
Audits Android TV and Google TV devices over ADB for BadBox, XCPE and Vo1d malware, with a VPN traffic sentinel, IoC blocklists and reversible remediation.
An autonomous, dual-layer forensic, malware removal, and reversible remediation suite (CLI + native Android TV app with local loopback VPN sentinel) engineered to audit, detect, verify, neutralize, and restore Android TV, Google TV, and generic streaming devices compromised by supply-chain trojans, botnets, and malware (BadBox, XCPE, Vo1d, Fuyao, Videostrong/SZBBOYS, and Residential Proxyware).
Language: English (Default) | Para a versão em Português (PT-PT), consulte README.pt.md | 📱 Tested Devices Registry | 🔬 Architecture Decisions (ADRs) | 📜 Third-Party Notices
Thousands of budget Android TV boxes and Smart TVs arrive pre-infected straight from the factory or become compromised via malicious streaming apps and trojanized firmware. The malware typically:
package_verifier_user_consent = -1./system or /vendor partition (surviving factory resets).This suite provides two complete audit & cleanup paths:
audit_security.py) featuring subnet discovery, token-saving pure JSON mode, mandatory backups, APK quarantining, and injection-safe parameterized rollback.Before running tools or dispatching an AI Agent, enable ADB Debugging on your TV so it can be audited over Wi-Fi:
[!WARNING] 🔒 Security Best Practice (Turn OFF Debugging when finished):
Leaving ADB / USB Debugging permanently enabled on your home Wi-Fi network exposes a management port (5555) that other devices or malicious apps on the local network could theoretically probe.
Always remember to turn OFF USB Debugging in Developer Options once the audit and remediation are complete!
This repository includes self-contained operational guidelines in AGENTS.md that modern AI coding agents automatically follow. You don't need to manually configure Python, install drivers, or write commands.
[!IMPORTANT] 🛡️ Safe Autonomous Workflow for AI Agents:
Always instruct the AI Agent to run--quick-scan(100% read-only diagnosis) first. Review the technical findings and the remediation plan before executing--remediate.
Simply open your AI Agent and send this single prompt:
Clone https://github.com/Mafarricos/badbox-tv-audit.git, read AGENTS.md, and audit my Android TV at IP <TV_IP>.
(Tip: If you do not know your TV's IP address, ask the agent: ...read AGENTS.md, discover my Android TV devices on subnet 192.168.1.0/24, and audit them.)
The agent will autonomously:
AGENTS.md for forensic instructions.community_audit_submission.json) ready to help update the infected devices database.app/)The application is localized automatically (Portuguese if device is set to Portuguese; English for all other languages) and is 100% D-Pad navigable:
/oem/, /vendor/, Play Protect integrity, trojan artifacts (Vo1d, XCPE), and inspects package signatures in seconds.[CRITICAL]): If the TV initiates connections to known C2 servers or malicious domains.[SUSPICIOUS]): If a connection is made to a known botnet/proxy port without matching a known C2 IP/domain.[ANOMALY]): If unusual high-port network traffic is detected while the TV is idle.Note on Network Connectivity during Monitoring: During the VPN Sentinel window, DNS queries (UDP port 53) are actively inspected and forwarded so malware domain lookups succeed, while direct TCP/UDP data packets are inspected and dropped in the loopback interface. Consequently, general internet browsing/streaming is temporarily paused while the monitor is active and immediately resumes once the VPN is stopped.