CVE-2026-48124
Description
A vulnerability in Cursor Desktop before 3.0.0 allows workspace-defined Claude hook commands to execute without user approval, enabling sandbox escape and local compromise.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
A vulnerability in Cursor Desktop before 3.0.0 allows workspace-defined Claude hook commands to execute without user approval, enabling sandbox escape and local compromise.
Vulnerability
Cursor Desktop versions prior to 3.0.0 could execute workspace-defined Claude hook commands from .claude/settings.local.json without dedicated user approval [1]. A malicious workspace or an agent-created file could configure hooks that run local commands in the user's context when an agent turn ends [1].
Exploitation
An attacker must convince the user to open a malicious workspace or an agent-created file that contains a crafted .claude/settings.local.json. The workspace can define hook commands that execute automatically when an agent turn ends, without any additional user interaction beyond opening the workspace [1]. This allows the attacker to execute arbitrary commands in the user's local context.
Impact
Successful exploitation enables sandbox escape, persistence across turns, local data access, or follow-on compromise [1]. The attacker achieves code execution in the user's context, potentially gaining access to files, credentials, and the ability to install malware or pivot to other systems.
Mitigation
The issue has been fixed in Cursor Desktop version 3.0.0 [1]. Users should update to the patched version immediately. Workspace-sourced hook commands now require appropriate approval and are subject to the same execution policy controls as other agent shell commands [1]. If upgrading is not possible, users should avoid opening workspaces or agent-created files from untrusted sources.
AI Insight generated on Jun 15, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.
Affected products
2Patches
0No patches discovered yet.
Vulnerability mechanics
AI mechanics synthesis has not run for this CVE yet.
References
1News mentions
0No linked articles in our index yet.