Researchers Highlight Google's Gemini AI Susceptibility to LLM Threats

Google's Gemini large language model LLM is susceptible to security threats that could cause it to divulge system prompts, generate harmful content, and carry out indirect injection attacks. The findings come from HiddenLayer, which said the issues impact consumers using Gemini Advanced with Goog...

A Taxonomy of Prompt Injection Attacks

Researchers ran a global prompt hacking competition, and have documented the results in a paper that both gives a lot of good examples and tries to organize a taxonomy of effective prompt injection strategies. It seems as if the most common successful strategy is the "compound instruction attack,...

Function Calling in Java and Spring AI using the latest Mistral AI API

UPDATE: As of March 13, 2024, Mistral AI has integrated support for parallel function calling into their large model, a feature that was absent at the time of this blog's initial publication. Mistral AI, a leading developer of open-source large language models, unveiled the addition of Function...

This Week in Spring - March 5th, 2024

Hi, Spring fans! Welcome to another exciting roundup of This Week in Spring! I expect many of you are reading this for the first time, especially with Facebook and Instagram being down. People have been exploring all the other lesser-known corners of the web, looking for their daily "doom scroll....

a-data-processing (0.0.1), ab-data-processing (0.0.1) +430 more potentially affected by CVE-2024-28088 via langchain (>=0.0.27, <=0.1.10)

langchain PYPI version =0.0.27, =0.1.7, =0.2.1, =0.1.0, =0.1.0, =0.0.6.dev1, =0.1.5, =0.0.2, =0.0.1, =0.0.1, =0.0.1, =0.1.0 and more Source cves: CVE-2024-28088 Source advisory: OSV:PYSEC-2024-43...

CVE-2024-0798

A privilege escalation vulnerability exists in mintplex-labs/anything-llm, allowing users with 'default' role to delete documents uploaded by 'admin'. Despite the intended restriction that prevents 'default' role users from deleting admin-uploaded documents, an attacker can exploit this...

CVE-2024-0798 Privilege Escalation in mintplex-labs/anything-llm

A privilege escalation vulnerability exists in mintplex-labs/anything-llm, allowing users with 'default' role to delete documents uploaded by 'admin'. Despite the intended restriction that prevents 'default' role users from deleting admin-uploaded documents, an attacker can exploit this...

CVE-2024-0798: Least Privilege Violation

A privilege escalation vulnerability exists in mintplex-labs/anything-llm, allowing users with 'default' role to delete documents uploaded by 'admin'. Despite the intended restriction that prevents 'default' role users from deleting admin-uploaded documents, an attacker can exploit this...

aqueduct-llm (0.0.1), fschat (>=0.1.1, <=0.2.15) +3 more potentially affected by CVE-2024-1729 via gradio (3.23.0)

gradio PYPI version 3.23.0 is affected by a known vulnerability. The following packages have a transitive dependency on gradio and may be impacted: - aqueduct-llm 0.0.1 - fschat =0.1.1, =0.0.1, =0.0.16 Source cves: CVE-2024-1729 Source advisory: OSV:GHSA-HMX6-R76C-85G9...

aqueduct-llm (0.0.1), fschat (>=0.1.1, <=0.2.15) +3 more potentially affected by CVE-2024-0964 via gradio (3.23.0)

gradio PYPI version 3.23.0 is affected by a known vulnerability. The following packages have a transitive dependency on gradio and may be impacted: - aqueduct-llm 0.0.1 - fschat =0.1.1, =0.0.1, =0.0.16 Source cves: CVE-2024-0964 Source advisory: OSV:GHSA-F3H9-8PHC-6GVH...

aqueduct-llm (0.0.1), fschat (>=0.1.1, <=0.2.15) +3 more potentially affected by CVE-2023-51449 via gradio (3.23.0)

gradio PYPI version 3.23.0 is affected by a known vulnerability. The following packages have a transitive dependency on gradio and may be impacted: - aqueduct-llm 0.0.1 - fschat =0.1.1, =0.0.1, =0.0.16 Source cves: CVE-2023-51449 Source advisory: OSV:PYSEC-2023-249...

Data Exfiltration Using Indirect Prompt Injection

Interesting attack on a LLM: In Writer, users can enter a ChatGPT-like session to edit or create their documents. In this chat session, the LLM can retrieve information from sources on the web to assist users in creation of their documents. We show that attackers can prepare websites that, when a...

aqueduct-llm (0.0.1), fschat (>=0.1.1, <=0.2.15) +3 more potentially affected by CVE-2023-6572 via gradio (3.23.0)

gradio PYPI version 3.23.0 is affected by a known vulnerability. The following packages have a transitive dependency on gradio and may be impacted: - aqueduct-llm 0.0.1 - fschat =0.1.1, =0.0.1, =0.0.16 Source cves: CVE-2023-6572 Source advisory: OSV:GHSA-GQVF-3HGP-5HXV...

aqueduct-llm (0.0.1), fschat (>=0.1.1, <=0.2.15) +3 more potentially affected by CVE-2023-6572 via gradio (3.23.0)

gradio PYPI version 3.23.0 is affected by a known vulnerability. The following packages have a transitive dependency on gradio and may be impacted: - aqueduct-llm 0.0.1 - fschat =0.1.1, =0.0.1, =0.0.16 Source cves: CVE-2023-6572 Source advisory: OSV:PYSEC-2023-255...

wafl-llm (>=0.0.80, <=0.1.0) potentially affected by CVE-2023-48299 via torchserve (0.7.1)

torchserve PYPI version 0.7.1 is affected by a known vulnerability. The following packages have a transitive dependency on torchserve and may be impacted: - wafl-llm =0.0.80, =0.1.0 Source cves: CVE-2023-48299 Source advisory: MEND:CVE-2023-48299...

wafl-llm (>=0.0.80, <=0.1.0) potentially affected by CVE-2023-48299 via torchserve (0.7.1)

torchserve PYPI version 0.7.1 is affected by a known vulnerability. The following packages have a transitive dependency on torchserve and may be impacted: - wafl-llm =0.0.80, =0.1.0 Source cves: CVE-2023-48299 Source advisory: OSV:GHSA-M2MJ-PR4F-H9JP...

wafl-llm (>=0.0.80, <=0.1.0) potentially affected by CVE-2023-48299 via torchserve (0.7.1)

torchserve PYPI version 0.7.1 is affected by a known vulnerability. The following packages have a transitive dependency on torchserve and may be impacted: - wafl-llm =0.0.80, =0.1.0 Source cves: CVE-2023-48299 Source advisory: GHSA-M2MJ-PR4F-H9JP...

CVE-2023-5833

Improper Access Control in GitHub repository mintplex-labs/anything-llm prior to 0.1.0...

CVE-2023-5833 Improper Access Control in mintplex-labs/anything-llm

Improper Access Control in GitHub repository mintplex-labs/anything-llm prior to 0.1.0...

CVE-2023-5833

The connected Huntr document provides concrete details for CVE-2023-5833: an improper access control flaw in mintplex-labs/anything-llm prior to 0.1.0 that allows overwriting backend environment variables via the /api/system/update-env endpoint. The vulnerability arises from how KEY_MAPPING expos...