PraisonAI's WSGI-based recipe registry server (server.py) prior to version 4.5.128 does not impose any limit on the size of HTTP request bodies it reads into memory, relying solely on the client-supplied Content-Length header. With authentication disabled by default, this allows any local process to send arbitrarily large POST requests, leading to memory exhaustion and denial of service. The Starlette-based server (serve.py) mitigates this with a 10MB request size limit via RequestSizeLimitMiddleware, but the WSGI server lacks equivalent throttling. The vulnerability is addressed in version 4.5.128.

An attacker with local access can cause a denial of service by exhausting server memory through sending large POST requests to the vulnerable WSGI-based server. There is no impact on confidentiality or integrity. The vulnerability does not require user interaction and has low attack complexity but requires local access.

Upgrade PraisonAI to version 4.5.128 or later, where this vulnerability is fixed by adding request size limits to the WSGI-based server. Until upgraded, restrict local access to the server to trusted users only to prevent exploitation. Patch status is not explicitly stated but the vendor fixed the issue in version 4.5.128.