Executive brief
NanoClaw, an agent management and messaging platform, contains a security flaw in how it handles administrative approvals for new communication channels. A user with limited administrative rights over one group can bypass restrictions to link messaging channels to other groups they are not authorized to manage. This could allow an unauthorized person to monitor or influence activities and communications within restricted parts of the organization.
Technical details
A privilege escalation vulnerability exists in NanoClaw's channel-registration approval flow due to improper authorization checks in the 'handleChannelApprovalResponse' function. The root cause is a failure to validate that a scoped administrator has administrative privileges over a target agent group when approving a channel connection. An attacker with 'scoped admin' privileges can submit forged or stale callback values to wire messaging channels into agent groups outside their authorized scope. This allows the attacker to observe or control activity in restricted agent groups. The issue is fixed in version 2.1.0 by enforcing target-group authorization at the time the approval is applied and filtering selectable options based on the approver's identity.
Affected products
- nanocoai NanoClaw < 2.1.0
Timeline
- 2026-05-20: other: Pull request submitted
- 2026-05-22: patched: Fix committed to main branch
- 2026-06-23: disclosed: CVE published