Executive brief
Dask Distributed, a tool for managing parallel computing tasks, contains a security flaw when used alongside Jupyter Lab. An attacker can trick a user into clicking a malicious link that executes unauthorized code within the user's Jupyter environment. This could allow an attacker to run arbitrary commands or access sensitive data on the user's local machine or server.
Technical details
A cross-site scripting (XSS) vulnerability exists in the Dask dashboard component of Dask Distributed. When deployed in an environment using Jupyter Lab and jupyter-server-proxy, an attacker can craft a malicious URL that leverages the proxy to bypass security boundaries. If a victim clicks this link, the XSS payload triggers an error page in the Dask dashboard that executes code within the default Jupyter Python kernel. The attack assumes the target is running these services on default ports (localhost). The vulnerability is rooted in improper neutralization of input (CWE-79/CWE-80) and is mitigated in version 2026.1.0.
Affected products
- Dask distributed < 2026.1.0
Timeline
- 2026-01-16: disclosed
- 2026-01-16: advisory
- 2026-01-16: patched: Fixed in version 2026.1.0
References
- https://github.com/dask/distributed/security/advisories/GHSA-c336-7962-wfj2
- https://github.com/dask/distributed/commit/ab72092a8a938923c2bb51a2cd14ca26614827fa
- https://github.com/pypa/advisory-database/tree/main/vulns/distributed/PYSEC-2026-169.yaml
- https://api.github.com/repos/dask/distributed/security-advisories/GHSA-c336-7962-wfj2