Skip to content
Threat Feed

Threat briefs

Use the filters to narrow the full brief archive. Filtered views, including Last 24 hours, are encoded in the URL so they can be bookmarked or shared.

October 2026 (30)

critical advisory

Apache Struts: File upload remote code execution (CVE-2024-53677)

A critical remote code execution vulnerability (CVE-2024-53677) in Apache Struts versions 2.0.0 through 6.3.0.2 allows attackers to leverage path traversal during file uploads to execute arbitrary code.

Struts apache-struts rce file-upload web-application-attack
1t 1c updated
low advisory

ACF City Selector: Arbitrary file upload remote code execution (CVE-2024-56264)

An arbitrary file upload vulnerability in the WordPress ACF City Selector plugin (<= 1.14.0) allows an authenticated administrator to achieve remote code execution by uploading a malicious PHP shell.

ACF City Selector wordpress web-shell rce file-upload
1r 2t 1c
high advisory

BlogEngine.NET: Arbitrary File Upload Remote Code Execution (CVE-2023-33404)

BlogEngine.NET versions 3.3.8.0 and earlier are vulnerable to arbitrary file upload via the /api/upload endpoint, allowing authenticated users with EditOwnPosts permissions to achieve remote code execution.

BlogEngine.NET
1r 2t 1c
high advisory

SQL Injection Vulnerability in BA-Workflow

An authenticated SQL injection vulnerability in YS-infoware BA-Workflow allows remote attackers to execute arbitrary database commands, compromising data integrity and confidentiality.

BA-Workflow sql-injection vulnerability
1t 1c
high advisory

Ghostscript -dSAFER Bypass via CVE-2026-112177

An incomplete security fix in Ghostscript 9.27 on RHEL 8 allows attackers to bypass -dSAFER restrictions, leading to arbitrary file access or command execution via crafted PostScript files.

Ghostscript vulnerability linux
1t 2c
high advisory

Insecure Deserialization in Uniong WebITR

Uniong WebITR contains an insecure deserialization vulnerability, CVE-2026-113404, allowing authenticated remote attackers to execute arbitrary code via malicious serialized data.

WebITR web-vulnerability rce deserialization
1t 1c
low advisory

Privilege Escalation in Webkul Krayin CRM via SVG XSS

CVE-2025-3568 allows an authenticated low-privileged user to perform an administrative account takeover in Krayin CRM 2.0.0 and 2.1.0 by leveraging SVG-based XSS to bypass password verification controls.

Krayin CRM web-application privilege-escalation xss csrf
1r 2t 1c
high advisory

IMAP Command Injection in Roundcube Webmail

Roundcube Webmail versions prior to 1.6.20 and 1.7.5 are vulnerable to IMAP command injection via CRLF sequences in the _uid parameter, allowing authenticated remote attackers to execute arbitrary IMAP commands.

Webmail web-vulnerability imap injection
1r 1t 1c
critical advisory

Unauthenticated Remote Code Execution in Mercury via /api/kernels

The Mercury server defaults to unauthenticated access and uses an ineffective blocklist mechanism, enabling remote, unauthenticated attackers to execute arbitrary Python code via the /api/kernels websocket endpoint.

PoC mercury remote-code-execution web-application-vulnerability jupyter
2t 1c updated
low advisory

Denial of Service via HTTP/2 Resource Exhaustion in Pingora

A vulnerability in Pingora versions prior to 0.8.1 allows unauthenticated remote attackers to trigger out-of-memory termination via HTTP/2 header-list and stream resource exhaustion.

Pingora denial-of-service vulnerability network
1t 1c
critical advisory

Insecure Default Configuration in Letta Leading to Unauthenticated RCE

Letta versions through 0.16.8 ship with a insecure compose.yaml that disables authentication middleware and exposes a PostgreSQL database with default credentials, enabling remote code execution and administrative account takeover.

Letta configuration-error remote-code-execution web-application authentication-bypass cloud webserver
4t 1c
high advisory

Insecure Direct Object Reference in Coze Studio workflow.go

Coze Studio versions up to 0.5.1 are vulnerable to an IDOR flaw in workflow.go allowing authenticated users to perform unauthorized operations on other tenants workflows.

Coze Studio idor web-vulnerability cloud
1t 1c
high advisory

Remote Code Execution in Coze Studio via Workflow Nodes

Coze Studio versions prior to 0.5.0 allow authenticated users to execute arbitrary commands by bypassing sandbox protections in Python workflow nodes.

Coze Studio remote-code-execution sandbox-escape python
1t 1c
high advisory

CVE-2026-77930 - Approval Bypass and Command Execution in EvoScientist

EvoScientist versions 0.1.0 through 0.2.4 contain an approval bypass vulnerability enabling attackers to execute arbitrary shell commands via unsanitized content ingestion.

EvoScientist vulnerability remote-code-execution
2t 1c
high advisory

Remote Code Execution in Aix-DB ChatBI Agent

Authenticated non-admin users can execute arbitrary system commands in Aix-DB version 1.2.4 by exploiting an unvalidated execute tool within the ChatBI agent via crafted HTTP requests.

