CVE-2026-2256-PoC

CVE-2026–2256 PoC Executive Summary A critical command injection vulnerability CVE-2026–2256, CVSS X.X was identified in the Shell tool component of the MS-Agent framework version 1.5.2. The vulnerability arises from improper sanitization of user-influenced input that is passed directly to a shel...

CVE-2026-75128

Langchain-Chatchat through 0.3.1.3 exposes a POST /tools/call API endpoint that invokes any registered tool by name without authentication, leaving the always-registered shell tool callable. Remote attackers can send a single request naming the shell tool to execute arbitrary OS commands as the...

EUVD-2026-104685

Langchain-Chatchat through 0.3.1.3 exposes a POST /tools/call API endpoint that invokes any registered tool by name without authentication, leaving the always-registered shell tool callable. Remote attackers can send a single request naming the shell tool to execute arbitrary OS commands as the...

CVE-2026-75128

Langchain-Chatchat through 0.3.1.3 exposes a POST /tools/call API endpoint that invokes any registered tool by name without authentication, leaving the always-registered shell tool callable. Remote attackers can send a single request naming the shell tool to execute arbitrary OS commands as the...

CVE-2026-75128 Langchain-Chatchat through 0.3.1.3 Unauthenticated RCE via /tools/call Endpoint

Langchain-Chatchat through 0.3.1.3 exposes a POST /tools/call API endpoint that invokes any registered tool by name without authentication, leaving the always-registered shell tool callable. Remote attackers can send a single request naming the shell tool to execute arbitrary OS commands as the...

CVE-2026-75128

Langchain-Chatchat versions through 0.3.1.3 are vulnerable to unauthenticated Remote Code Execution (RCE) due to a missing authentication check on the POST /tools/call API endpoint. This endpoint allows the invocation of any registered tool by name; specifically, the shell tool is always register...

CVE-2026-75128 Langchain-Chatchat through 0.3.1.3 Unauthenticated RCE via /tools/call Endpoint

Langchain-Chatchat through 0.3.1.3 exposes a POST /tools/call API endpoint that invokes any registered tool by name without authentication, leaving the always-registered shell tool callable. Remote attackers can send a single request naming the shell tool to execute arbitrary OS commands as the...

PT-2026-117937

Name of the Vulnerable Software and Affected Versions piflow versions 2.0.0 through 2.0.1 Description The DeepAgents chat agent contains a flaw allowing remote code execution. The 'POST /chat' and 'POST /chat/stream' endpoints lack authentication, enabling attackers to instruct the agent to call...

PT-2026-117947

Name of the Vulnerable Software and Affected Versions Langchain-Chatchat versions prior to 0.3.1.4 Description The software exposes a POST '/tools/call' API endpoint that allows the invocation of any registered tool by name without authentication. Because the shell tool is registered by default,...

Chatchat-Space Langchain-Chatchat Authorization Issue Vulnerability

Chatchat-Space Langchain-Chatchat 0.3.1.3 and earlier versions have an authorization issue vulnerability. This vulnerability arises from the POST /tools/call API endpoint, which allows arbitrary registered tools to be called without authentication. This enables remote attackers to execute arbitra...

Model-As-A-Trojan: Self-Activating Covert DNS Exfiltration Baked into the Weights of an LLM Agent

Open-weight models are downloaded from public hubs and run inside agent harnesses that grant them tools, most powerfully, a shell, under the assumption that a sandbox contains whatever the model might do. This paper demonstrates that the assumption has a blind spot: the weights themselves can be...

ModelScope MS-Agent 1.5.1 < 1.6.0 Command Injection (CVE-2026-2256)

The version of ModelScope MS-Agent installed on the remote host is 1.5.1 or later but prior to 1.6.0. It is, therefore, affected by a command injection vulnerability. The Shell tool runs commands on the host through a shell, guarded only by a regex denylist that can be bypassed, allowing an...

Malicious code in 9remote (npm)

Source: amazon-inspector 4160b2e2774d56c9e7373db28a4073d456248d0fe9b55979fa507f0910b65ce0 9remote is a remote-shell tool: its CLI launches an HTTP/WebSocket server and provisions a Cloudflare Quick Tunnel that exposes the host to the public internet at a 9remote.cc-fronted URL. Incoming WebSocket...

Improper Neutralization of Input Used for LLM Prompting

Overview strands-agents-tools is an A collection of specialized tools for Strands Agents Affected versions of this package are vulnerable to Improper Neutralization of Input Used for LLM Prompting via the shell tool. An attacker can execute arbitrary operating system commands by crafting a prompt...

CVE-2026-18733

A prompt injection vulnerability in the shell tool in Amazon Strands Agents Tools before 0.8.0 might allow remote actors to execute arbitrary operating system commands on the agent's host via a crafted prompt that sets the noninteractive parameter to true, bypassing the human consent gate. To...

CVE-2026-18733

CVE-2026-18733 affects the Shell tool in Amazon Strands Agents Tools prior to v0.8.0. A crafted prompt can set non_interactive to true, bypassing the consent gate and potentially allowing remote execution of arbitrary OS commands on the agent host. Impact appears to involve remote command executi...

CVE-2026-18733 Prompt injection bypasses shell tool consent gate in Strands Agents Tools

A prompt injection vulnerability in the shell tool in Amazon Strands Agents Tools before 0.8.0 might allow remote actors to execute arbitrary operating system commands on the agent's host via a crafted prompt that sets the noninteractive parameter to true, bypassing the human consent gate. To...

CVE-2026-18733

A prompt injection vulnerability in the shell tool in Amazon Strands Agents Tools before 0.8.0 might allow remote actors to execute arbitrary operating system commands on the agent's host via a crafted prompt that sets the noninteractive parameter to true, bypassing the human consent gate. To...

Malicious code in 0x2ai-multi-mq (npm)

Source: amazon-inspector 7d056f067b0af2084bd7777fcdb2ae6e2c06bb67f40929ba9900b5aa9cb83649 When the documented invocation npx 0x2ai-multi-mq is run, bin/start.cjs copies chatroom-mcp-lite-patched.cjs and chatroom-monitor.cjs into the user's current working directory, writes a .mcp.json containing ...

CVE-2026-57142

PraisonAI recipe execution has a dangerous-tool policy that is supposed to block default-denied tools unless the caller explicitly passes "allowdangeroustools=True". That policy only checks tools declared in "TEMPLATE.yaml" "requires.tools". For steps-based recipes, the actual execution path load...