PraisonAI versions before 4.5.128 have an unauthenticated information disclosure vulnerability (CWE-200) in the AgentOS deployment platform. The GET /api/agents endpoint returns sensitive agent metadata including names, roles, and the first 100 characters of system instructions to any unauthenticated caller. The FastAPI application does not enforce authentication or API key validation and uses a permissive CORS policy (allow_origins=["*"]) with host set to "0.0.0.0", exposing the endpoint to any network origin by default. This vulnerability allows unauthorized actors to access sensitive configuration data. The issue is fixed in version 4.5.128.

An unauthenticated attacker can access partial sensitive information about deployed agents, including their names, roles, and partial system instructions. This exposure could aid attackers in understanding the system's configuration or behavior, potentially facilitating further attacks or reconnaissance. There is no indication of integrity or availability impact. The CVSS score is 5.3 (medium severity), reflecting the network accessibility and confidentiality impact without requiring privileges or user interaction.

Upgrade PraisonAI to version 4.5.128 or later, where this vulnerability is fixed. Prior to upgrading, restrict network access to the AgentOS deployment platform to trusted users only and implement external access controls if possible. Since no official remediation level or patch link is provided beyond the version fix, rely on the vendor's version update as the primary mitigation.