VYPR
← All advisories
Moderate severityNVD Advisory· Published Mar 20, 2025· Updated Oct 15, 2025

Denial of Service in run-llama/llama_index

CVE-2024-12910

Description

A vulnerability in the KnowledgeBaseWebReader class of the run-llama/llama_index repository, version latest, allows an attacker to cause a Denial of Service (DoS) by controlling a URL variable to contain the root URL. This leads to infinite recursive calls to the get_article_urls method, exhausting system resources and potentially crashing the application.

AI Insight

LLM-synthesized narrative grounded in this CVE's description and references.

Controlling a URL variable to contain the root URL in llama_index's KnowledgeBaseWebReader causes infinite recursive calls, leading to Denial of Service.

Vulnerability

Overview

CVE-2024-12910 describes a Denial of Service (DoS) vulnerability in the KnowledgeBaseWebReader class of the run-llama/llama_index repository [1][2]. The issue arises when an attacker can control a URL variable to contain the root URL. This causes the get_article_urls method to enter an infinite recursive loop, exhausting system resources and potentially crashing the application [2][3].

Exploitation

Method

An attacker can exploit this vulnerability by providing a crafted URL that points to the root of the knowledge base. No special authentication is mentioned as a prerequisite; the attacker only needs to be able to supply a URL to the reader. When the load_data method is called, it invokes get_article_urls with the attacker-controlled root URL. Because the method lacks a maximum depth or loop detection, it recursively calls itself indefinitely when the current URL equals the root URL [3].

Impact

Successful exploitation results in a Denial of Service condition. The uncontrolled recursion consumes CPU and memory resources, which can degrade performance or crash the application, making it unavailable for legitimate users [2].

Mitigation

Status

The vulnerability has been addressed in a commit [3] that adds a max_depth parameter (defaulting to 100) to get_article_urls. This change limits recursion depth and prevents infinite loops. Users should update to a version of llama_index that includes this fix. Additionally, the PyPI advisory database references this issue [4], confirming that a patched version is advised.

References
  1. GitHub - run-llama/llama_index: LlamaIndex is the leading document agent and OCR platform
  2. NVD - CVE-2024-12910
  3. fix: prevent infinite recursion in get_article_urls (#17360) · run-llama/llama_index@159ce48
  4. advisory-database/vulns/llama-index/PYSEC-2025-11.yaml at main · pypa/advisory-database

AI Insight generated on May 20, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.

Affected packages

Versions sourced from the GitHub Security Advisory.

PackageAffected versionsPatched versions
llama-indexPyPI
< 0.12.90.12.9

Affected products

3

Patches

1
159ce485a116

fix: prevent infinite recursion in get_article_urls (#17360)

https://github.com/run-llama/llama_indexMassimiliano PippiDec 24, 2024via ghsa
commit
2 files changed · +14 −9
  • llama-index-integrations/readers/llama-index-readers-web/llama_index/readers/web/knowledge_base/base.py+13 −8 modified
    @@ -5,7 +5,8 @@
 
 
 class KnowledgeBaseWebReader(BaseReader):
-    """Knowledge base reader.
+    """
+    Knowledge base reader.
 
     Crawls and reads articles from a knowledge base/help center with Playwright.
     Tested on Zendesk and Intercom CMS, may work on others.
@@ -36,6 +37,7 @@ def __init__(
         title_selector: Optional[str] = None,
         subtitle_selector: Optional[str] = None,
         body_selector: Optional[str] = None,
+        max_depth: int = 100,
     ) -> None:
         """Initialize with parameters."""
         self.root_url = root_url
@@ -44,6 +46,7 @@ def __init__(
         self.title_selector = title_selector
         self.subtitle_selector = subtitle_selector
         self.body_selector = body_selector
+        self.max_depth = max_depth
 
     def load_data(self) -> List[Document]:
         """Load data from the knowledge base."""
@@ -54,9 +57,7 @@ def load_data(self) -> List[Document]:
 
             # Crawl
             article_urls = self.get_article_urls(
-                browser,
-                self.root_url,
-                self.root_url,
+                browser, self.root_url, self.root_url, self.max_depth
             )
 
             # Scrape
@@ -82,7 +83,8 @@ def scrape_article(
         browser: Any,
         url: str,
     ) -> Dict[str, str]:
-        """Scrape a single article url.
+        """
+        Scrape a single article url.
 
         Args:
             browser (Any): a Playwright Chromium browser.
@@ -125,9 +127,10 @@ def scrape_article(
         return {"title": title, "subtitle": subtitle, "body": body, "url": url}
 
     def get_article_urls(
-        self, browser: Any, root_url: str, current_url: str
+        self, browser: Any, root_url: str, current_url: str, max_depth: int = 100
     ) -> List[str]:
-        """Recursively crawl through the knowledge base to find a list of articles.
+        """
+        Recursively crawl through the knowledge base to find a list of articles.
 
         Args:
             browser (Any): a Playwright Chromium browser.
@@ -158,7 +161,9 @@ def get_article_urls(
 
         for link in links:
             url = root_url + page.evaluate("(node) => node.getAttribute('href')", link)
-            article_urls.extend(self.get_article_urls(browser, root_url, url))
+            article_urls.extend(
+                self.get_article_urls(browser, root_url, url, max_depth)
+            )
 
         page.close()
  • llama-index-integrations/readers/llama-index-readers-web/pyproject.toml+1 −1 modified
    @@ -45,7 +45,7 @@ license = "MIT"
 maintainers = ["HawkClaws", "Hironsan", "NA", "an-bluecat", "bborn", "jasonwcfan", "kravetsmic", "pandazki", "ruze00", "selamanse", "thejessezhang"]
 name = "llama-index-readers-web"
 readme = "README.md"
-version = "0.3.2"
+version = "0.3.3"
 
 [tool.poetry.dependencies]
 python = ">=3.9,<4.0"

Vulnerability mechanics

Generated on May 9, 2026. Inputs: CWE entries + fix-commit diffs from this CVE's patches. Citations validated against bundle.

References

5

News mentions

0

No linked articles in our index yet.