PraisonAI Has Sandbox Escape via shell=True and Bypassable Blocklist in SubprocessSandbox

Summary SubprocessSandbox in all modes BASIC, STRICT, NETWORKISOLATED calls subprocess.run with shell=True and relies solely on string-pattern matching to block dangerous commands. The blocklist does not include sh or bash as standalone executables, allowing trivial sandbox escape in STRICT mode...

Command Injection

Overview PraisonAI is a PraisonAI is an AI Agents Framework with Self Reflection. PraisonAI application combines PraisonAI Agents, AutoGen, and CrewAI into a low-code solution for building and managing multi-agent LLM systems, focusing on simplicity, customisation, and efficient human-agent...

Arbitrary Code Injection

Overview praisonaiagents is a Praison AI agents for completing complex tasks with Self Reflection Agents Affected versions of this package are vulnerable to Arbitrary Code Injection via the executecode method. An attacker can execute arbitrary operating system commands by passing a crafted str...

CVE-2026-34528 File Browser's Signup Grants Execution Permissions When Default Permissions Includes Execution

File Browser is a file managing interface for uploading, deleting, previewing, renaming, and editing files within a specified directory. Prior to version 2.62.2, the signupHandler in File Browser applies default user permissions via d.settings.Defaults.Applyuser, then strips only Admin. The Execu...

EUVD-2026-17935

A vulnerability in the configuration backup feature of Cisco Nexus Dashboard could allow an attacker who has the encryption password and access to Full or Config-only backup files to access sensitive information. This vulnerability exists because authentication details are included in the encrypt...

CVE-2026-20160

Cisco Smart Software Manager On-Prem (SSM On-Prem) is affected by CVE-2026-20160 due to an unintended exposure of an internal service. An unauthenticated, remote attacker could send a crafted request to the exposed service API and execute commands on the underlying OS with root-level privileges. ...

CVE-2026-20160 Cisco Smart Software Manager On-Prem Arbitrary Command Execution Vulnerability

A vulnerability in Cisco Smart Software Manager On-Prem SSM On-Prem could allow an unauthenticated, remote attacker to execute arbitrary commands on the underlying operating system of an affected SSM On-Prem host. This vulnerability is due to the unintentional exposure of an internal service. An...

CVE-2026-20096 Cisco Integrated Management Controller Command Injection Vulnerability

A vulnerability in the web-based management interface of Cisco IMC could allow an authenticated, remote attacker with admin-level privileges to perform command injection attacks on an affected system and execute arbitrary commands as the root user. This vulnerability is due to improper validation...

freerdp: FreeRDP: Arbitrary code execution via heap buffer overflow in GDI surface pipeline

A flaw was found in FreeRDP, a free implementation of the Remote Desktop Protocol RDP. A malicious RDP server can exploit a heap buffer overflow vulnerability by sending a specially crafted graphics command to a FreeRDP client. This allows the server to write data outside of its intended memory...

CVE-2026-20042 Cisco Nexus Dashboard Configuration REST API Unauthorized Access Vulnerability

A vulnerability in the configuration backup feature of Cisco Nexus Dashboard could allow an attacker who has the encryption password and access to Full or Config-only backup files to access sensitive information. This vulnerability exists because authentication details are included in the encrypt...

freerdp: FreeRDP: Arbitrary code execution via heap buffer overflow in GDI surface pipeline

A flaw was found in FreeRDP, a free implementation of the Remote Desktop Protocol RDP. A malicious RDP server can exploit a heap buffer overflow vulnerability by sending a specially crafted graphics command to a FreeRDP client. This allows the server to write data outside of its intended memory...

CLSA-2026-1775059689 python: Fix of CVE-2025-15367

CVE-2025-15367: reject control characters in POP3 commands to prevent command injection via newlines...

CLSA-2026-1775058454 python: Fix of CVE-2025-15366

CVE-2025-15366: reject control characters in IMAP commands to prevent command injection...

CLSA-2026-1775058202 python: Fix of CVE-2025-15366

CVE-2025-15366: reject control characters in IMAP commands to prevent command injection...

freerdp: FreeRDP: Arbitrary code execution via heap buffer overflow in GDI surface pipeline

A flaw was found in FreeRDP, a free implementation of the Remote Desktop Protocol RDP. A malicious RDP server can exploit a heap buffer overflow vulnerability by sending a specially crafted graphics command to a FreeRDP client. This allows the server to write data outside of its intended memory...

CVE-2026-34430 ByteDance DeerFlow LocalSandboxProvider Host Bash Escape

ByteDance DeerFlow versions prior to commit 92c7a20 contain a sandbox escape vulnerability in bash tool handling that allows attackers to execute arbitrary commands on the host system by bypassing regex-based validation using shell features such as directory changes and relative paths. Attackers...

CVE-2026-34430

CVE-2026-34430 affects ByteDance DeerFlow versions before commit 92c7a20. A sandbox escape exists in the bash tool handling, allowing an attacker to bypass regex-based validation via shell features (e.g., directory changes, relative paths) and exploit incomplete shell semantics modeling to read/m...

Exploit for CVE-2025-48757

Vibe Coding Security Scanner A security audit skill for AI-as...

CVE-2026-4399

Prompt injection vulnerability in 1millionbot Millie chatbot that occurs when a user manages to evade chat restrictions using Boolean prompt injection techniques formulating a question in such a way that, upon receiving an affirmative response 'true', the model executes the injected instruction,...

CVE-2026-30310

In its design for automatic terminal command execution, Sixth offers two options: Execute safe commands and Execute all commands. The description for the former states that commands determined by the model to be safe will be automatically executed, whereas if the model judges a command to be...