Researchers at Pillar Security demonstrated sandbox escapes across four widely used AI coding agents: Cursor, OpenAI Codex CLI, Google Gemini CLI, and Antigravity. In nearly every case the agent never broke the sandbox directly; it only had to write a file that a trusted component outside the sandbox would later run, load, or scan. Failure modes included hook abuse, editing a virtual environment interpreter the editor then ran itself, planting Git metadata outside a .git folder to fire execution through fsmonitor, and a command allowlist that trusted a tool by name while the real invocation was not read only. Prompt injection in workspace content was the trigger.
Security firm Mindgard disclosed an unpatched flaw in Cursor, the widely used AI code editor, that runs code automatically when a developer opens a malicious repository on Windows. When loading a project, Cursor looks for a Git binary in several places, including the workspace itself, so a file named git.exe planted in the repository root gets executed with no click, prompt, or warning, and it re-runs on a cadence while the project stays open. In a proof of concept, a renamed Windows Calculator launched on its own. Mindgard reported it in December and went public after seven months with no fix, so there is no patch and no CVE, only workarounds.
Researchers at Cato AI Labs detailed two flaws, dubbed DuneSlide, in the AI code editor Cursor that let a prompt-injection attack break out of the sandbox Cursor uses to contain the commands its agent runs. The attacker never types anything: they plant instructions in content the agent reads on the user's behalf, such as a connected MCP service or a web page. One flaw abuses a working-directory setting to get an attacker path added to the allowed-write list, letting injected commands overwrite the sandbox helper itself and then run with no sandbox. Both are rated 9.8 and are fixed in Cursor 3.0; every earlier version is affected, so users should update.
The second day of Pwn2Own Berlin 2026 added $385,750 across 15 unique zero-days, bringing the running total to $908,750 across 39 zero-days. The headline was Orange Tsai of DEVCORE chaining three bugs to gain SYSTEM-level remote code execution on Microsoft Exchange Server, taking the $200,000 top prize and pushing his event total past $375,000. Other day-two wins included a Windows 11 integer-overflow LPE, a Red Hat Enterprise Linux for Workstations root, a use-after-free in NVIDIA Container Toolkit, and AI-category exploits against LM Studio, Cursor, OpenAI Codex, and Anthropic Claude Desktop (the last as a collision with a previously known bug).
Google patched a critical flaw in Gemini CLI, the command-line tool developers use to interact with Gemini models from CI pipelines and dev workstations. CVSS 10.0. The bug let an attacker execute arbitrary code on the developer's machine by feeding crafted input to the CLI - specifically through the same pattern that compromised LiteLLM and several other AI tools recently. A separate but related set of flaws in Cursor, the AI-powered IDE, also enables code execution. The pattern across all these AI dev tools is the same: input validation gaps where attacker-controlled prompts or model output reach a shell or code execution path.
Socket reported 73 newly identified malicious extensions on Open VSX, the marketplace used by VS Code, Cursor, and Windsurf editors. The extensions impersonate popular developer tools - same name, same icon, but published by newly-created GitHub accounts with empty repositories. Instead of being malicious from day one, they sit harmlessly for weeks gathering downloads and trust, then push a 'normal' update that silently installs malware. Six of the 73 extensions have already activated; the rest are still in the sleeper phase. The campaign is part of GlassWorm, an ongoing supply-chain attack family that has been working its way through npm, GitHub, and editor extension marketplaces since 2025.
Backfill from April 15: OX Security disclosed an architectural flaw in the official Model Context Protocol SDKs (Python, TypeScript, Java, Rust) that lets attacker-controlled JSON config trigger arbitrary OS commands via the STDIO transport. Roughly 200,000 publicly reachable MCP servers and 150 million SDK downloads inherit the issue. OX has tied 14 CVEs to the same root cause across LiteLLM (patched), Bisheng (patched), Windsurf (zero-click RCE in Cursor-style IDEs, still reported), Flowise, LangFlow, GPT Researcher, Agent Zero, and DocsGPT. Anthropic declined to patch the protocol, calling the behavior 'expected.'
TrustStrike Labs