Raw vector
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/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:XSummary
CVE-2026-25481 is a critical-severity Code Injection (CWE-94) vulnerability in Langroid Langroid. Its CVSS base score is 9.4 (Critical).
Operationally, exploitation aligns with the MITRE ATT&CK technique Command and Scripting Interpreter (T1059); ranked at the 47th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog; a public proof-of-concept is referenced.
The strongest mitigations our analysis identified map to SA-11 (Developer Testing and Evaluation) and SI-10 (Information Input Validation) — 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-2026-25481 is a code injection vulnerability affecting Langroid, a framework for building large-language-model-powered applications. In versions prior to 0.59.32, the TableChatAgent component can invoke the pandas_eval tool to evaluate expressions, bypassing a Web Application Firewall (WAF) introduced in langroid/utils/pandas_utils.py to mitigate CVE-2025-46724. The bypass occurs because the _literal_ok() function returns False on invalid input instead of raising an UnsafeCommandError, while dangerous dunder attributes such as __init__, __globals__, and __builtins__ remain unrestricted. This enables attackers to chain whitelisted DataFrame methods to access the eval builtin and execute arbitrary code.
The vulnerability carries a CVSS v3.1 base score of 9.6 (AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H), indicating exploitation over the network with low complexity, no privileges required, but user interaction needed, and a changed scope with high impacts on confidentiality, integrity, and availability. Attackers can exploit it by crafting malicious input that tricks users into processing it via a vulnerable Langroid application using TableChatAgent, such as in an LLM-driven interface handling table data. Successful exploitation grants arbitrary code execution on the host system.
The issue has been addressed in Langroid version 0.59.32. Official advisories and the patching commit are available on GitHub, including the security advisory GHSA-jqq5-wc57-f8hj, related advisory GHSA-x34r-63hx-w57f for the prior CVE, and commit 30abbc1a854dee22fbd2f8b2f575dfdabdb603ea detailing the fix.
This vulnerability is particularly relevant to AI/ML deployments, as Langroid targets LLM-powered applications, and it stems from CWE-94 (improper control of code generation). No public reports of real-world exploitation are noted as of the CVE publication on 2026-02-04.
OWASP Top 10 for Web (2025)
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2026-5347
Vulnerability Data
Langroid is a framework for building large-language-model-powered applications. Prior to version 0.59.32, there is a bypass to the fix for CVE-2025-46724. TableChatAgent can call pandas_eval tool to evaluate the expression. There is a WAF in langroid/utils/pandas_utils.py introduced to block code…
more
injection CVE-2025-46724. However it can be bypassed due to _literal_ok() returning False instead of raising UnsafeCommandError on invalid input, combined with unrestricted access to dangerous dunder attributes (__init__, __globals__, __builtins__). This allows chaining whitelisted DataFrame methods to leak the eval builtin and execute arbitrary code. This issue has been patched in version 0.59.32.
- CWE(s)
Related Threats
MITRE ATT&CK Enterprise Techniques
CVEs Like This One
Affected Assets
Mitigating Controls
Control response
—
—
—
V1.3.1
Mitigating Controls (NIST 800-53 r5) AI
Developer testing and evaluation finds code paths that accept and execute externally influenced strings.
Input validation directly stops untrusted data from being used to construct executable code without neutralization.
Least privilege limits the damage an injected code fragment can perform once executed.
Requiring documented secure development standards and tools enforces use of safe code-generation APIs and escaping.
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.
PR.PS-06's SDLC practices directly target injection flaws via secure coding and testing (mostly), yet as a single broad outcome it leaves many code-generation specifics unaddressed (partial).
PR.DS-10 protects runtime data confidentiality/integrity but has no bearing on neutralizing externally influenced input during code generation, so neither direction shows any preventive effect.
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.
Banning unapproved code samples and unauthenticated web services, combined with secure-coding standards and SAST, prevents the dynamic generation or inclusion of attacker-supplied code.
Controls that restrict unauthorized or malicious code from being introduced via external networks or removable media limit opportunities for an attacker to inject and execute arbitrary code.