CVE-2026-7159 is a path traversal vulnerability in the douinc mkdocs-mcp-plugin versions up to 0.4.1. It affects the read_document and list_documents functions in server.py, allowing remote attackers to manipulate the docs_dir or file_path arguments to access unauthorized filesystem locations. The vulnerability has a CVSS 4.0 base score of 6.9 (medium severity). The vendor has acknowledged the issue and stated that a fix will be published within a few days, but no patch is currently available. Exploit details have been made public, though no known exploits in the wild are reported yet.
CVE-2026-7158 is a server-side request forgery (SSRF) vulnerability in the dmitryglhf mcp-url-downloader project affecting the _validate_url_safe function in src/mcp_url_downloader/server.py. This vulnerability allows remote attackers to manipulate the URL argument, potentially causing the server to make unauthorized requests. The issue has been publicly disclosed, but no official vendor response or patch is currently available. The product uses a rolling release model, so specific affected or fixed versions are not clearly identified. The CVSS 4.0 base score is 6.9, indicating a medium severity level.
The npm package '@pyme-web/ui-base' version 99.0.4 has been identified as containing malicious code. Installation or execution of this package can lead to full compromise of the affected computer, potentially exposing all stored secrets and keys. Removal of the package does not guarantee elimination of all malicious software introduced. Immediate rotation of all secrets and keys from a separate, uncompromised machine is strongly advised.
CVE-2026-7066 is an OS command injection vulnerability in the function exec_openstack of the file server.py in choieastsea simple-openstack-mcp up to commit 767b2f4a8154cca344344b9725537a58399e6036. This vulnerability allows remote attackers to execute arbitrary OS commands. The project uses rolling releases, so specific affected and fixed versions are not clearly defined. The issue was reported early but the vendor has not responded or issued a patch. The CVSS 4.0 base score is 6.9, indicating a medium severity level. Exploit code has been publicly disclosed, but no known exploits in the wild have been confirmed.
CVE-2026-7059 is a path traversal vulnerability in the 666ghj MiroFish application versions up to 0.1.2. It affects the get_simulation_posts function in the backend/app/api/simulation.py file, specifically the Query Parameter Handler component. An attacker can remotely manipulate the Platform argument to perform path traversal. The vulnerability has a CVSS 4.0 base score of 6.9, indicating medium severity. No official patch or remediation guidance is currently available, and no known exploits are reported in the wild.
CVE-2026-41664 is an integer overflow vulnerability in Samsung Open Source ONE prior to version 1.30.0. The flaw occurs during memory copy size calculation when handling large tensor shapes, potentially causing invalid memory operations. The vulnerability has a CVSS 3.1 base score of 6.6, indicating medium severity. There is no official patch or remediation level currently confirmed by the vendor. Exploitation requires local access with limited privileges and user interaction. The impact includes limited confidentiality and integrity loss, with a high impact on availability.
A vulnerability was detected in langflow-ai langflow up to 1.8.3. The impacted element is the function get_client_ip/install_mcp_config of the file src/backend/base/langflow/api/v1/mcp_projects.py of the component Model Context Protocol Configuration API. Performing a manipulation of the argument X-Forwarded-For results in injection. The attack may be initiated remotely. The exploit is now public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
A security vulnerability has been detected in langflow-ai langflow up to 1.8.3. The affected element is the function create_project/encrypt_auth_settings of the file src/backend/base/Langflow/api/v1/projects.py of the component Project Creation Endpoint. Such manipulation of the argument auth_settings leads to cleartext storage in a file or on disk. The attack can be launched remotely. The exploit has been disclosed publicly and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
A weakness has been identified in langflow-ai langflow up to 1.8.3. Impacted is the function remove_api_keys/has_api_terms of the file src/backend/base/langflow/api/utils/core.py of the component Flow Using API. This manipulation causes unprotected storage of credentials. The attack can be initiated remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.
A security flaw has been discovered in langflow-ai langflow up to 1.1.0. This issue affects the function create_upload_file of the file src/backend/base/Langflow/api/v1/endpoints.py of the component API Endpoint. The manipulation results in unrestricted upload. It is possible to launch the attack remotely. The exploit has been released to the public and may be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.
CVE-2026-5631 is a code injection vulnerability in the assafelovic gpt-researcher software versions up to 3.4.3. It affects the extract_command_data function in the backend/server/server_utils.py file, specifically in the ws Endpoint component. The vulnerability allows remote attackers to manipulate the args argument to inject code. The issue has been publicly disclosed, but the vendor has not yet responded or provided a fix. The CVSS 4.0 base score is 6.9, indicating a medium severity level.
CVE-2026-5630 is a cross-site scripting (XSS) vulnerability in the assafelovic gpt-researcher product up to version 3.4.3. The flaw exists in an unknown function within the Report API component located in backend/server/app.py. This vulnerability can be exploited remotely and requires user interaction. Although the issue was reported early to the project, no response or fix has been provided yet. The CVSS 4.0 base score is 5.3, indicating a medium severity level.
