CVE-2019-1010200

9.8 CRITICAL

📋 TL;DR

This CVE describes a critical OS command injection vulnerability in Google Voice Builder that allows remote attackers to execute arbitrary commands on affected servers. The vulnerability exists in three endpoints across two web servers and can be exploited via specially crafted GET requests. Any organization running vulnerable versions of Voice Builder is affected.

💻 Affected Systems

Products:
  • Google Voice Builder
Versions: All versions prior to commit f6660e6d8f0d1d931359d591dbdec580fef36d36
Operating Systems: Any OS running Voice Builder
Default Config Vulnerable: ⚠️ Yes
Notes: The vulnerability affects two web servers within the project: merlin_model_server and festival_model_server.

📦 What is this software?

⚠️ Risk & Real-World Impact

🔴

Worst Case

Full server compromise leading to data exfiltration, lateral movement within the network, and persistent backdoor installation.

🟠

Likely Case

Remote code execution allowing attackers to steal sensitive data, disrupt services, or use the server for further attacks.

🟢

If Mitigated

Attack blocked at network perimeter or detected before successful exploitation.

🌐 Internet-Facing: HIGH - The vulnerable endpoints are exposed via web servers and can be exploited remotely without authentication.
🏢 Internal Only: HIGH - Even internally accessible servers can be exploited by attackers who gain internal network access.

🎯 Exploit Status

Public PoC: ✅ No
Weaponized: LIKELY
Unauthenticated Exploit: ⚠️ Yes
Complexity: LOW

The vulnerability is straightforward to exploit via HTTP GET requests with malicious query parameters.

🛠️ Fix & Mitigation

✅ Official Fix

Patch Version: After commit f6660e6d8f0d1d931359d591dbdec580fef36d36

Vendor Advisory: https://github.com/google/voice-builder/commit/f6660e6d8f0d1d931359d591dbdec580fef36d36

Restart Required: Yes

Instructions:

1. Update to the latest Voice Builder version after commit f6660e6d8f0d1d931359d591dbdec580fef36d36. 2. Restart all Voice Builder services. 3. Verify the fix by checking that the vulnerable endpoints no longer accept malicious input.

🔧 Temporary Workarounds

Network Access Restriction

linux

Restrict access to the vulnerable endpoints using firewall rules or network segmentation.

iptables -A INPUT -p tcp --dport [VOICE_BUILDER_PORT] -s [TRUSTED_IPS] -j ACCEPT
iptables -A INPUT -p tcp --dport [VOICE_BUILDER_PORT] -j DROP

Web Application Firewall

all

Deploy a WAF with command injection detection rules to block exploitation attempts.

🧯 If You Can't Patch

  • Implement strict input validation and sanitization for all user inputs in the affected endpoints.
  • Run Voice Builder services with minimal privileges and in isolated containers or virtual machines.

🔍 How to Verify

Check if Vulnerable:

Check if your Voice Builder version is prior to commit f6660e6d8f0d1d931359d591dbdec580fef36d36 by examining the git commit history or version metadata.

Check Version:

git log --oneline -1

Verify Fix Applied:

Test the /tts and /alignment endpoints with command injection payloads (e.g., '; ls' or '| cat /etc/passwd') to ensure they are properly sanitized.

📡 Detection & Monitoring

Log Indicators:

  • Unusual command execution patterns in system logs
  • HTTP requests to /tts or /alignment endpoints with suspicious query parameters containing shell metacharacters

Network Indicators:

  • HTTP GET requests to vulnerable endpoints with command injection payloads in query strings

SIEM Query:

source="web_server_logs" AND (uri_path="/tts" OR uri_path="/alignment") AND query_string MATCHES "[;|&`$()]"

🔗 References

📤 Share & Export