The vulnerability exists in the global policy read endpoint (`GET /api/latest/fleet/policies/{policy_id}`) of Fleet DM where authorization is performed against an empty policy struct with a nil TeamID, allowing any user with observer-level access on any team to bypass team isolation. The database fetch does not filter policies by team ownership when TeamID is nil, returning any policy by ID regardless of team. This allows unauthorized reading of policies belonging to other teams. Properly secured endpoints enforce team scope checks, but this endpoint lacks a post-fetch verification to confirm the policy is global (TeamID nil).

An authenticated user with observer-level access on a single team can read SQL queries, compliance pass/fail counts, software installer details, and script metadata from policies belonging to other teams. This exposure leaks sensitive security monitoring and compliance information across team boundaries, violating the intended team isolation model and potentially revealing internal infrastructure and security posture details.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. The recommended fix is to add a post-fetch check in the vulnerable endpoint to verify that the returned policy is actually a global policy (with nil TeamID), consistent with other endpoints that enforce team scope. Until a patch is available, restrict observer-level access carefully and monitor for unauthorized access attempts.