Empresa
¿Quiénes somos? Visión y Valores
Herramientas
Email Checker Vigía DNS SSL Checker Password Strength HTTP Headers
Alertas
Todas las alertas Vulnerabilidades Incidentes Solo críticas En CISA KEV
Editorial
Análisis técnico ¿Cuál es mi IP?
Blog
Blog 2MCI ISO 27001 Amenazas LATAM Recursos Gratuitos eBook Gratuito Newsletter Podcast / YouTube
Empresa
Servicios Contacto Suscribirse al Newsletter
🆕 Nuevo en 2MCI
✨ Crear cuenta gratis 🛠️ Ver herramientas sin registro
Ya tengo cuenta
🔒 Iniciar sesión
Equipo
🏠 Portal interno 2MCI
Crítico CVE-2026-55743 Multiple Vendors

[CVE-2026-55743] The shell tool command allowlist in the SecurityPolicy of OpenHuman desktop agent through 0.54.0 (de…

Vulnerabilidad · Publicado 17/06/2026 · Actualizado 01/08/2026

9.6
CVSS 3.x
04 Medio7 Alto9 Crítico10
Severidad: Crítico
Resumen ejecutivo

The shell tool command allowlist in the SecurityPolicy of OpenHuman desktop agent through 0.54.0 (default Supervised security policy) can be bypassed to execute arbitrary OS commands with the privileges of the desktop user. Two flaws in src/openhuman/security/policy.rs combine: (1) is_args_safe() blocks the find flags -exec and -ok but not the functionally identical -execdir and -okdir, which also…

Descripción técnica

Descripción técnica

The shell tool command allowlist in the SecurityPolicy of OpenHuman desktop agent through 0.54.0 (default Supervised security policy) can be bypassed to execute arbitrary OS commands with the privileges of the desktop user. Two flaws in src/openhuman/security/policy.rs combine: (1) is_args_safe() blocks the find flags -exec and -ok but not the functionally identical -execdir and -okdir, which also execute an arbitrary command for each matched file; and (2) skip_env_assignments() strips leading inline KEY=value environment-variable assignments before allowlist validation, so a command such as GIT_EXTERNAL_DIFF=<cmd> git diff is validated as the allowed git diff but, when executed via the shell, runs <cmd> through git's environment-driven hooks (for example GIT_EXTERNAL_DIFF or GIT_SSH_COMMAND). Because the sandbox is the primary trust boundary between untrusted LLM-processed content and the host operating system, an attacker can achieve remote code execution via indirect prompt injection: a malicious document, email, calendar event, or web page ingested by the agent instructs it to run a benign-looking allowlisted command, resulting in arbitrary command execution, data exfiltration, arbitrary file read/write, and lateral movement on the user's machine. The issue was fixed in commit 60050aa09a870f53ed7e4cd40ed41fd2860329e7 (first released in 0.54.22-staging; first stable release 0.56.0), which blocks -execdir/-okdir for find.

Puntuación CVSS

Score: 9.6/10 — Severidad: CRITICAL — Estado NIST: Deferred

CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H

Debilidades (CWE)

CWE-78, CWE-184

Fuente oficial

Publicado en NIST NVD.

¿Qué hacer?
  • Revisar el aviso oficial del fabricante y aplicar los parches o mitigaciones disponibles
  • Consultar las referencias en NIST NVD para actualizaciones de estado
  • Priorizar según el score CVSS y la exposición de los sistemas afectados
Esta alerta fue generada automáticamente a partir del NVD del NIST.
← Volver a todas las alertas