Lucene search
+L

1258 matches found

Schneier on Security
Schneier on Security
added 2025/04/29 11:03 a.m.13 views

Applying Security Engineering to Prompt Injection Security

This seems like an important advance in LLM security against prompt injection: Google DeepMind has unveiled CaMeL CApabilities for MachinE Learning, a new approach to stopping prompt-injection attacks that abandons the failed strategy of having AI models police themselves. Instead, CaMeL treats...

7.4AI score
SaveExploits0
Packet Storm News
Packet Storm News
added 2025/04/29 12:00 a.m.29 views

CachePrune: Neural-Based Attribution Defense against Indirect Prompt Injection Attacks

Large Language Models LLMs are identified as being susceptible to indirect prompt injection attack, where the model undesirably deviates from user-provided instructions by executing tasks injected in the prompt context. This vulnerability stems from LLMs' inability to distinguish between data and...

7.2AI score
SaveExploits0
Packet Storm News
Packet Storm News
added 2025/04/29 12:00 a.m.38 views

Token-Efficient Prompt Injection Attack: Provoking Cessation in LLM Reasoning Via Adaptive Token Compression

While reasoning large language models LLMs demonstrate remarkable performance across various tasks, they also contain notable security vulnerabilities. Recent research has uncovered a "thinking-stopped" vulnerability in DeepSeek-R1, where model-generated reasoning tokens can forcibly interrupt th...

7.5AI score
SaveExploits0
Packet Storm News
Packet Storm News
added 2025/04/29 12:00 a.m.12 views

Robustness Via Referencing: Defending against Prompt Injection Attacks by Referencing the Executed Instruction

Large language models LLMs have demonstrated impressive performance and have come to dominate the field of natural language processing NLP across various tasks. However, due to their strong instruction-following capabilities and inability to distinguish between instructions and data content, LLMs...

7.4AI score
SaveExploits0
RedhatCVE
RedhatCVE
added 2025/04/26 5:01 a.m.25 views

CVE-2025-31363

Mattermost versions 10.4.x = 10.4.2, 10.5.x = 10.5.0, 9.11.x = 9.11.9 fail to restrict domains the LLM can request to contact upstream which allows an authenticated user to exfiltrate data from an arbitrary server accessible to the victim via performing a prompt injection in the AI plugin's Jira...

6.5CVSS7AI score0.00272EPSS
SaveExploits0References1
Packet Storm News
Packet Storm News
added 2025/04/25 12:00 a.m.23 views

Adversarial Attacks on LLM-As-A-Judge Systems: Insights from Prompt Injections

LLM as judge systems used to assess text quality code correctness and argument strength are vulnerable to prompt injection attacks. We introduce a framework that separates content author attacks from system prompt attacks and evaluate five models Gemma 3.27B Gemma 3.4B Llama 3.2 3B GPT 4 and Clau...

7.7AI score
SaveExploits0
CNVD
CNVD
added 2025/04/25 12:00 a.m.15 views

Mattermost Information Disclosure Vulnerability

Mattermost is an open source collaboration platform from Mattermost, Inc. in the United States. Mattermost suffers from an information disclosure vulnerability. The vulnerability stems from an under-restricted LLM request domain. An attacker can exploit the vulnerability to perform prompt injecti...

6.5CVSS3.5AI score0.00272EPSS
SaveExploits0References1
Veracode
Veracode
added 2025/04/24 2:08 a.m.12 views

Prompt Injection

github.com/mattermost/mattermost-server is vulnerable to prompt injection. The vulnerability is due to insufficient domain restriction to the AI plugin's Jira tool, allowing authenticated users to exfiltrate data from arbitrary servers via crafted prompts...

6.5CVSS6.9AI score0.00272EPSS
SaveExploits0References3Affected Software1
Tenable Nessus
Tenable Nessus
added 2025/04/24 12:00 a.m.22 views

Mattermost Server 9.11.x < 9.11.10 / 10.4.x < 10.4.3 / 10.5.x < 10.5.1 / 10.6.0 (MMSA-2024-00401)

The version of Mattermost Server installed on the remote host is prior to 9.11.10, 10.4.3, or 10.5.1 / 10.6.0. It is, therefore, affected by a vulnerability as referenced in the MMSA-2024-00401 advisory. - Mattermost versions 10.4.x = 10.4.2, 10.5.x = 10.5.0, 9.11.x = 9.11.9 fail to restrict...

6.5CVSS4.9AI score0.00272EPSS
SaveExploits0References2
Packet Storm News
Packet Storm News
added 2025/04/20 12:00 a.m.7 views

Breaking the Prompt Wall (I): a Real-World Case Study of Attacking ChatGPT Via Lightweight Prompt Injection

Whitepaper called Breaking The Prompt Wall I: A Real-World Case Study Of Attacking ChatGPT Via Lightweight Prompt Injection...

