PraisonAI versions before 4.5.115 contain an information exposure vulnerability (CWE-200) due to lack of authentication on the A2U event stream server endpoints. The create_a2u_routes() function registers multiple endpoints that allow access to agent activity data without any authentication checks. This results in unauthorized disclosure of sensitive information. The CVSS 3.1 base score is 7.5, reflecting network attack vector, no privileges required, no user interaction, and high confidentiality impact. The issue is resolved in PraisonAI version 4.5.115.

Unauthorized actors can access sensitive agent activity information from the PraisonAI A2U event stream server without authentication. This leads to a high confidentiality impact, as sensitive data is exposed. There is no impact on integrity or availability reported. No known active exploitation has been documented.

Upgrade PraisonAI to version 4.5.115 or later, where this vulnerability is fixed by adding authentication checks to the affected endpoints. Since no official patch link or vendor advisory is provided, confirm the fix by verifying the version in use. Patch status is not explicitly confirmed beyond the stated fix version, so check vendor resources for current remediation guidance.