Vulnerabilidad · Publicado 15/09/2026
PraisonAI versiones 1.4.0 a 1.7.2 contienen una falla de control de acceso en createAgentLoop() que ejecuta herramientas antes de validar permisos. Un atacante puede forzar la ejecución de funciones no autorizadas que generen efectos secundarios irreversibles, comprometiendo sistemas de IA generativa en empresas mexicanas y latinoamericanas. El CVSS 8.8 refleja impacto alto en confidencialidad e integridad de datos.
PraisonAI is a multi-agent teams system. From 1.4.0 until 1.7.2, createAgentLoop() in src/praisonai-ts/src/ai/agent-loop.ts passes executable tools to generateText() before invoking the onToolCall approval callback. Because the wrapped AI SDK executes tool handlers during generation, a callback that returns false records tool_rejected only after the denied tool has already produced side effects and populated toolResults. Applications using onToolCall as a human or policy approval boundary can therefore execute rejected file, command, API, or data-modifying operations. This issue is fixed in version 1.7.2.
Score: 8.8/10 — Severidad: HIGH — Estado NIST: Received
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
CWE-693, CWE-862, CWE-863
Publicado en NIST NVD.