Interfaz en español. Las fichas, los resúmenes y las fuentes conservan su idioma original.

Todos los avisosActividad de amenazas
Hablar con Oktsec
Cloud identities and exposed applications

ARTEX: exposed credentials turn into cloud access

Actualizado 14 nov 2026
Actuar ahoraInforme de investigación

Seguí las novedades de seguridad

Alertas prioritarias y resúmenes de Radar por correo.

Análisis de amenazas

Microsoft, Google and Mandiant document ARTEX use in intrusions involving exposed credentials, cloud permission checks and data access. Cisco Talos adds evidence of associated infrastructure and connected tools; these reports cover different actors and outcomes.

Quiénes deben revisar: Cloud identity, application security and incident response teams.

Acción recomendada

Revoke exposed credentials and tokens, check for added application credentials and roles, and investigate cloud activity and affected hosts.

¿Qué ocurrió y por qué es relevante?

Microsoft’s November 5 report describes public repository secret discovery followed by cloud access, including Azure Key Vault data theft. Its November 6 report separately identifies Storm-3358 activity against financial organizations, mainly in Southeast Asia. Shared use of ARTEX does not make every activity the same campaign or actor.

Google and Mandiant’s October 31 investigation adds cloud and endpoint evidence from incident response. In some cases, attackers supplied already compromised credentials to agents, which tested permissions and attempted privileged operations. One AWS sequence recorded 27 events in 23 seconds: that measures a burst of permission checks, not the duration of a complete intrusion. Failed actions and blocked movement also appear in the investigation.

Cisco Talos’s November 13 analysis explains that ARTEX coordinates a planner, workers and connected tools. Traditional command and control frameworks are not bundled natively; operators can add tools through MCP. Talos identified more than 2,500 IP addresses associated with ARTEX command and control infrastructure, with other tooling present on parts of that infrastructure. This is an infrastructure count, not a count of victims or proof that every listed tool participated in one intrusion.

For affected teams, the investigation must follow the identity beyond its first login: application credentials and roles, cloud operations, data access and endpoint activity. Rotating the original secret alone may leave another credential or persistence mechanism in place. The linked Research article develops these checks using the FBI, Google/Mandiant and Talos reports. These sources do not establish compromise of all users of Azure or AWS, or victims in Latin America and the Caribbean for the activity described here.

Sector afectado
Cloud identities and exposed applications
Divulgado
5 nov 2026
Última revisión de evidencia
14 nov 2026 UTC

¿Qué evidencia respalda este aviso?

Evidencia del incidente

Informe de investigación

Fundamento de la prioridad

Researchers document successful credential abuse and cloud intrusions. Act immediately on exposed credentials or evidence of unauthorized access; the reports do not establish compromise of every organization using these platforms.

¿Qué implica para los entornos de IA?

Agents can automate permission checks and tool use after access is obtained. Their reach depends on credentials, permissions and connected tools; findings do not establish unrestricted autonomy.

Fuentes de este aviso

Abrir evidencia de respaldo y registro técnico
Historial editorial2 entradas

Added vendor scope, corrections and response guidance.

Added Google/Mandiant and Cisco Talos findings, AWS investigation scope and checks for application persistence and connected tools.

Acciones de respuesta

Prepará tu respuesta

0 de 5 acciones registradas como completadas

Usá esta lista para evaluar si el aviso afecta a tus sistemas, registrar las acciones de tu equipo y compartir el resultado. Radar no escanea ni verifica tu entorno.

Los cambios se guardan automáticamente en este navegador. No se sincronizan con tu equipo ni se envían a Oktsec.

1. Evaluá si aplica a tu entorno

Leé el aviso y sus fuentes; después registrá tu decisión. Si tenés dudas, elegí «Requiere verificación».

Evaluación de aplicabilidadSin evaluar

2. Registrá la respuesta

Indicá el sistema y la persona o equipo responsable. Actualizá cada acción y agregá una referencia al ticket o notas sobre lo revisado.

  1. Revoke exposed credentials

    Revoke exposed secrets at their issuer and invalidate affected tokens or sessions. Removing a secret from a repository does not revoke it. Preserve sign-in and resource access records.

    Agregar evidencia o notasUsá solo referencias. No guardes credenciales ni datos confidenciales acá.
  2. Reduce what identities can access

    Review application permissions, service principal credentials, newly assigned roles and identities able to deploy workloads. Prefer managed identities and short lived credentials where supported.

    Agregar evidencia o notasUsá solo referencias. No guardes credenciales ni datos confidenciales acá.
  3. Review exposed applications

    The separate Storm-3358 report includes anonymous uploads, SSRF and remote access tools. Review relevant exposed services and investigate persistence.

    Agregar evidencia o notasUsá solo referencias. No guardes credenciales ni datos confidenciales acá.
  4. Trace cloud activity beyond the first credential

    Correlate authentication, cloud API and endpoint records. Investigate unexpected permission probes, storage writes or deletes, new compute instances and application credential changes.

    Agregar evidencia o notasUsá solo referencias. No guardes credenciales ni datos confidenciales acá.
  5. Contain the access that remains

    Oktsec recommends checking whether an intruder added another credential, role or remote access tool before closing the incident. Isolate affected hosts and disable compromised applications where appropriate.

    Agregar evidencia o notasUsá solo referencias. No guardes credenciales ni datos confidenciales acá.

3. Compartí el plan

Descargá un archivo Markdown con tu decisión, los estados de las acciones, las notas y las fuentes. Compartilo con tu equipo o adjuntalo a un ticket.

VERIFICÁ TU ENTORNO

Sistemas afectados y correcciones

Abrir plan de respuesta

Compará la versión en ejecución y la configuración con las indicaciones del proveedor.

ProductoAfectadoVersión corregida / acción

Confirmá con el proveedor las correcciones adaptadas a tu distribución. Coincidir en el producto no basta para establecer exposición.

CoberturaInforme de investigación
EVIDENCIA CON FUENTES

Evidencia y fuentes

Revisado 14 nov 2026 UTC

Por qué tiene esta prioridad

Researchers document successful credential abuse and cloud intrusions. Act immediately on exposed credentials or evidence of unauthorized access; the reports do not establish compromise of every organization using these platforms.

Criterios de prioridad

Fuentes

Los datos técnicos provienen de estas fuentes. Las prioridades y las acciones recomendadas son el análisis de Oktsec.

  1. Microsoft: exposed secrets and cloud accessMicrosoft · Aviso del proveedor · Revisado 14 nov 2026 UTC
  2. Microsoft: Storm-3358 financial sector activityMicrosoft · Informe de investigación · Revisado 14 nov 2026 UTC
  3. Google and Mandiant: ARTEX and other offensive frameworksGoogle Threat Intelligence Group / Mandiant · Informe de investigación · Revisado 14 nov 2026 UTC
  4. Cisco Talos: ARTEX infrastructure and connected toolsCisco Talos · Informe de investigación · Revisado 14 nov 2026 UTC
Registro técnicoProductos, identificadores y detalles de respaldo
Registro del incidente
Servicios atacadosMicrosoft Entra ID, Azure, AWS, GitHub repositoriesCloud identities and exposed applications
Evidencia de investigación
Informe de investigación
Alcance de la investigaciónCloud identity, application security and incident response teams.
Reportado porMicrosoft, Google Threat Intelligence Group / Mandiant, Cisco Talos
Divulgado
5 nov 2026
Registro revisado
14 nov 2026 UTC

Para entornos con agentes de IA

Agents can automate permission checks and tool use after access is obtained. Their reach depends on credentials, permissions and connected tools; findings do not establish unrestricted autonomy.