AnythingLLM is an application that turns pieces of content into context that any LLM can use as references during chatting. In 1.11.1 and earlier, AnythingLLM Desktop contains a Streaming Phase XSS vulnerability in the chat rendering pipeline that escalates to Remote Code Execution on the host OS due to insecure Electron configuration. This works with default settings and requires no user interaction beyond normal chat usage. The custom markdown-it image renderer in frontend/src/utils/chat/markdown.js interpolates token.content directly into the alt attribute without HTML entity escaping. The PromptReply component renders this output via dangerouslySetInnerHTML without DOMPurify sanitization — unlike HistoricalMessage which correctly applies DOMPurify.sanitize().
LangChain is a framework for building LLM-powered applications. Prior to version 1.1.8, a redirect-based Server-Side Request Forgery (SSRF) bypass exists in `RecursiveUrlLoader` in `@langchain/community`. The loader validates the initial URL but allows the underlying fetch to follow redirects automatically, which permits a transition from a safe public URL to an internal or metadata endpoint without revalidation. This is a bypass of the SSRF protections introduced in 1.1.14 (CVE-2026-26019). Users should upgrade to `@langchain/community` 1.1.18, which validates every redirect hop by disabling automatic redirects and re-validating `Location` targets before following them. In this version, automatic redirects are disabled (`redirect: "manual"`), each 3xx `Location` is resolved and validated with `validateSafeUrl()` before the next request, and a maximum redirect limit prevents infinite loops.
New API is a large language mode (LLM) gateway and artificial intelligence (AI) asset management system. Prior to version 0.10.8-alpha.9, a potential unsafe operation occurs in component `MarkdownRenderer.jsx`, allowing for Cross-Site Scripting(XSS) when the model outputs items containing `<script>` tag. Version 0.10.8-alpha.9 fixes the issue.
LangChain is a framework for building LLM-powered applications. Prior to 1.1.14, the RecursiveUrlLoader class in @langchain/community is a web crawler that recursively follows links from a starting URL. Its preventOutside option (enabled by default) is intended to restrict crawling to the same site as the base URL. The implementation used String.startsWith() to compare URLs, which does not perform semantic URL validation. An attacker who controls content on a crawled page could include links to domains that share a string prefix with the target, causing the crawler to follow links to attacker-controlled or internal infrastructure. Additionally, the crawler performed no validation against private or reserved IP addresses. A crawled page could include links targeting cloud metadata services, localhost, or RFC 1918 addresses, and the crawler would fetch them without restriction. This vulnerability is fixed in 1.1.14.
llama.cpp is an inference of several LLM models in C/C++. In commits 55d4206c8 and prior, the n_discard 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 fills up, llama_memory_seq_rm/add receives a reversed range and negative offset, causing out-of-bounds memory writes in the token evaluation loop. This deterministic memory corruption can crash the process or enable remote code execution (RCE). There is no fix at the time of publication.
GitHub Kanban MCP Server is a Model Context Protocol (MCP) server for managing GitHub issues in Kanban board format and streamlining LLM task management. Version 0.3.0 of the MCP Server is written in a way that is vulnerable to command injection vulnerability attacks as part of some of its MCP Server tool definition and implementation. The MCP Server exposes the tool `add_comment` which relies on Node.js child process API `exec` to execute the GitHub (`gh`) command, is an unsafe and vulnerable API if concatenated with untrusted user input. As of time of publication, no known patches are available.
iOS Simulator MCP Server (ios-simulator-mcp) is a Model Context Protocol (MCP) server for interacting with iOS simulators. Versions prior to 1.3.3 are written in a way that is vulnerable to command injection vulnerability attacks as part of some of its MCP Server tool definition and implementation. The MCP Server exposes the tool `ui_tap` which relies on Node.js child process API `exec` which is an unsafe and vulnerable API if concatenated with untrusted user input. LLM exposed user input for `duration`, `udid`, and `x` and `y` args can be replaced with shell meta-characters like `;` or `&&` or others to change the behavior from running the expected command `idb` to another command. When LLMs are tricked through prompt injection (and other techniques and attack vectors) to call the tool with input that uses special shell characters such as `; rm -rf /tmp;#` and other payload variations, the full command-line text will be interepted by the shell and result in other commands except of `ps` executing on the host running the MCP Server. Version 1.3.3 contains a patch for the issue.
AstrBot is a large language model chatbot and development framework. A path traversal vulnerability present in versions 3.4.4 through 3.5.12 may lead to information disclosure, such as API keys for LLM providers, account passwords, and other sensitive data. The vulnerability has been addressed in Pull Request #1676 and is included in version 3.5.13. As a workaround, users can edit the `cmd_config.json` file to disable the dashboard feature as a temporary workaround. However, it is strongly recommended to upgrade to version v3.5.13 or later to fully resolve this issue.
The desktop application in Dot through 0.9.3 allows XSS and resultant command execution because user input and LLM output are appended to the DOM with innerHTML (in render.js), and because the Electron window can access Node.js APIs.
A vulnerability in mintplex-labs/anything-llm prior to version 1.2.2 allows for Prisma injection. The issue exists in the API endpoint "/embed/:embedId/stream-chat" where user-provided JSON is directly taken to the Prisma library's where clause. An attacker can exploit this by providing a specially crafted JSON object, such as {"sessionId":{"not":"a"}}, causing Prisma to return all data from the table. This can lead to unauthorized access to all user queries in embedded chat mode.
A vulnerability was found in the ilab model serve component, where improper handling of the best_of parameter in the vllm JSON web API can lead to a Denial of Service (DoS). The API used for LLM-based sentence or chat completion accepts a best_of parameter to return the best completion from several options. When this parameter is set to a large value, the API does not handle timeouts or resource exhaustion properly, allowing an attacker to cause a DoS by consuming excessive system resources. This leads to the API becoming unresponsive, preventing legitimate users from accessing the service.