CVE-2026-5625 is a cross-site scripting (XSS) vulnerability in the assafelovic gpt-researcher product versions up to 3.4.3. The issue arises from improper handling of the 'task' argument in the WebSocket Interface component, specifically in the file gpt_researcher/skills/researcher.py. This vulnerability can be exploited remotely and the exploit code is publicly available. The vendor has been informed but has not yet responded or provided a fix. The CVSS 4.0 base score is 5.3, indicating a medium severity risk.
The local filesystem memory tool in the Anthropic Python SDK created memory files with mode 0o666, leaving them world-readable on systems with a standard umask and world-writable in environments with a permissive umask such as many Docker base images. A local attacker on a shared host could read persisted agent state, and in containerized deployments could modify memory files to influence subsequent model behavior. Both the synchronous and asynchronous memory tool implementations were affected.
Users on the affected versions are advised to update to the latest version.
Claude SDK for Python thanks [`lucasfutures`](https://hackerone.com/lucasfutures) on HackerOne for the report.
Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to version 1.5.1, the `_read_flow` helper in `src/backend/base/langflow/api/v1/flows.py` branched on the `AUTO_LOGIN` setting to decide whether to filter by `user_id`. When `AUTO_LOGIN` was `False` (i.e., authentication was enabled), neither branch enforced an ownership check — the query returned any flow matching the given UUID regardless of who owned it. This allowed any authenticated user to read any other user's flow, including embedded plaintext API keys; modify the logic of another user's AI agents, and/or delete flows belonging to other users. The vulnerability was introduced by the conditional logic that was meant to accommodate public/example flows (those with `user_id = NULL`) under auto-login mode, but inadvertently left the authenticated path without an ownership filter. The fix in version 1.5.1 removes the `AUTO_LOGIN` conditional entirely and unconditionally scopes the query to the requesting user.
A vulnerability was detected in vanna-ai vanna up to 2.0.2. Affected by this vulnerability is the function ask of the file vanna\legacy\base\base.py. Performing a manipulation results in sql injection. The attack is possible to be carried out remotely. The exploit is now public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
CVE-2026-3954 is a path traversal vulnerability in OpenBMB XAgent version 1.0.0, specifically in the workspace function of the file XAgentServer/application/routers/workspace.py. This flaw allows remote attackers to manipulate the file_name argument to access files outside the intended directory. The vulnerability requires no authentication or user interaction and has a CVSS 4.0 base score of 6.9, indicating medium severity. Although the exploit code is publicly available, no known active exploitation has been reported. The OpenBMB project has been informed but has not yet issued a patch.
WeKnora is an LLM-powered framework designed for deep document understanding and semantic retrieval. Prior to version 0.3.0, a cross-tenant authorization bypass in the knowledge base copy endpoint allows any authenticated user to clone (duplicate) another tenant’s knowledge base into their own tenant by knowing/guessing the source knowledge base ID. This enables bulk data exfiltration (document/FAQ content) across tenants. This issue has been patched in version 0.3.0.
### Summary
The `<base>` tag passes through the default `Cleaner` configuration. While `page_structure=True` removes `html`, `head`, and `title` tags, there is no specific handling for `<base>`, allowing an attacker to inject it and hijack relative links on the page.
### Details
The `<base>` tag is not currently in the `page_structure` kill set. Even though the specification says `<base>` must be inside `<head>`, browsers accept `<base>` tags outside of the head.
If an attacker injects a `<base>` tag, it changes the base URL for all relative URLs on the page (links, images, scripts) to a domain controlled by the attacker.
### PoC
```python
from lxml_html_clean import clean_html
# The base tag is preserved in the output
result = clean_html('<base href="http://evil.com/"><a href="/account">Account</a>')
print(result)
# Output: <div><base href="http://evil.com/">...<a href="/account">Account</a></div>
```
### Impact
The injection of a `<base>` tag allows an attacker to hijack the resolution of **all** relative URLs on the page. This results in three critical attack vectors:
1. **Phishing & Redirection:** Attackers can redirect user navigation (e.g., `<a href="/login">`) and form submissions (e.g., `<form action="/auth">`) to an attacker-controlled domain, effectively stealing credentials or sensitive data without the user realizing they have left the legitimate site.
2. **Cross-Site Scripting (XSS):** If the victim application loads JavaScript files using relative paths (e.g., `<script src="assets/app.js">`), the browser will attempt to fetch the script from the attacker's domain. This upgrades the vulnerability from HTML injection to full Stored XSS.
3. **Defacement:** Relative references to images (`<img>`) and stylesheets (`<link>`) will be loaded from the attacker's server, allowing for UI redressing or defacement.
The import hook in CPython that handles legacy *.pyc files (SourcelessFileLoader) is incorrectly handled in FileLoader (a base class) and so does not use io.open_code() to read the .pyc files. sys.audit handlers for this audit event therefore do not fire.