AI Threat AlertCVE-2026-115456: GCC libstdc++: from_chars OOB read causes DoS
LOWA low-severity flaw in GCC's libstdc++ lets std::from_chars() read past the end of a buffer when given a string that is not null-terminated. This can crash an application or cause invalid data parsing. It matters mainly because Red Hat lists many OpenShift AI (rhoai) container images and CoreOS images as affected, since they ship the toolchain runtime. The CVSS score is only 3.7 (network, high complexity, availability impact only). There is no known exploit, no Nuclei template, no KEV listing and no EPSS score, so exploitation looks unlikely. Treat it as routine patching: apply Red Hat's updated gcc and libstdc++ packages and rebuild images when fixes are published, and don't raise a breaking alert.
What is the risk?
Low risk. Exploitation needs a network-reachable application that passes untrusted, unterminated input to std::from_chars(), and attack complexity is high. Impact is limited to availability (crash) or malformed numeric parsing, with no confidentiality or integrity impact. No public exploit, KEV listing or SSVC decision exists, and EPSS is unavailable. The package-level risk score of 81/100 and the 137 other CVEs reflect the broad gcc footprint, not this specific flaw.
How does the attack unfold?
What systems are affected?
How severe is it?
What is the attack surface?
What should I do?
1 step1) Track Red Hat advisory CVE-2026-115456 and Bugzilla 2563044 for fixed gcc, libstdc++ and image errata. 2) Apply updated packages and pull rebuilt RHOAI and OpenShift images through normal patch cycles. 3) Rebuild internally compiled C++ components against the fixed toolchain. 4) As a workaround, ensure buffers passed to std::from_chars() are bounded by an explicit end pointer and validated before parsing. 5) For detection, monitor for unexpected crashes or restarts (SIGSEGV) in C++ services and controllers, and use SBOM scanning to find affected base images.
How is it classified?
Which compliance frameworks are affected?
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2026-115456?
A low-severity flaw in GCC's libstdc++ lets std::from_chars() read past the end of a buffer when given a string that is not null-terminated. This can crash an application or cause invalid data parsing. It matters mainly because Red Hat lists many OpenShift AI (rhoai) container images and CoreOS images as affected, since they ship the toolchain runtime. The CVSS score is only 3.7 (network, high complexity, availability impact only). There is no known exploit, no Nuclei template, no KEV listing and no EPSS score, so exploitation looks unlikely. Treat it as routine patching: apply Red Hat's updated gcc and libstdc++ packages and rebuild images when fixes are published, and don't raise a breaking alert.
Is CVE-2026-115456 actively exploited?
No confirmed active exploitation of CVE-2026-115456 has been reported, but organizations should still patch proactively.
How to fix CVE-2026-115456?
1) Track Red Hat advisory CVE-2026-115456 and Bugzilla 2563044 for fixed gcc, libstdc++ and image errata. 2) Apply updated packages and pull rebuilt RHOAI and OpenShift images through normal patch cycles. 3) Rebuild internally compiled C++ components against the fixed toolchain. 4) As a workaround, ensure buffers passed to std::from_chars() are bounded by an explicit end pointer and validated before parsing. 5) For detection, monitor for unexpected crashes or restarts (SIGSEGV) in C++ services and controllers, and use SBOM scanning to find affected base images.
What systems are affected by CVE-2026-115456?
This vulnerability affects the following AI/ML architecture patterns: model serving, MLOps platforms, Kubernetes operators for ML, training pipelines, container base images.
What is the CVSS score for CVE-2026-115456?
CVE-2026-115456 has a CVSS v3.1 base score of 3.7 (LOW).
What is the AI security impact?
Affected AI Architectures
MITRE ATLAS Techniques
AML.T0010.001AI SoftwareAML.T0029Denial of AI ServiceCompliance Controls Affected
What are the technical details?
Original Advisory
A flaw was found in gcc. An attacker could cause a Denial of Service (DoS) by passing a string that is not null-terminated to the std::from_chars() function in the libstdc++ library. This issue triggers an out-of-bounds read beyond the allocated buffer, potentially leading to invalid data parsing or an application crash.
Exploitation Scenario
An attacker sends crafted input to a network-facing C++ service in an ML platform, such as a model-serving or controller component. The service passes the input to std::from_chars() without a proper bound, so the function reads beyond the allocated buffer. The out-of-bounds read produces garbage numeric values or crashes the process. Repeated requests could cause a repeated crash, degrading model serving or pipeline orchestration until the service is restarted. Exploitation depends on specific application code and is difficult to achieve.
Weaknesses (CWE)
CWE-125 — Out-of-bounds Read: The product reads data past the end, or before the beginning, of the intended buffer.
- [Implementation] Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylis
- [Architecture and Design] Use a language that provides appropriate memory abstractions.
Source: MITRE CWE corpus.
CVSS Vector
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:LReferences
- access.redhat.com/security/cve/CVE-2026-115456vdb-entryx_refsource_REDHAT
- bugzilla.redhat.com/show_bug.cgiissue-trackingx_refsource_REDHAT
Timeline
Related Vulnerabilities
CVE-2025-1537910.0MLflow: RCE via unsanitized model dependency specs
Same package: mlflowCVE-2023-376510.0MLflow: path traversal allows arbitrary file read
Same package: mlflowCVE-2023-11779.8MLflow: path traversal allows arbitrary file read/write
Same package: mlflowCVE-2026-26359.8mlflow: security flaw enables exploitation
Same package: mlflowCVE-2023-27809.8MLflow: path traversal allows arbitrary file read/write
Same package: mlflow