Aix-DB remote-code-execution web-application cve
1r 1t 1c
high advisory

OpenSRE Approval Gate Bypass Vulnerability

OpenSRE versions through 0.1.2026.10.19 are vulnerable to an approval gate bypass that enables arbitrary command execution via manipulated alert or messaging inputs.

OpenSRE vulnerability rce cve-2026-109444
2t 1c
high advisory

Remote Code Execution in JoySafeter via Unsandboxed Execute Tool

JoySafeter version 1.0.0 and earlier contains a remote code execution vulnerability allowing authenticated users to execute shell commands via an unsandboxed agent tool.

JoySafeter remote-code-execution cloud-security vulnerability
2t 1c
high advisory

Indirect Prompt Injection and Remote Code Execution in LangAlpha

LangAlpha versions through 2026.09.07 are vulnerable to indirect prompt injection that permits remote attackers to execute arbitrary code and exfiltrate credentials via insecure agent functions.

LangAlpha prompt-injection rce ai-security
3t 1c
critical advisory

Unauthenticated RCE in DB-GPT via Insecure Mock Authentication

DB-GPT versions 0.8.2 and earlier contain an unauthenticated remote code execution vulnerability in the ReAct agent endpoint that allows attackers to execute arbitrary system commands as root.

DB-GPT remote-code-execution web-application container-security vulnerability credential-exposure mysql
1r 3t 1c
critical advisory

Unauthenticated Remote Code Execution in Langchain-Chatchat

Langchain-Chatchat versions up to 0.3.1.3 contain a critical vulnerability in the /tools/call endpoint that permits unauthenticated remote code execution via arbitrary tool invocation.

Langchain-Chatchat
1r 2t 1c
critical advisory

Remote Code Execution in piflow DeepAgents

An unauthenticated remote code execution vulnerability in piflow versions 2.0.0 through 2.0.1 allows attackers to execute arbitrary system commands via the DeepAgents chat agent.

piflow rce vulnerability cve-2026-109442
1r 2t 1c
medium advisory

CVE-2026-109395 - Out-of-Bounds Memory Access in Ning compress-lzf

Ning compress-lzf versions prior to 1.2.0 are vulnerable to an out-of-bounds memory access flaw in the UnsafeChunkDecoder, which can be exploited via crafted LZF streams to leak JVM heap memory or cause application crashes.

compress-lzf vulnerability memory-safety java
1c
high advisory

Arbitrary Code Execution in Mozilla Sccache via LD_PRELOAD

Mozilla Sccache versions prior to 0.4.0 are vulnerable to arbitrary code execution due to improper handling of the LD_PRELOAD environment variable when invoking compilers.

Sccache vulnerability code-execution build-security
2t 1c
low advisory

Arbitrary File Read in Apache Superset via CVE-2024-34693

CVE-2024-34693 is an arbitrary file read vulnerability in Apache Superset (versions < 3.1.3 and 4.0.1) that allows authenticated attackers to leverage rogue MariaDB servers and the LOAD DATA LOCAL INFILE command to exfiltrate sensitive files.

Superset +1 apache cve-2024-34693 arbitrary-file-read authenticated-access
1t 1c
medium advisory

Denial of Service Vulnerability in FFmpeg avpriv_h264_has_num_reorder_frames

A NULL pointer dereference vulnerability in FFmpeg version N-119856-gbe46370941 allows remote attackers to trigger a Denial of Service through specially crafted media files.

FFmpeg denial-of-service vulnerability
high advisory

Privilege Escalation in IBM Verify Identity and Security Access via Hard-coded Credentials

IBM Verify Identity Access 11.0.3.1 and IBM Security Verify Access 10.0.9.3 contain hard-coded credentials that allow a physically proximate attacker to gain elevated administrative privileges.

Verify Identity Access +1 privilege-escalation identity-management hardware
1t 1c
high advisory

IBM Langflow OSS Authorization Bypass Vulnerability

IBM Langflow OSS versions 1.0.0 through 1.12.4 are vulnerable to an authorization bypass in message endpoints that enables unauthorized session hijacking, message deletion, and prompt injection.

Langflow OSS web-application authentication-bypass authorization-bypass cve-2026-107366
high advisory

Privilege Escalation Vulnerability in IBM Sterling Connect Direct for Unix

IBM Sterling Connect Direct for Unix is susceptible to local privilege escalation due to improper privilege management in versions 6.3.0.0 through 6.4.0.6, identified as CVE-2026-105453.

Sterling Connect Direct for Unix privilege-escalation vulnerability local-attack
1t
high advisory

Remote Backdoor Vulnerability in Netis EAP930

A vulnerability in the Netis EAP930 CGI endpoint allows remote attackers to trigger a backdoor via crafted manipulation of the SIGNATURE argument in the /cgi-bin/debug file.

EAP930 vulnerability remote-code-execution network-infrastructure
2t 1c
high advisory

Integer Overflow Vulnerability in GIMP file-pvr Plug-in

An integer overflow vulnerability in the GIMP file-pvr plug-in allows heap-based buffer overflow via malicious .pvr files, potentially leading to arbitrary code execution or application crashes.

GIMP vulnerability remote-code-execution memory-corruption cve-2026-112258 cve-2026-112261 code-execution
2t 1c updated