CVE-2026-27941
Published: 26 February 2026
Summary
CVE-2026-27941 is a critical-severity Inclusion of Functionality from Untrusted Control Sphere (CWE-829) vulnerability in Openlit Openlit Software Development Kit. Its CVSS base score is 9.9 (Critical).
Operationally, exploitation aligns with the MITRE ATT&CK technique Compromise Software Dependencies and Development Tools (T1195.001); ranked at the 21.5th 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 Other AI Platforms.
The strongest mitigations our analysis identified are NIST 800-53 AC-6 (Least Privilege) and CM-5 (Access Restrictions for Change).
Threat & Defense at a Glance
Threat & Defense Details
Mitigating Controls (NIST 800-53 r5)AI
Establishes and enforces secure configuration settings for GitHub Actions workflows to prevent use of pull_request_target event with untrusted code checkout in privileged contexts.
Enforces least privilege on GITHUB_TOKEN permissions and limits exposure of sensitive secrets in workflows, directly mitigating privilege escalation from untrusted PR code execution.
Restricts access to modify workflow configuration files, preventing introduction or persistence of vulnerable pull_request_target configurations in the repository.
MITRE ATT&CK Enterprise TechniquesAI
Why these techniques?
Vulnerability enables execution of attacker-controlled code from forks inside privileged GitHub Actions workflows (pull_request_target), directly mapping to poisoned pipeline execution (T1677) and compromise of development tools / software supply chain (T1195.001/002).
NVD Description
OpenLIT is an open source platform for AI engineering. Prior to version 1.37.1, several GitHub Actions workflows in OpenLIT's GitHub repository use the `pull_request_target` event while checking out and executing untrusted code from forked pull requests. These workflows run with…
more
the security context of the base repository, including a write-privileged `GITHUB_TOKEN` and numerous sensitive secrets (API keys, database/vector store tokens, and a Google Cloud service account key). Version 1.37.1 contains a fix.
Deeper analysisAI
CVE-2026-27941 is a high-severity vulnerability (CVSS 9.9, CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H) affecting GitHub Actions workflows in the OpenLIT open-source platform for AI engineering. Prior to version 1.37.1, several workflows in the OpenLIT GitHub repository improperly used the `pull_request_target` event trigger. This allowed the workflows to check out and execute untrusted code from pull requests originating from forked repositories, while running in the elevated security context of the base repository. The context included a write-privileged `GITHUB_TOKEN` and exposure to sensitive secrets such as API keys, database and vector store tokens, and a Google Cloud service account key. The issue stems from CWE-829 (Inclusion of Functionality from Untrusted Control Sphere).
An attacker with low privileges (PR:L), such as a user able to fork the repository and submit a pull request, can exploit this vulnerability remotely over the network with no user interaction required. By crafting a malicious pull request containing arbitrary code in the workflow, the attacker triggers execution in the base repository's privileged context. This grants full read/write access to the repository via the `GITHUB_TOKEN`, as well as exfiltration or misuse of all exposed secrets, potentially leading to compromise of linked services like databases, vector stores, and Google Cloud resources. The changed scope (S:C) amplifies impact across confidentiality, integrity, and availability.
The vulnerability was addressed in OpenLIT version 1.37.1, which contains a fix detailed in commit 4a62039a1659d6cbb8913172693f587b5fc2546c. Security practitioners should upgrade to this version or later and review workflows for similar misuse of `pull_request_target`. Additional guidance is available in the GitHub Security Advisory at GHSA-9jgv-x8cq-296q.
Details
- CWE(s)
Affected Products
AI Security AnalysisAI
- AI Category
- Other AI Platforms
- Risk Domain
- N/A
- OWASP Top 10 for LLMs 2025
- None mapped
- Classification Reason
- Matched keywords: ai