Evidence · DiggingBeagle record
MCP Python SDK experimental task handlers allowed cross-client task access
Maintainer advisory for experimental task ownership. The remediation section distinguishes SDK-generated run_task IDs from explicit IDs, direct store writes and custom handlers; those qualifications are material to deployment review.
- Published
- Jun 5, 2026
- Accessed
- Oct 1, 2026
- Publisher
- Model Context Protocol Python SDK
- Source role
- primary disclosure
- Version
- 2026-06-05
- Rights
- Public web source; citation and short excerpt/paraphrase only.
Evidence record
Maintainer advisory for experimental task ownership. The remediation section distinguishes SDK-generated run_task IDs from explicit IDs, direct store writes and custom handlers; those qualifications are material to deployment review.
Claim-level citations (4)
- supportsMCP Python SDK tasks crossed client-session boundaries: The default experimental task handlers did not bind a task to the session that created it, so another connected client could observe, read results from or cancel that task.
Summary and details
- supportsMCP Python SDK tasks crossed client-session boundaries: The MCP Python SDK advisory lists version 1.27.2 as the patched version for CVE-2026-52870.
Patched versions
- supportsMCP Python SDK tasks crossed client-session boundaries: CVE-2026-52870 affected MCP Python SDK versions 1.23.0 through 1.27.1 when applications explicitly enabled the experimental tasks feature.
Affected versions and Am I affected?
- supportsMCP Python SDK tasks crossed client-session boundaries: The advisory says the patch binds SDK-generated run_task IDs to the creating session; explicit task IDs and direct TaskStore writes do not receive that automatic binding and remain reachable by ID.
Remediation: explicit task_id, direct TaskStore writes and custom handlers