7AI score
SaveExploits0
RedhatCVE
RedhatCVE
added 2025/04/17 1:35 p.m.22 views

CVE-2025-3579

In versions prior to Aidex 1.7, an authenticated malicious user, taking advantage of an open registry, could execute unauthorised commands within the system. This includes executing operating system Unix commands, interacting with internal services such as PHP or MySQL, and even invoking native...

9.3CVSS7.6AI score0.00609EPSS
SaveExploits0References3
Github Security Blog
Github Security Blog
added 2025/04/16 12:31 p.m.21 views

Mattermost doesn't restrict domains LLM can request to contact upstream

Mattermost versions 10.4.x = 10.4.2, 10.5.x = 10.5.0, 9.11.x = 9.11.9 fail to restrict domains the LLM can request to contact upstream which allows an authenticated user to exfiltrate data from an arbitrary server accessible to the victim via performing a prompt injection in the AI plugin's Jira...

6.5CVSS6.7AI score0.00272EPSS
SaveExploits0References4Affected Software1
OSV
OSV
added 2025/04/16 12:31 p.m.11 views

GHSA-9H6J-4FFX-CM84 Mattermost doesn't restrict domains LLM can request to contact upstream

Mattermost versions 10.4.x = 10.4.2, 10.5.x = 10.5.0, 9.11.x = 9.11.9 fail to restrict domains the LLM can request to contact upstream which allows an authenticated user to exfiltrate data from an arbitrary server accessible to the victim via performing a prompt injection in the AI plugin's Jira...

3CVSS6.7AI score0.00272EPSS
SaveExploits0References4
NVD
NVD
added 2025/04/16 10:15 a.m.41 views

CVE-2025-31363

Mattermost versions 10.4.x = 10.4.2, 10.5.x = 10.5.0, 9.11.x = 9.11.9 fail to restrict domains the LLM can request to contact upstream which allows an authenticated user to exfiltrate data from an arbitrary server accessible to the victim via performing a prompt injection in the AI plugin's Jira...

6.5CVSS0.00272EPSS
SaveExploits0References1
CVE
CVE
added 2025/04/16 9:14 a.m.235 views

CVE-2025-31363

Mattermost CVE-2025-31363 affects Mattermost Server in versions 10.4.x &lt;= 10.4.2, 10.5.x &lt;= 10.5.0, and 9.11.x

6.5CVSS4AI score0.00272EPSS
SaveExploits0References1Affected Software1
OSV
OSV
added 2025/04/16 9:14 a.m.10 views

CVE-2025-31363 Data exfiltration via AI plugin Jira tool

Mattermost versions 10.4.x = 10.4.2, 10.5.x = 10.5.0, 9.11.x = 9.11.9 fail to restrict domains the LLM can request to contact upstream which allows an authenticated user to exfiltrate data from an arbitrary server accessible to the victim via performing a prompt injection in the AI plugin's Jira...

3CVSS6AI score
SaveExploits0References3
Cvelist
Cvelist
added 2025/04/16 9:14 a.m.54 views

CVE-2025-31363 Data exfiltration via AI plugin Jira tool

Mattermost versions 10.4.x = 10.4.2, 10.5.x = 10.5.0, 9.11.x = 9.11.9 fail to restrict domains the LLM can request to contact upstream which allows an authenticated user to exfiltrate data from an arbitrary server accessible to the victim via performing a prompt injection in the AI plugin's Jira...

3CVSS0.00272EPSS
SaveExploits0References1
Packet Storm News
Packet Storm News
added 2025/04/16 12:00 a.m.9 views

Bypassing Prompt Injection and Jailbreak Detection in LLM Guardrails

Large Language Models LLMs guardrail systems are designed to protect against prompt injection and jailbreak attacks. However, they remain vulnerable to evasion techniques. We demonstrate two approaches for bypassing LLM prompt injection and jailbreak detection systems via traditional character...

7.4AI score
SaveExploits0
CNNVD
CNNVD
added 2025/04/16 12:00 a.m.9 views

Mattermost 安全漏洞

Mattermost is an open source collaboration platform from Mattermost, Inc. in the United States. Mattermost suffers from an information disclosure vulnerability. The vulnerability stems from an under-restricted LLM request domain. An attacker can exploit the vulnerability to perform prompt injecti...

6.5CVSS4AI score0.00272EPSS
SaveExploits0References1
NVD
NVD
added 2025/04/15 9:15 a.m.20 views

CVE-2025-3579

In versions prior to Aidex 1.7, an authenticated malicious user, taking advantage of an open registry, could execute unauthorised commands within the system. This includes executing operating system Unix commands, interacting with internal services such as PHP or MySQL, and even invoking native...

9.3CVSS0.00609EPSS
SaveExploits0References1
Rows per page
Query Builder