CVE-2025-46070 & CVE-2025-46066: Critical Automai Vulnerabilities Enable Remote Code Execution and Full Administrative Compromise

Vendor: Automai
Affected Products: Automai BotManager, Automai Director
Affected Version: 25.2.0
Risk Level: Critical
Attack Surface: Network / Application / Privilege Management

Executive Summary

Two critical security vulnerabilities have been identified in Automai automation products. When combined or individually exploited, these flaws can result in full system compromise, including remote code execution and administrator-level privilege escalation.

These vulnerabilities are particularly dangerous because Automai components often run with elevated permissions and are embedded deep within enterprise automation environments.

Immediate defensive action and monitoring are strongly advised until an official vendor patch is applied.

Affected Products and Components

Product	Component	Description
Automai BotManager	BotManager.exe	Core automation execution service
Automai Director	Web / Orchestration Engine	Central management and workflow control

CVE-2025-46070

Automai BotManager — Remote Code Execution

Vulnerability Type

Remote Code Execution (RCE)
Command Injection

Severity

Critical (CVSS ~9.8–10.0)

Attack Prerequisites

Network access to the BotManager service
No authentication required in common deployment scenarios

Technical Description

Automai BotManager improperly validates incoming data used during task execution or update-related communication. User-supplied input is passed directly into system-level execution functions without sufficient sanitization.

As a result, attackers can inject operating system commands into otherwise legitimate requests. The service executes these commands with the privileges of the BotManager service account, which is frequently SYSTEM or Administrator.

Exploitation Flow

Attacker identifies a host running Automai BotManager.
A crafted network request is sent containing embedded shell commands.
BotManager parses the request without validation.
Injected commands are executed directly by the operating system.
Attacker gains persistent access, installs backdoors, or pivots laterally.

Impact

Full takeover of the automation host
Arbitrary command execution
Credential theft
Persistence via services or scheduled tasks
Lateral movement across the environment

Proof-of-Concept Status

Public proof-of-concept material exists.
Exploitation techniques are publicly documented.
PoC usage should be restricted strictly to educational and defensive research purposes only.

Detection and Monitoring

Primary Indicators of Compromise

BotManager spawning unexpected child processes
Execution of cmd.exe, powershell.exe, wmic.exe, or scripting engines
Creation of scheduled tasks or new services
Unexpected outbound network connections
Suspicious command-line arguments containing separators or encoded payloads

Splunk Detection Logic

Process Execution Detection

index=windows EventCode=4688
Parent_Process_Name="*BotManager.exe"
(New_Process_Name="*cmd.exe" OR New_Process_Name="*powershell.exe" OR New_Process_Name="*wmic.exe")
| stats count by host, Parent_Process_Name, New_Process_Name, Command_Line

Suspicious Command Line Injection

index=windows EventCode=4688
Parent_Process_Name="*BotManager.exe"
Command_Line="*;*" OR Command_Line="*&*" OR Command_Line="*|*" OR Command_Line="*-EncodedCommand*"

Log Source Requirements

Log Source	Purpose
Windows Security Logs (4688)	Process creation visibility
Sysmon (Event ID 1, 3)	Child processes and network connections
Automai Application Logs	Error messages and execution traces
Firewall / Proxy Logs	Suspicious inbound and outbound traffic
EDR Telemetry	Behavioral detection and alerting

Interim Mitigation (Until Patch)

Restrict network access to BotManager
Enforce outbound allow-listing
Run service under least-privileged account
Enable aggressive process execution monitoring
Isolate affected systems if suspicious activity is detected

Official Patch / Upgrade

Apply only the official vendor update when released:
https://www.automai.com/

CVE-2025-46066

Automai Director — Privilege Escalation

Vulnerability Type

Privilege Escalation
Improper Access Control

Severity

Critical (CVSS ~9.9)

Attack Prerequisites

Low-privileged authenticated access or internal foothold

Technical Description

Automai Director does not consistently enforce authorization checks for sensitive administrative actions. Certain crafted requests or workflow interactions allow users with limited privileges to execute actions reserved for administrators.

Once elevated, attackers can fully control automation workflows, modify jobs, execute commands indirectly, and access sensitive orchestration data.

Exploitation Flow

Attacker gains a basic Director user account.
Crafted requests bypass role validation.
User privileges are escalated silently.
Administrative capabilities are abused for persistence or execution.

Impact

Unauthorized administrative access
Workflow and job manipulation
Indirect command execution
Data exposure and system compromise

Proof-of-Concept Status

Publicly documented exploitation techniques exist
Use limited to defensive testing and research only

Detection and Monitoring

Indicators of Exploitation

Non-admin users performing admin-only actions
Sudden role changes without approval
Unapproved automation job modifications
Unexpected script uploads or workflow changes

Splunk Detection Logic

Privilege Escalation Detection

index=automai
event_type="role_change"
new_role="admin"
previous_role!="admin"
| stats count by user, host, _time

Unauthorized Administrative Action

index=automai
user_role!="admin"
(action="create_job" OR action="modify_workflow" OR action="upload_script")

Log Source Requirements

Log Source	Purpose
Automai Director Audit Logs	Role and permission changes
Web Server Logs	Unauthorized requests
Authentication Logs	User session tracking
SIEM Correlation	Privilege abuse detection
EDR	Script and execution monitoring

Interim Mitigation (Until Patch)

Restrict Director access to trusted users
Enforce MFA for all users
Minimize number of admin accounts
Monitor role changes in real time
Disable unnecessary upload features if possible

Official Patch / Upgrade

Apply only the official vendor update when released:
https://www.automai.com/

Overall Risk Assessment

These vulnerabilities represent a severe enterprise risk, especially in environments where Automai orchestrates critical workflows. A successful exploit can rapidly escalate from a single host to full automation platform compromise.

Final Takeaway

Treat exposed Automai systems as high-risk
Monitor process execution and privilege changes
Apply vendor patches immediately upon release
Perform threat hunting if Automai 25.2.0 is present