Raw vector
CVSS:3.1/AV:N/AC:H/PR:H/UI:N/S:C/C:H/I:H/A:HSummary
CVE-2026-26268 is a high-severity Missing Authorization (CWE-862) vulnerability in Anysphere Cursor. Its CVSS base score is 8.0 (High).
Operationally, exploitation aligns with the MITRE ATT&CK technique Exploitation for Privilege Escalation (T1068); ranked at the 48th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog.
This vulnerability is AI-related — categorised as Enterprise AI Assistants; in the LLM/Generative AI Risks risk domain.
The strongest mitigations our analysis identified map to AC-24 (Access Control Decisions) and AC-25 (Reference Monitor) — 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-2026-26268 is a sandbox escape vulnerability in Cursor, a code editor built for programming with AI, affecting versions prior to 2.5. The flaw stems from improper protection of .git configuration files, enabling unauthorized writes to settings such as git hooks (CWE-862: Missing Authorization). Published on 2026-02-13, it carries a CVSS v3.1 base score of 8.0 (AV:N/AC:H/PR:H/UI:N/S:C/C:H/I:H/A:H).
A malicious agent, for example via prompt injection, can exploit this by writing to the .git configuration from within the sandboxed environment. When Git automatically triggers these commands, such as executing hooks, it results in remote code execution (RCE) outside the sandbox. No user interaction is required, though exploitation demands network access, high attack complexity, and high privileges.
The vulnerability is fixed in Cursor version 2.5. Additional details on mitigation are available in the GitHub security advisory at https://github.com/cursor/cursor/security/advisories/GHSA-8pcm-8jpx-hv8r.
OWASP Top 10 for Web (2025)
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2026-7368
Vulnerability Data
Cursor is a code editor built for programming with AI. Sandbox escape via writing .git configuration was possible in versions prior to 2.5. A malicious agent (ie prompt injection) could write to improperly protected .git settings, including git hooks, which…
more
may cause out-of-sandbox RCE next time they are triggered. No user interaction was required as Git executes these commands automatically. Fixed in version 2.5.
- CWE(s)
AI Security AnalysisAI
- AI Category
- Enterprise AI Assistants
- Risk Domain
- LLM/Generative AI Risks
- OWASP Top 10 for LLMs 2025
- None mapped
- Classification Reason
- Matched keywords: ai, prompt injection
Related Threats
MITRE ATT&CK Enterprise Techniques
CVEs Like This One
Affected Assets
Mitigating Controls
Control response
—
—
- 5 hardening rules · 3 OS baselines
—
Mitigating Controls (NIST 800-53 r5) AI
AC-3 directly requires enforcement of authorization decisions on every access request, structurally eliminating missing authorization checks.
AC-24 mandates that access control decisions are applied to each request, preventing bypass of authorization logic.
AC-25 requires a tamperproof, always-invoked reference monitor that ensures authorization checks cannot be omitted.
AC-6 reduces the set of actions reachable without proper authorization, limiting blast radius of missing checks.
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.
Explicitly requires defining, enforcing, and reviewing authorizations and least privilege, directly preventing missing authorization checks.
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.
Requiring authentication and credentials before any access occurs eliminates the absence of authorization checks that would otherwise allow an unauthenticated actor to reach protected resources.
Requiring formal authorization of every access request before rights are granted ensures that checks for required permissions are performed, preventing missing authorization checks from being introduced.
Mandatory authorization checks and central records of granted rights ensure that every access attempt is preceded by an explicit decision rather than relying on missing checks.
Segregating the approval of access rights from their implementation provides an independent check that reduces the impact of missing authorization checks in the resulting system configuration.
By requiring competent outsiders to verify that every function enforces the need-to-know principle, the control lowers the likelihood that missing authorization checks persist undetected.
Defining authorization responsibilities and reviewing risk-treatment progress throughout the project lifecycle catches missing authorization checks before the system is deployed.
Hardening callouts derived
Configuration rules from DISA STIG baselines that bear on weaknesses of the type cited by this CVE. Each rule is shown with the relationship its mapping actually records, against the CWE it was authored against. Derived via CVE→CWE over `controls_xwalks` (authoritative rows only; rows rated `none` are excluded).
Oracle Linux 8 (2 rules)
- V-248581 OL 8 must require users to provide a password for privilege escalation. prevents CWE-862
- V-252656 The OL 8 operating system must not be configured to bypass password requirements for privilege escalation. prevents CWE-862
RHEL 7 (2 rules)
- V-251704 The Red Hat Enterprise Linux operating system must not be configured to bypass password requirements for privilege escalation. prevents CWE-862
- V-204429 The Red Hat Enterprise Linux operating system must be configured so that users must provide a password for privilege escalation. prevents CWE-862