Directory Traversal

Overview llama-index-readers-obsidian is a llama-index readers obsidian integration Affected versions of this package are vulnerable to Directory Traversal via the ObsidianReader process. An attacker can access arbitrary files outside the intended directory by creating symbolic links that point t...

Uncontrolled Recursion

Overview llama-index is an Interface between LLMs and your data Affected versions of this package are vulnerable to Uncontrolled Recursion via the JSONReader process. An attacker can cause the application to crash by submitting deeply nested JSON structures, resulting in a stack overflow and...

PYSEC-2025-252

A vulnerability in the ObsidianReader class of the run-llama/llamaindex repository, specifically in version 0.12.27, allows for hardlink-based path traversal. This flaw permits attackers to bypass path restrictions and access sensitive system files, such as /etc/passwd, by exploiting hardlinks. T...

PYSEC-2025-251

The JSONReader in run-llama/llamaindex versions 0.12.28 is vulnerable to a stack overflow due to uncontrolled recursive JSON parsing. This vulnerability allows attackers to trigger a Denial of Service DoS by submitting deeply nested JSON structures, leading to a RecursionError and crashing...

CVE-2025-3225

An XML Entity Expansion vulnerability, also known as a 'billion laughs' attack, exists in the sitemap parser of the run-llama/llamaindex repository, specifically affecting version v0.12.21. This vulnerability allows an attacker to supply a malicious Sitemap XML, leading to a Denial of Service DoS...

PYSEC-2025-250

An XML Entity Expansion vulnerability, also known as a 'billion laughs' attack, exists in the sitemap parser of the run-llama/llamaindex repository, specifically affecting version v0.12.21. This vulnerability allows an attacker to supply a malicious Sitemap XML, leading to a Denial of Service DoS...

CVE-2025-3044

A vulnerability in the ArxivReader class of the run-llama/llamaindex repository, versions up to v0.12.22.post1, allows for MD5 hash collisions when generating filenames for downloaded papers. This can lead to data loss as papers with identical titles but different contents may overwrite each othe...

CVE-2025-6210 Hardlink-Based Path Traversal in run-llama/llama_index

A vulnerability in the ObsidianReader class of the run-llama/llamaindex repository, specifically in version 0.12.27, allows for hardlink-based path traversal. This flaw permits attackers to bypass path restrictions and access sensitive system files, such as /etc/passwd, by exploiting hardlinks. T...

CVE-2025-3046 Path Traversal via Symbolic Links in run-llama/llama_index

A vulnerability in the ObsidianReader class of the run-llama/llamaindex repository, versions 0.12.23 to 0.12.28, allows for arbitrary file read through symbolic links. The ObsidianReader fails to resolve symlinks to their real paths and does not validate whether the resolved paths lie within the...

CVE-2025-3044 MD5 Hash Collision in run-llama/llama_index

A vulnerability in the ArxivReader class of the run-llama/llamaindex repository, versions up to v0.12.22.post1, allows for MD5 hash collisions when generating filenames for downloaded papers. This can lead to data loss as papers with identical titles but different contents may overwrite each othe...

CVE-2025-3225 XML Entity Expansion vulnerability in run-llama/llama_index

An XML Entity Expansion vulnerability, also known as a 'billion laughs' attack, exists in the sitemap parser of the run-llama/llamaindex repository, specifically affecting version v0.12.21. This vulnerability allows an attacker to supply a malicious Sitemap XML, leading to a Denial of Service DoS...

CVE-2025-3225 XML Entity Expansion vulnerability in run-llama/llama_index

An XML Entity Expansion vulnerability, also known as a 'billion laughs' attack, exists in the sitemap parser of the run-llama/llamaindex repository, specifically affecting version v0.12.21. This vulnerability allows an attacker to supply a malicious Sitemap XML, leading to a Denial of Service DoS...

CVE-2025-3225 XML Entity Expansion vulnerability in run-llama/llama_index

An XML Entity Expansion vulnerability, also known as a 'billion laughs' attack, exists in the sitemap parser of the run-llama/llamaindex repository, specifically affecting version v0.12.21. This vulnerability allows an attacker to supply a malicious Sitemap XML, leading to a Denial of Service DoS...

CVE-2025-3225

The CVE-2025-3225 issue affects the run-llama/llama_index project, specifically its sitemap parser, where an XML Entity Expansion (billion laughs) vulnerability exists in version v0.12.21. This can be triggered by a malicious Sitemap XML to cause Denial of Service via memory exhaustion, potential...

PT-2025-28163

Name of the Vulnerable Software and Affected Versions: run-llama/llama index versions 0.12.27 through 0.12.40 Description: A path traversal vulnerability exists, specifically within the encode image function in generic utils.py, allowing an attacker to manipulate the image path input to read...

PT-2025-28149

Name of the Vulnerable Software and Affected Versions: llama index versions v0.12.21 through v0.12.28 Description: An XML Entity Expansion vulnerability, also known as a 'billion laughs' attack, exists in the sitemap parser of the run-llama/llama index repository. This issue allows an attacker to...

Deserialization of Untrusted Data

Overview llama-index-core is an Interface between LLMs and your data Affected versions of this package are vulnerable to Deserialization of Untrusted Data via the JsonPickleSerializer process. An attacker can execute arbitrary code by submitting specially crafted serialized data that triggers the...

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 JsonPickleSerializer process. An attacker can execute arbitrary code by submitting specially crafted serialized data that triggers the...

PYSEC-2025-249

A critical deserialization vulnerability exists in the run-llama/llamaindex library's JsonPickleSerializer component, affecting versions v0.12.27 through v0.12.40. This vulnerability allows remote code execution due to an insecure fallback to Python's pickle module. JsonPickleSerializer prioritiz...

Model Inversion Attacks on Llama 3: Extracting PII from Large Language Models

Large language models LLMs have transformed natural language processing, but their ability to memorize training data poses significant privacy risks. This paper investigates model inversion attacks on the Llama 3.2 model, a multilingual LLM developed by Meta. By querying the model with carefully...