GHSA-9VHV-G5HF-M7M7 fast-agent: `fast-agent serve` binds HTTP to 0.0.0.0 with no authentication by default; with --shell this is unauthenticated remote command execution

Summary fast-agent serve exposes the agent over MCP. By default it uses HTTP transport bound to 0.0.0.0 all interfaces on port 8000 with no authentication auth=None and prints no warning. Any host that can reach the port can call the exposed agent tool. When the server is started with --shell whi...

CVE-2026-109440

LangAlpha through 2026.09.07 contains an indirect prompt injection vulnerability that allows remote attackers to execute shell commands or Python by planting instructions in web content retrieved by the agent. Because Bash and ExecuteCode lack an allow-list and human-in-the-loop approval, attacke...

PT-2026-104374

Name of the Vulnerable Software and Affected Versions AgentScope version 1.0.20 Description Code injection exists in the execute shell command function located in src/agentscope/tool/ coding/ shell.py. Depending on the exposed entry, an attacker can trigger the execution of attacker-controlled co...

Malicious code in integer-execute-shell-runtime-protected (npm)

The package integer-execute-shell-runtime-protected was found to contain malicious code. --- Credit: OpenSSF source...

Malicious code in bnpl-blocks-atom-bnpl-text (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 1f37374f9e046ed30b7430e942c4e4f23ddc6ef55e76a3a00b2f40506c0f2bb9 On require of bnpl-blocks-atom-bnpl-text, index.js loads polyfill.js, which selects an OS-specific asset and fetches a binary over HTTPS from...

Malicious code in claims-patch-url (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector c07292652ca72654d945a2833f8cf8915ea79f1ce665a0445938ff2605394c08 [email protected] presents an empty ClaimsPatchUrl shell class in index.js and unconditionally requires./compat.js on load. On require, compat....

flyto-core has Unauthenticated Command Execution via HTTP MCP `execute_module`

Unauthenticated Command Execution via HTTP MCP executemodule SummaryThe HTTP MCP endpoint POST /mcp in flyto-core accepts unauthenticated JSON-RPC tools/call requests and dispatches them to arbitrary registered modules, including sandbox.executeshell, which passes attacker-controlled input direct...

PYSEC-2026-2482 flyto-core has Unauthenticated Command Execution via HTTP MCP `execute_module`

Unauthenticated Command Execution via HTTP MCP executemodule Summary The HTTP MCP endpoint POST /mcp in flyto-core accepts unauthenticated JSON-RPC tools/call requests and dispatches them to arbitrary registered modules, including sandbox.executeshell, which passes attacker-controlled input...

AgentScope Vulnerable to Remote Code Injection

A vulnerability was determined in modelscope agentscope up to 1.0.18. Affected by this vulnerability is the function executepythoncode/executeshellcommand of the file src/AgentScope/tool/coding/python.py. This manipulation causes code injection. The attack is possible to be carried out remotely...

GHSA-H9F9-H6GM-WC85 flyto-core has Unauthenticated Command Execution via HTTP MCP `execute_module`

Unauthenticated Command Execution via HTTP MCP executemodule Summary The HTTP MCP endpoint POST /mcp in flyto-core accepts unauthenticated JSON-RPC tools/call requests and dispatches them to arbitrary registered modules, including sandbox.executeshell, which passes attacker-controlled input...

CVE-2026-34112 Guardian Language-System Unauthenticated OS Command Injection via id Parameter in speechmac.php

Guardian language-system passes the id GET parameter directly into a PHP exec call in speechmac.php line 18 without sanitization: exec"php jobs/speechaudiomac.php ".$loginsession." ".$GET'id'." ...". No authentication is required. An unauthenticated remote attacker can append shell...

PT-2026-53330

Summary agentic-flow versions = 2.0.13 MCP server tools interpolated attacker-influenceable tool parameters e.g. agent, task, name, language, agentdb arguments directly into shell command strings passed to execSync. A malicious value reaching any of the affected MCP tools could break out of the...

EUVD-2026-27009

Evolver is a GEP-powered self-evolving engine for AI agents. Prior to version 1.69.3, a command injection vulnerability in the extractLLM function allows attackers to execute arbitrary shell commands on the server. The function constructs a curl command using string concatenation and passes it to...

AgentScope Vulnerable to Remote Code Injection

A vulnerability was determined in modelscope agentscope up to 1.0.18. Affected by this vulnerability is the function executepythoncode/executeshellcommand of the file src/AgentScope/tool/coding/python.py. This manipulation causes code injection. The attack is possible to be carried out remotely...

GHSA-CR24-FV3H-8CJM AgentScope Vulnerable to Remote Code Injection

A vulnerability was determined in modelscope agentscope up to 1.0.18. Affected by this vulnerability is the function executepythoncode/executeshellcommand of the file src/AgentScope/tool/coding/python.py. This manipulation causes code injection. The attack is possible to be carried out remotely...

Arbitrary Code Injection

Overview agentscope is an AgentScope: A Flexible yet Robust Multi-Agent Platform. Affected versions of this package are vulnerable to Arbitrary Code Injection via the executeshellcommand function. An attacker can execute arbitrary code by supplying crafted input remotely. Remediation There is no...

CVE-2026-6603 modelscope agentscope _python.py execute_shell_command code injection

A vulnerability was determined in modelscope agentscope up to 1.0.18. Affected by this vulnerability is the function executepythoncode/executeshellcommand of the file src/AgentScope/tool/coding/python.py. This manipulation causes code injection. The attack is possible to be carried out remotely...

AgentScope security vulnerabilities

AgentScope is an open-source application developed by ModelScope. It simplifies the development of multi-agent applications based on LLMs. Versions of AgentScope prior to 1.0.18 contain security vulnerabilities, which stem from incorrect operations on the function...

CVE-2020-5722

The HTTP interface of the Grandstream UCM6200 series is vulnerable to an unauthenticated remote SQL injection via crafted HTTP request. An attacker can use this vulnerability to execute shell commands as root on versions before 1.0.19.20 or inject HTML in password recovery emails in versions befo...

OESA-2024-1159 aops-ceres security update

An agent which needs to be adopted in client, it managers some plugins, such as gala-gopherkpi collection, fluentdlog collection and so on. Security Fixes: In versions 1.3.0-1.4.1 of the ceres software package, the executeshellcommand function does not properly verify or filter the command or...