CVE-2026-47391
Description détaillée
PraisonAI is a multi-agent teams system. Prior to version 4.6.40, PraisonAI's first-party A2A server example exposes an unauthenticated A2A JSON-RPC endpoint and registers a `calculate(expression)` tool implemented with Python `eval()`. The example also binds to `0.0.0.0`. A remote unauthenticated attacker can send `message/send` to `/a2a`; the request reaches `agent.chat()`, and a real LLM can invoke the registered `calculate` tool. In testing with `gemini/gemini-2.5-flash-lite`, this resulted in arbitrary Python execution in the server process, confirmed by creation of a marker file from an unauthenticated HTTP request. The issue affects deployments following the official A2A example or similar unauthenticated public A2A deployments with unsafe tools. The default unauthenticated A2A surface also exposes task history and task cancellation APIs, increasing confidentiality and integrity impact. Version 4.6.40 patches the issue.
Vecteur d'attaque (CVSS)
Dernières Vulnérabilités
CVE-2026-64877
An authenticated non-admin user can exploit a SQL injection flaw in the ticketing REST API to access sensitive data stored in the appliance database.
CVE-2026-63454
An authenticated path traversal vulnerability exists in AOS-CX. Successful exploitation of this vulnerability allows an attacker to copy arbitrary files to a user readable location from the command line interface of the underlying operating system, which could lead to remote code execution.
CVE-2026-63453
Buffer overflow vulnerabilities exist in the command line interface of AOS-CX. Successful exploitation of these vulnerabilities could allow a remote high-privileged user to execute arbitrary code as a privileged user on the underlying operating system.