Loading story…
Security researchers have demonstrated new attack methods targeting AI-powered browsers and integrated development environments (IDEs) by exploiting their integration with external servers and AI sidebars. In one case, a proof-of-concept attack showed that a rogue Model Context Protocol (MCP) server could inject malicious JavaScript into Cursor’s built-in browser, allowing attackers to replace login pages, harvest credentials, and potentially compromise the victim’s workstation by leveraging the IDE’s privileges. The attack leverages the client-server architecture of MCP, which is increasingly used in AI agent workflows, and highlights the risks of using unvetted or custom MCP servers in developer environments.
Separately, researchers have revealed an "AI sidebar spoofing" technique that targets AI browsers such as Comet by Perplexity and Atlas by OpenAI. This attack exploits users’ trust in AI-generated instructions by manipulating the AI sidebar interface, potentially leading to credential theft or other malicious outcomes. Both attack vectors underscore the expanding attack surface introduced by AI integrations in browsers and development tools, and the need for heightened scrutiny of third-party server integrations and user interface trust boundaries in AI-powered applications.
Track how attackers are adapting to this technology.
2 events from the most recent confirmed update back to the earliest known activity.
Security researchers from Knostic.ai demonstrated a proof-of-concept attack showing that a malicious Model Context Protocol server can inject JavaScript into Cursor's built-in browser. The attack can replace login pages, steal credentials and cookies, and potentially lead to full workstation compromise because the MCP server inherits the IDE's privileges.
Researchers described a new attack technique called AI sidebar spoofing in which a malicious browser extension injects a fake AI assistant sidebar into AI-powered browsers such as Perplexity Comet and OpenAI Atlas. The method could be used to deliver phishing prompts, malicious links, or device-compromise instructions by abusing user trust in browser-integrated AI.
Vulnerabilities, threat actors, malware, products, organizations, breaches, and observables Mallory has linked to this story. Indicator values are masked here and available in full in the app.
Indicator values are masked on this page. See the values in Mallory Domains, IPs, hashes, and URLs are exportable to your SIEM.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
2 references tracked. Mallory keeps watching after this page renders.
Map indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Researchers and security firms disclosed multiple ways attackers can manipulate AI agents and coding assistants into exposing sensitive data or executing malicious actions. In one case, Invariant Labs showed GitHub integrations built on the **Model Context Protocol (MCP)** could be abused to access private repositories, while Adversa AI reported a **two-click** exploit in Cursor that can trick developers into installing a malicious MCP server through a deeplink argument-injection flaw and a misleading confirmation dialog. Because MCP servers run with the user’s privileges, a successful attack could lead to arbitrary command execution, theft of source code and secrets, and full compromise of a developer workstation. Separate academic research found that AI agents can also be subverted through poisoned but trusted-looking data rather than explicit prompt injection. A study covered by *Nature* showed agents from OpenAI, Anthropic, and Google often accepted doctored datasets and altered README files, producing manipulated conclusions, while another paper introduced **agent data injection (ADI)**, in which attackers tamper with fields such as sender names, button IDs, or tool history to trigger misclicks, unauthorized purchases, fraudulent merge approvals, or attacker-command execution. Researchers said current defenses remain weak because agents do not reliably distinguish trusted from untrusted data, although measures such as MCP allowlisting, randomized identifiers, and stronger provenance tracking can reduce risk.
Jul 22, 2026
A new class of vulnerabilities, termed "IDEsaster," has been discovered affecting a wide range of AI-powered Integrated Development Environments (IDEs) and coding assistants. Research revealed that over 30 security vulnerabilities, including 24 assigned CVEs, impact more than 10 leading products such as GitHub Copilot, Claude Code, and others, potentially exposing millions of users. The vulnerabilities stem from the integration of AI agents into IDEs, which were not originally designed with such capabilities in mind, leading to attack chains that can result in data exfiltration and remote code execution. Major vendors have issued advisories and updated documentation in response to these findings. Further research highlights the risks associated with the Model Context Protocol (MCP) sampling feature, commonly used in coding copilot applications. Without adequate safeguards, malicious MCP servers can exploit this feature to perform resource theft, hijack conversations, exfiltrate sensitive data, and covertly invoke tools. Proof-of-concept attacks demonstrate that the implicit trust model and lack of robust security controls in MCP can be leveraged for persistent and covert attacks, underscoring the urgent need for improved security measures in AI-driven development environments.
Jun 29, 2026
Researchers and vendors have disclosed serious security weaknesses in AI-enabled browsers, warning that agentic features can undermine core browser protections such as the same-origin policy, tab isolation, and safe handling of cross-origin content. A study cited by *Live Science* found inconsistent security behavior across seven AI browsers and concluded that more capable products often introduced greater risk, including exposure through prompt injection, multi-tab visibility, and unsafe memory handling. Brave separately detailed indirect prompt injection risks in its discussion of Perplexity Comet, highlighting how malicious web content can manipulate browser agents through embedded instructions. A separate vulnerability chain in OpenAI’s ChatGPT Atlas browser showed how these design risks can translate into account-compromise scenarios. According to Hacktron AI, exposed Chromium Mojo IPC interfaces were reachable from broadly allowlisted OpenAI web origins, and researchers chained a `postMessage`-based XSS on `forums.openai.com` with a login CSRF flaw to execute JavaScript on an authenticated OpenAI domain. From there, they were able to control Atlas browser functions, enumerate and manipulate tabs, and continuously read live tab URLs, enabling theft of OAuth authorization codes that could lead to takeover of accounts such as GitHub, Reddit, and Facebook. OpenAI said it fixed the issue in Atlas version `1.2025.288.15` and awarded a $5,000 bounty.
Jul 24, 2026