CVE-2026-30615

A prompt injection vulnerability in Windsurf version 1.9544.26 allows remote attackers to execute arbitrary commands on a victim system. The flaw is triggered when the application processes attacker-controlled HTML content , which can lead to unauthorized modification of the local MCP configurati...

Windsurf 安全漏洞

Windsurf is an AI programming software developed by the Windsurf company. Version 1.9544.26 of Windsurf contains a security vulnerability. This vulnerability stems from prompt injection, and it could allow remote attackers to execute arbitrary commands on the victim’s system...

CVE-2026-39884

The CVE-2026-39884 entry concerns mcp-server-kubernetes (Model Context Protocol server for Kubernetes) with a vulnerability in the port_forward tool (src/tools/port_forward.ts). The code builds a kubectl command by string concatenation using user-controlled input and naively splits on spaces befo...

CVE-2026-40112

PraisonAI is a multi-agent teams system. Prior to 4.5.128, the Flask API endpoint in src/praisonai/api.py renders agent output as HTML without effective sanitization. The sanitizehtml function relies on the nh3 library, which is not listed as a required or optional dependency in pyproject.toml...

CVE-2026-4369

A maliciously crafted HTML payload in an assembly variant name, when displayed during the delete confirmation dialog and clicked by a user, can trigger a Stored Cross-site Scripting XSS vulnerability in the Autodesk Fusion desktop application. A malicious actor may leverage this vulnerability to...

JLSEC-2026-102 Interactive `run` permission prompt spoofing via improper ANSI neutralization

Summary Arbitrary program names without any ANSI filtering allows any malicious program to clear the first 2 lines of a opspawnchild or opkill prompt and replace it with any desired text. Details The main entry point comes down to the ability to override what the API control says 40process.js...

JLSEC-2026-100 Deno is vulnerable to race condition via interactive permission prompt spoofing

Impact Multi-threaded programs were able to spoof interactive permission prompt by rewriting the prompt to suggest that program is waiting on user confirmation to unrelated action. A malicious program could clear the terminal screen after permission prompt was shown and write a generic message li...

JLSEC-2026-105 Deno's `deno_runtime` vulnerable to interactive permission prompt spoofing via improper ANSI stripping

Summary A maliciously crafted permission request can show the spoofed permission prompt by inserting a broken ANSI escape sequence into the request contents. Details In the patch for CVE-2023-28446, Deno is stripping any ANSI escape sequences from the permission prompt, but permissions given to t...

CVE-2026-34225

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Versions 0.7.2 and below contain a Blind Server Side Request Forgery in the functionality that allows editing an image via a prompt. The affected function performs a GET request to a user-provided U...

GHSA-P4H8-56QP-HPGV SSH/SCP option injection allowing local RCE in @aiondadotcom/mcp-ssh

Impact A crafted hostAlias argument such as -oProxyCommand=... was passed to ssh/scp without an argument terminator. SSH interprets arguments starting with - as options regardless of position, so the option-injection caused SSH to execute the attacker-supplied ProxyCommand locally on the machine...

LogicEval: A Systematic Framework for Evaluating Automated Repair Techniques for Logical Vulnerabilities in Real-World Software

Logical vulnerabilities in software stem from flaws in program logic rather than memory safety, which can lead to critical security failures. Although existing automated program repair techniques primarily focus on repairing memory corruption vulnerabilities, they struggle with logical...

LLM-Guided Prompt Evolution for Password Guessing

Passwords still remain a dominant authentication method, yet their security is routinely subverted by predictable user choices and large-scale credential leaks. Automated password guessing is a key tool for stress-testing password policies and modeling attacker behavior. This paper applies...

PT-2026-34828

Impact A crafted hostAlias argument such as -oProxyCommand=... was passed to ssh/scp without an argument terminator. SSH interprets arguments starting with - as options regardless of position, so the option-injection caused SSH to execute the attacker-supplied ProxyCommand locally on the machine...

CVE-2026-40087

A flaw was found in LangChain. A missing validation of f-string prompt templates in some classes, specifically in DictPromptTemplate and ImagePromptTemplate, can cause the evaluation of attribute access or indexing expressions during template formatting. Also, f-string validation based on parsed...

PraisonAIAgents: Arbitrary File Read via read_skill_file Missing Workspace Boundary and Approval Gate

Summary readskillfile in skilltools.py allows reading arbitrary files from the filesystem by accepting an unrestricted skillpath parameter. Unlike filetools.readfile which enforces workspace boundary confinement, and unlike runskillscript which requires critical-level approval, readskillfile has...

EUVD-2026-21152

PraisonAIAgents has an OS Command Injection via shell=True in Memory Hooks Executor memory/hooks.py...

PT-2026-33895

Summary The approval system in PraisonAI Agents caches tool approval decisions by tool name only, not by invocation arguments. Once a user approves execute command for any command e.g., ls -la, all subsequent execute command calls in that execution context bypass the approval prompt entirely...

CLIP-Inspector: Model-Level Backdoor Detection for Prompt-Tuned CLIP Via OOD Trigger Inversion

Organisations with limited data and computational resources increasingly outsource model training to Machine Learning as a Service MLaaS providers, who adapt vision-language models VLMs such as CLIP to downstream tasks via prompt tuning rather than training from scratch. This semi-honest setting...

Anthropic Claude Code < 2.0.65 API Key Leak via Project Settings (CVE-2026-21852)

The version of Anthropic Claude Code installed on the remote host is prior to 2.0.65. It is, therefore, affected by an information disclosure vulnerability. A vulnerability in the project-load flow allowed malicious repositories to exfiltrate data including Anthropic API keys before users confirm...

CVE-2026-40111

PraisonAIAgents is a multi-agent teams system. Prior to 1.5.128, he memory hooks executor in praisonaiagents passes a user-controlled command string directly to subprocess.run with shell=True at src/praisonai-agents/praisonaiagents/memory/hooks.py. No sanitization is performed and shell...