker:DescribePipeline, since that API is the key-exposure vector, and review recent pipeline executions in the account for unexpected or unapproved function payloads. Note the flaw requires an already-authenticated, high-privilege user in the same AWS account, so cross-account exposure is not indicated by the data.

Affected
Amazon Web Services Amazon SageMaker Python SDK (2.x line)all versions before 2.256.0
Amazon Web Services Amazon SageMaker Python SDK (3.x line)all versions before 3.11.0
Estimated exposure
largetens of thousands of development/CI environments (≈10k–100k), limited to accounts actively using the @step/@remote pipeline decorators — The SageMaker Python SDK is among the most widely adopted ML toolchains (millions of monthly PyPI downloads and SageMaker's position as a leading managed ML service), but only the subset of users who build pipelines with the @step/@remote…

Order-of-magnitude estimate by the model from install counts, market share and public scan data it knows; verify before quoting.

Description

Cleartext storage of sensitive information in the @step and @remote decorator pipeline component in Amazon SageMaker Python SDK before v3.11.0 and v2.256.0 might allow an authenticated remote user to extract the HMAC signing key from SageMaker DescribePipeline API responses and forge valid integrity signatures for specially crafted function payloads, achieving code execution in another user's pipeline execution context within the same AWS account.

Weakness
CWE-312
Vector
CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:N/VA:N/SC:H/SI:H/SA:H/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

In the news

No ingested article mentions this CVE yet.