Claude

35 exploits Active since May 2025
CVE-2025-46725 WRITEUP CRITICAL WRITEUP
langroid < 0.53.15 - Remote Code Execution via LanceDocChatAgent QueryPlan.dataframe_calc
Langroid is a Python framework to build large language model (LLM)-powered applications. Prior to version 0.53.15, `LanceDocChatAgent` uses pandas eval() through `compute_from_docs()`. As a result, an attacker may be able to make the agent run malicious commands through `QueryPlan.dataframe_calc]`) compromising the host system. Langroid 0.53.15 sanitizes input to the affected function by default to tackle the most common attack vectors, and added several warnings about the risky behavior in the project documentation.
CVSS 9.8
CVE-2025-55750 WRITEUP MEDIUM WRITEUP
Gitpod < main-gha.33628 - Authenticated OAuth Token Exposure via Bitbucket Redirect Flow
Gitpod is a developer platform for cloud development environments. In versions before main-gha.33628 for both Gitpod Classic and Gitpod Classic Enterprise, OAuth integration with Bitbucket in certain conditions allowed a crafted link to expose a valid Bitbucket access token via the URL fragment when clicked by an authenticated user. This resulted from how Bitbucket returned tokens and how Gitpod handled the redirect flow. The issue was limited to Bitbucket (GitHub and GitLab integrations were not affected), required user interaction, and has been mitigated through redirect handling and OAuth logic hardening. The issue was resolved in main-gha.33628 and later. There are no workarounds.
CVSS 6.5
CVE-2025-58177 WRITEUP MEDIUM WRITEUP
n8n 1.24.0-1.106.9 - Authenticated Stored Cross-Site Scripting via LangChain Chat Trigger Initial Messages
n8n is an open source workflow automation platform. From 1.24.0 to before 1.107.0, there is a stored cross-site scripting (XSS) vulnerability in @n8n/n8n-nodes-langchain.chatTrigger. An authorized user can configure the LangChain Chat Trigger node with malicious JavaScript in the initialMessages field and enable public access so that the payload is executed in the browser of any user who visits the resulting public chat URL. This can be used for phishing or to steal cookies or other sensitive data from users accessing the public chat link. The issue is fixed in version 1.107.0. Updating to 1.107.0 or later is recommended. As a workaround, the affected chatTrigger node can be disabled. No other workarounds are known.
CVSS 5.4
CVE-2025-58444 WRITEUP HIGH WRITEUP
MCP Inspector < 0.16.6 - Cross-Site Scripting via Malicious Redirect URI
The MCP inspector is a developer tool for testing and debugging MCP servers. A cross-site scripting issue was reported in versions of the MCP Inspector local development tool prior to 0.16.6 when connecting to untrusted remote MCP servers with a malicious redirect URI. This could be leveraged to interact directly with the inspector proxy to trigger arbitrary command execution. Users are advised to update to 0.16.6 to resolve this issue.
CVE-2025-64712 WRITEUP CRITICAL WRITEUP
unstructured < 0.18.18 - Path Traversal and Arbitrary File Write via MSG Attachment Processing
The unstructured library provides open-source components for ingesting and pre-processing images and text documents, such as PDFs, HTML, Word docs, and many more. Prior to version 0.18.18, a path traversal vulnerability in the partition_msg function allows an attacker to write or overwrite arbitrary files on the filesystem when processing malicious MSG files with attachments. This issue has been patched in version 0.18.18.
CVSS 9.8
CVE-2025-66414 WRITEUP HIGH WRITEUP
MCP TypeScript SDK < 1.24.0 - DNS Rebinding Local Server Tool Invocation
MCP TypeScript SDK is the official TypeScript SDK for Model Context Protocol servers and clients. Prior to 1.24.0, The Model Context Protocol (MCP) TypeScript SDK does not enable DNS rebinding protection by default for HTTP-based servers. When an HTTP-based MCP server is run on localhost without authentication with StreamableHTTPServerTransport or SSEServerTransport and has not enabled enableDnsRebindingProtection, a malicious website could exploit DNS rebinding to bypass same-origin policy restrictions and send requests to the local MCP server. This could allow an attacker to invoke tools or access resources exposed by the MCP server on behalf of the user in those limited circumstances. Note that running HTTP-based MCP servers locally without authentication is not recommended per MCP security best practices. This issue does not affect servers using stdio transport. This vulnerability is fixed in 1.24.0.
CVSS 8.1
CVE-2025-68475 WRITEUP HIGH WRITEUP
Fedify < 1.6.13, 1.7.0-1.7.13, 1.8.0-1.8.14, 1.9.0-1.9.1 - Regular Expression Denial of Service in HTML Parser
Fedify is a TypeScript library for building federated server apps powered by ActivityPub. Prior to versions 1.6.13, 1.7.14, 1.8.15, and 1.9.2, a Regular Expression Denial of Service (ReDoS) vulnerability exists in Fedify's document loader. The HTML parsing regex at packages/fedify/src/runtime/docloader.ts:259 contains nested quantifiers that cause catastrophic backtracking when processing maliciously crafted HTML responses. This issue has been patched in versions 1.6.13, 1.7.14, 1.8.15, and 1.9.2.
CVSS 7.5
CVE-2026-1979 WRITEUP MEDIUM WRITEUP
mruby < 3.4.0 - Use-After-Free in JMPNOT-to-JMPIF Optimization
A flaw has been found in mruby up to 3.4.0. This affects the function mrb_vm_exec of the file src/vm.c of the component JMPNOT-to-JMPIF Optimization. Executing a manipulation can lead to use after free. The attack needs to be launched locally. The exploit has been published and may be used. This patch is called e50f15c1c6e131fa7934355eb02b8173b13df415. It is advisable to implement a patch to correct this issue.
CVSS 5.3
CVE-2026-22701 WRITEUP MEDIUM WRITEUP
filelock < 3.20.3 - TOCTOU Race Condition in SoftFileLock _acquire Method
filelock is a platform-independent file lock for Python. Prior to version 3.20.3, a TOCTOU race condition vulnerability exists in the SoftFileLock implementation of the filelock package. An attacker with local filesystem access and permission to create symlinks can exploit a race condition between the permission validation and file creation to cause lock operations to fail or behave unexpectedly. The vulnerability occurs in the _acquire() method between raise_on_not_writable_file() (permission check) and os.open() (file creation). During this race window, an attacker can create a symlink at the lock file path, potentially causing the lock to operate on an unintended target file or leading to denial of service. This issue has been patched in version 3.20.3.
CVSS 5.3
CVE-2026-25481 WRITEUP CRITICAL WRITEUP
langroid < 0.59.32 - Remote Code Execution via Pandas Eval Bypass
Langroid is a framework for building large-language-model-powered applications. Prior to version 0.59.32, there is a bypass to the fix for CVE-2025-46724. TableChatAgent can call pandas_eval tool to evaluate the expression. There is a WAF in langroid/utils/pandas_utils.py introduced to block code injection CVE-2025-46724. However it can be bypassed due to _literal_ok() returning False instead of raising UnsafeCommandError on invalid input, combined with unrestricted access to dangerous dunder attributes (__init__, __globals__, __builtins__). This allows chaining whitelisted DataFrame methods to leak the eval builtin and execute arbitrary code. This issue has been patched in version 0.59.32.
CVSS 9.6