CVE-2026-27905
Bentoml ≤ 1.4.36
Raw vector
CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/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:XSummary
CVE-2026-27905 is a high-severity Link Following (CWE-59) vulnerability in Bentoml Bentoml. Its CVSS base score is 8.6 (High).
Operationally, exploitation aligns with the MITRE ATT&CK technique Path Interception (T1034); ranked at the 17th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog; a public proof-of-concept is referenced.
This vulnerability is AI-related — categorised as Machine Learning Libraries; in the Supply Chain and Deployment risk domain.
The strongest mitigations our analysis identified map to AC-3 (Access Enforcement) and AC-6 (Least Privilege) — see the control section below for these in your framework.
Deeper analysis AI-assisted summary
Synthesised by an AI model from the NVD description and linked references — a reading aid, not an authoritative source.
CVE-2026-27905 is a vulnerability in BentoML, a Python library for building online serving systems optimized for AI applications and model inference. In versions prior to 1.4.36, the safe_extract_tarfile() function performs path validation to ensure tar members stay within the destination directory, but it only checks the symlink's own path for symlink members, neglecting the symlink's target path. This flaw, classified as CWE-59 (Improper Link Resolution Before File Access), enables symlink traversal during tar extraction.
Exploitation requires local access (AV:L) with low complexity (AC:L), no privileges (PR:N), and user interaction (UI:R), as reflected in its CVSS v3.1 base score of 7.8 (S:U/C:H/I:H/A:H). An attacker can craft a malicious bento or model tar file containing a symlink that points outside the extraction directory, followed by a regular file whose contents are written through the symlink. This achieves arbitrary file writes on the host filesystem.
The issue was addressed in BentoML 1.4.36. Mitigation involves updating to this version or later. Details on the patch are provided in the BentoML GitHub commit (https://github.com/bentoml/BentoML/commit/4e0eb007765ac04c7924220d643f264715cc9670) and security advisory (https://github.com/bentoml/BentoML/security/advisories/GHSA-m6w7-qv66-g3mf).
OWASP Top 10 for Web (2025)
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2026-9343
Vulnerability Data
BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.36, the safe_extract_tarfile() function validates that each tar member's path is within the destination directory, but for symlink members it only…
more
validates the symlink's own path, not the symlink's target. An attacker can create a malicious bento/model tar file containing a symlink pointing outside the extraction directory, followed by a regular file that writes through the symlink, achieving arbitrary file write on the host filesystem. This vulnerability is fixed in 1.4.36.
- CWE(s)
AI Security AnalysisAI
- AI Category
- Machine Learning Libraries
- Risk Domain
- Supply Chain and Deployment
- OWASP Top 10 for LLMs 2025
- None mapped
- Classification Reason
- Matched keywords: ai, bentoml
Related Threats
MITRE ATT&CK Enterprise Techniques
CVEs Like This One
Affected Assets
Mitigating Controls
Control response
—
—
—
V15.4.2
Mitigating Controls (NIST 800-53 r5) AI
Proper enforcement of access authorizations on the resolved target resource stops a link from reaching an unintended object.
Least-privilege limits the damage an attacker can cause after following an unintended link.
Validating file-name inputs can reject or canonicalize names that resolve to links before access occurs.
Mitigating Controls (NIST CSF 2.0) AI
Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→CSF cross-walk (authority under review) — links open the control.
Secure SDLC practices directly require code to validate paths and avoid unsafe link following.
Mitigating Controls (ISO/IEC 27001:2022 Annex A) AI
Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→ISO cross-walk (authority under review) — links open the control.
Security testing can detect link-following flaws before release.
Secure SDLC practices can mandate link-resolution checks and canonicalization before file access.
Application security requirements can explicitly require safe handling of symbolic links and path traversal.
Secure architecture principles include input validation and safe file-access design patterns.
Secure coding standards directly address canonicalization and symlink attacks during implementation.
Access-control rules can limit which files are reachable, reducing exposure to malicious links.