llama.cpp 缓冲区错误漏洞

llama.cpp is a multimodal model by the individual developer Georgi Gerganov. A buffer error vulnerability exists in llama.cpp 55d4206c8 and earlier versions, which stems from an unvalidated non-negative value of the ndiscard parameter, which can lead to out-of-bounds memory writes and remote code...

Out-of-bounds Write

Overview Affected versions of this package are vulnerable to Out-of-bounds Write via the ndiscard parameter in the server's completion endpoints. An attacker can cause memory corruption, process crashes, or potentially execute arbitrary code by supplying a negative value that leads to out-of-boun...

CVE-2026-21869 llama.cpp has Out-of-bounds Write in llama-server

llama.cpp is an inference of several LLM models in C/C++. In commits 55d4206c8 and prior, the ndiscard parameter is parsed directly from JSON input in the llama.cpp server's completion endpoints without validation to ensure it's non-negative. When a negative value is supplied and the context fill...

CVE-2026-21869 llama.cpp has Out-of-bounds Write in llama-server

llama.cpp is an inference of several LLM models in C/C++. In commits 55d4206c8 and prior, the ndiscard parameter is parsed directly from JSON input in the llama.cpp server's completion endpoints without validation to ensure it's non-negative. When a negative value is supplied and the context fill...

CVE-2026-21869

llama.cpp is an inference of several LLM models in C/C++. In commits 55d4206c8 and prior, the ndiscard parameter is parsed directly from JSON input in the llama.cpp server's completion endpoints without validation to ensure it's non-negative. When a negative value is supplied and the context fill...

CVE-2026-1234

A security issue was identified in the Llama Stack server when PGVector is used as a vector store provider. During initialization, the server logs print the PGVector database password in clear text. This occurs due to insufficient redaction of sensitive configuration fields. As a result, anyone...

Emoji-Based Jailbreaking of Large Language Models

Large Language Models LLMs are integral to modern AI applications, but their safety alignment mechanisms can be bypassed through adversarial prompt engineering. This study investigates emoji-based jailbreaking, where emoji sequences are embedded in textual prompts to trigger harmful and unethical...

Low Rank Comes with Low Security: Gradient Assembly Poisoning Attacks against Distributed LoRA-Based LLM Systems

Low-Rank Adaptation LoRA has become a popular solution for fine-tuning large language models LLMs in federated settings, dramatically reducing update costs by introducing trainable low-rank matrices. However, when integrated with frameworks like FedIT, LoRA introduces a critical vulnerability:...

SQL Injection

llamaindex is vulnerable to SQL Injection. The vulnerability is due to unsafe construction of SQL queries without prepared statements in the duckdbretriever component, which allows an attacker to inject arbitrary SQL commands and execute malicious code...

Persistent Backdoor Attacks under Continual Fine-Tuning of LLMs

Backdoor attacks embed malicious behaviors into Large Language Models LLMs, enabling adversaries to trigger harmful outputs or bypass safety controls. However, the persistence of the implanted backdoors under user-driven post-deployment continual fine-tuning has been rarely examined. Most prior...

Llama-Based Source Code Vulnerability Detection: Prompt Engineering Vs Fine Tuning

The significant increase in software production, driven by the acceleration of development cycles over the past two decades, has led to a steady rise in software vulnerabilities, as shown by statistics published yearly by the CVE program. The automation of the source code vulnerability detection...

Securing Large Language Models (LLMs) from Prompt Injection Attacks

Large Language Models LLMs are increasingly being deployed in real-world applications, but their flexibility exposes them to prompt injection attacks. These attacks leverage the model's instruction-following ability to make it perform malicious tasks. Recent work has proposed JATMO, a task-specif...

Evaluating LLMs for One-Shot Patching of Real and Artificial Vulnerabilities

Automated vulnerability patching is crucial for software security, and recent advancements in Large Language Models LLMs present promising capabilities for automating this task. However, existing research has primarily assessed LLMs using publicly disclosed vulnerabilities, leaving their...

TASO: Jailbreak LLMs Via Alternative Template and Suffix Optimization

Many recent studies showed that LLMs are vulnerable to jailbreak attacks, where an attacker can perturb the input of an LLM to induce it to generate an output for a harmful question. In general, existing jailbreak techniques either optimize a semantic template intended to induce the LLM to produc...

Researchers Find Serious AI Bugs Exposing Meta, Nvidia, and Microsoft Inference Frameworks

Cybersecurity researchers have uncovered critical remote code execution vulnerabilities impacting major artificial intelligence AI inference engines, including those from Meta, Nvidia, Microsoft, and open-source PyTorch projects such as vLLM and SGLang. "These vulnerabilities all traced back to t...

Insecure Temporary File Usage

llama-index-core is vulnerable to Insecure Temporary File Usage. The vulnerability is due to the use of a predictable hardcoded cache directory /tmp/llamaindex in getcachedir, where attackers on multi-user Linux systems can steal cached model data, poison embeddings, or exploit symlink race...

EUVD-2025-117312

Malicious code in ill-chocolate-llama npm...

EUVD-2025-117162

Malicious code in prepared-red-llama npm...

Malicious code in ill-chocolate-llama (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector b5555ac2c54f83f5d6054d0c5ca3c988693b9e111b21e1e7d6e7350584dd3c05 This package appears to be part of the tea.xyz token reward campaign that flooded npm. These packages typically contain autopublish scripts auto.js,...

MAL-2025-138940 Malicious code in prepared-red-llama (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector d1a7781454329c1c6b148e3d24e92ef63ac0a34da6bd11cae790b3057bd11cfa This package appears to be part of the tea.xyz token reward campaign that flooded npm. These packages typically contain autopublish scripts auto.js,...