Deserialization of Untrusted Data

Overview llama-index is an Interface between LLMs and your data Affected versions of this package are vulnerable to Deserialization of Untrusted Data via the loadfromdisk function. An attacker can execute arbitrary code by supplying a crafted multiembedstore.pkl file in a user-controlled director...

Deserialization of Untrusted Data

Overview llama-index-indices-managed-bge-m3 is a LlamaIndex Indices Integration: Bge-M3 Affected versions of this package are vulnerable to Deserialization of Untrusted Data via the loadfromdisk function. An attacker can execute arbitrary code by supplying a crafted multiembedstore.pkl file in a...

Use of NullPointerException Catch to Detect NULL Pointer Dereference

Overview Affected versions of this package are vulnerable to Use of NullPointerException Catch to Detect NULL Pointer Dereference in the MultimodalTokenize function that improperly processes NULL from mtmdhelperbitmapinitfrombuf function of vendored llama.cpp. An attacker can cause the applicatio...

agentbuddy (0.1.1.dev33), athina (>=1.0.2, <=1.6.33) +58 more potentially affected by CVE-2024-14021 via llama-index (>=0.4.17, <=0.11.4)

llama-index PYPI version =0.4.17, =1.0.2, =0.0.1, =0.0.33, =0.18.2, =0.3.2, =0.1.1, =0.3.8.0, =1.2.3, =0.1.0, =0.0.8, =0.1.1, =0.1.0, =0.1.0, =0.1.3 and more Source cves: CVE-2024-14021 Source advisory: OSV:PYSEC-2026-85...

agentbuddy (0.1.1.dev33), airunner (>=3.0.0, <=3.1.7) +75 more potentially affected by CVE-2024-58339 via llama-index (>=0.4.17, <=0.12.2)

llama-index PYPI version =0.4.17, =3.0.0, =1.0.2, =0.0.1, =0.0.33, =0.18.2, =0.3.2, =0.1.1, =0.3.8.0, =1.2.3, =0.1.0, =0.0.8, =0.1.1, =0.1.0, =0.4.0 and more Source cves: CVE-2024-58339 Source advisory: OSV:PYSEC-2026-86...

PYSEC-2026-85

LlamaIndex run-llama/llamaindex versions up to and including 0.11.6 contain an unsafe deserialization vulnerability in BGEM3Index.loadfromdisk in llamaindex/indices/managed/bgem3/base.py. The function uses pickle.load to deserialize multiembedstore.pkl from a user-supplied persistdir without...

CVE-2024-58339

LlamaIndex run-llama/llamaindex versions up to and including 0.12.2 contain an uncontrolled resource consumption vulnerability in the VannaPack VannaQueryEngine implementation. The customquery logic generates SQL statements from a user-supplied prompt and executes them via vn.runsql without...

SUSE 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-2026-21869

A flaw was found in llama.cpp. A remote attacker can exploit an input validation vulnerability in the server's completion endpoints. By supplying a negative value for the ndiscard parameter in JSON input, an attacker can cause out-of-bounds memory writes. This can lead to a process crash or enabl...

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...

UBUNTU-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-2026-21869

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...

llama.cpp Buffer Error Vulnerability

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 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 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-2026-1234

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:...