Skip to content
CursorGHSA-3v8f-48vw-3mjx

Cursor Desktop sandbox escape via symlink and failed path canonicalization

CriticalCVE-2026-50549 · Published Jun 5, 2026

### Summary Cursor runs agent terminal commands in a sandbox by default. Before a `Write`, the agent canonicalizes the target path to confirm it stays inside the workspace, but when canonicalization fails it falls back to the original path and writes without approval. A malicious agent can create an in-workspace symlink that points outside the workspace and force canonicalization to fail — either because the target does not exist or because read permission is removed from the path — so the agent writes through the symlink to an arbitrary location without approval. ### Impact A malicious agent could write arbitrary files outside the workspace under the user's privileges. This enables non-sandboxed Remote Code Execution — for example by overwriting the `cursorsandbox` helper so later commands run unsandboxed — with no user interaction beyond a benign prompt. ### Remediation Update Cursor to version 3.0. Canonicalization failures are now treated as untrusted: when a target path cannot be resolved — including when read permission is removed to defeat resolution — the write is blocked instead of falling back to the original in-workspace path. ### Credit Cato AI Labs Maor Dokhanian (m...

GitHub advisory

Affected versions

PackageAffectedFixed in
cursor
Product
< 3.03.0
Details and references

### Summary Cursor runs agent terminal commands in a sandbox by default. Before a `Write`, the agent canonicalizes the target path to confirm it stays inside the workspace, but when canonicalization fails it falls back to the original path and writes without approval. A malicious agent can create an in-workspace symlink that points outside the workspace and force canonicalization to fail — either because the target does not exist or because read permission is removed from the path — so the agent writes through the symlink to an arbitrary location without approval. ### Impact A malicious agent could write arbitrary files outside the workspace under the user's privileges. This enables non-sandboxed Remote Code Execution — for example by overwriting the `cursorsandbox` helper so later commands run unsandboxed — with no user interaction beyond a benign prompt. ### Remediation Update Cursor to version 3.0. Canonicalization failures are now treated as untrusted: when a target path cannot be resolved — including when read permission is removed to defeat resolution — the write is blocked instead of falling back to the original in-workspace path. ### Credit Cato AI Labs Maor Dokhanian (maordok)

Severity from
GitHub (reviewed advisory)

More Cursor advisories

All Cursor

Critical advisories by email

Wednesdays: the week’s critical and high advisories in the AI and data stack, with the fixed versions. Only in weeks that have some.

Double opt-in. Unsubscribe any time.