Incidents tagged LLM

Evidence indicates that Large Language Model (LLM) is affected by a security issue. Exploitation evidence is classified as actively exploited.

A critical vulnerability (CVE-2026-59726) in Ruflo's MCP bridge exposed unauthenticated shell command execution via default network binding (0.0.0.0:3001). This flaw enabled attackers to achieve remote code execution, steal LLM API keys, and poison AI model memory.

Evidence indicates that Custom Code Guardrails is affected by remote code execution.

Novee has introduced an AI Red Teaming platform to proactively identify security vulnerabilities in LLM-powered applications. Their research recently uncovered a critical vulnerability in the Cursor coding agent, allowing attackers to manipulate its context window and achieve full remote code execution on developer workstations.

Evidence indicates that AI systems is affected by remote code execution. Exploitation evidence is classified as confirmed in the wild.

Evidence indicates that AI agents is affected by a security issue. Exploitation evidence is classified as confirmed in the wild.

Evidence indicates that Meta is affected by remote code execution.

Evidence indicates that Claude is affected by remote code execution.

Evidence indicates that Large Language Model is affected by prompt injection. Exploitation evidence is classified as demonstrated.

Evidence indicates that Claude is affected by a security issue.

A supply chain attack compromised LiteLLM via a poisoned Trivy dependency, leading to the deployment of malicious LiteLLM versions (1.82.7/1.82.8) that exfiltrated 153GB of corporate credentials. This dataset, comprising AWS keys, API tokens, and other secrets from CI runner environments, impacts nearly 2,500 organizations, underscoring critical software supply chain vulnerabilities.

Evidence indicates that LiteLLM proxy is affected by a security issue.

Evidence indicates that OpenShift AI is affected by a security issue.

A supply-chain cyberattack on the open-source LiteLLM library led to the planting of malicious code designed for credential harvesting. This incident resulted in a significant data breach at AI startup Mercor, potentially exposing sensitive company data, user information, and confidential AI project details.

A security breach at AI data vendor Mercor potentially exposed sensitive AI training data, including proprietary methodologies and competitive intelligence, impacting clients like Meta. The incident is suspected to involve a supply chain attack via malicious code injected into the LiteLLM open-source library, used to steal credentials and facilitate data exfiltration.

A sophisticated multi-stage supply chain attack, initiated by compromising open-source security scanner Trivy to steal LiteLLM PyPI credentials, injected malicious versions (1.82.7, 1.82.8) of LiteLLM into customer environments. This enabled data exfiltration, system persistence, and lateral movement within Kubernetes clusters, notably leveraging AI coding assistants with unrestricted permissions as an unwitting infection vector.

AI firm Mercor confirmed a breach stemming from a supply chain attack involving the open-source LiteLLM PyPI package, where attackers published malicious versions after compromising maintainer credentials. This incident led to the alleged theft of 4TB of sensitive data, including candidate profiles, PII, source code, and API keys, subsequently listed by the Lapsus$ extortion group.

Attackers executed a supply-chain attack on the open-source library LiteLLM by exploiting stolen credentials to inject malicious code into its PyPI distribution pipeline. This malware actively harvested sensitive information, including API keys and cloud credentials, from affected systems, potentially leading to the compromise of up to 4TB of data from companies like Mercor.

The article outlines a comprehensive AI security roadmap addressing unique threats to LLMs and AI agents, such as prompt injection, data poisoning, model inversion, and data leakage, which exploit probabilistic system behaviors across the full AI lifecycle. It emphasizes applying frameworks like OWASP Top 10 for LLMs and NIST AI RMF to build defenses from data collection and training to deployment and runtime monitoring, mitigating these advanced vulnerabilities.

An extortion group executed a supply chain attack by compromising the open-source LiteLLM project, which serves as a widely-used AI model API proxy. This breach led to the theft of sensitive data from AI recruiting startup Mercor, underscoring systemic vulnerabilities in the AI industry's reliance on unvetted open-source dependencies.