Loading vulnerability…
CVE-2025-54135 (CurXecute) is a remote code execution vulnerability in the Cursor AI code editor affecting versions prior to 1.3.9. According to the provided content, Cursor could write files within the workspace without user approval. While editing an existing dotfile required approval, creating a new dotfile did not. This allowed an attacker, when the target sensitive MCP configuration file did not already exist, to chain an indirect prompt injection issue to cause Cursor to create and write a sensitive configuration file such as .cursor/mcp.json without user approval. By manipulating MCP configuration in this way, the attacker could register a malicious MCP server or otherwise alter execution behavior so that code would run on the victim host. The issue was fixed in Cursor 1.3.9.
Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
If you can’t patch tonight, do this now.
Patch, then assume compromise.
No valid public exploits. Mallory filtered out 2 candidates as fakes, detection scripts, or README-only repos.
All candidate exploits were filtered out by Mallory's validation.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.
20 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
An earlier Cursor prompt-injection-related vulnerability that allowed command execution by rewriting Cursor MCP configuration even after the user rejected the edit.
A vulnerability/attack chain affecting AI coding assistant and MCP-integrated developer environments, where untrusted repository files such as hidden AI configuration files can trigger malicious command execution or credential exfiltration when a project is opened.
A vulnerability/attack in AI coding assistant and MCP-integrated developer environments where untrusted repository files, including hidden AI configuration files, can trigger malicious command execution or credential exfiltration when a project is opened.
A Cursor IDE remote code execution vulnerability caused by MCP configuration manipulation.
Query your assets running an affected version, and investigate the blast radius.
Every observed campaign linking this CVE to a named adversary.
Malware families riding this exploit, with evidence and IOCs.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Cross-references every affected SKU, including bundled OEM variants.
Community discussion across Reddit, Mastodon, and other social sources.