An Empirical Investigation of Backdoors and Undocumented Functionality through IoT Real-Device Verification
Authors/Creators
Description
Backdoors and undocumented functions embedded in IoT devices pose serious supply chain security threats, yet the actual prevalence of such functions in commercial products remains poorly understood. To investigate this, we conducted an empirical study on the firmware of 19 commercial IoT products from 11 vendors, spanning routers, access points, cameras, network switches, Wi-Fi range extenders, and UTM appliances. Combining LLM-assisted static analysis with dynamic verification on real devices, we surfaced 272 candidate functions, of which 184 (67.6%) were completely undocumented in product specifications. Dynamic verification onfirmed 14 candidates as functional, of which 9 fell within the scope of this study—4 enabling nauthorized access and 5 enabling information disclosure—including Telnet activation via magic packets, undocumented management interfaces, and plaintext disclosure of Wi-Fi credentials and device identifiers.We additionally validated the underlying analysis on the DarkWrt benchmark (100% precision/recall on source code, 98.3%/77.3% on compiled firmware). All findings will be reported through coordinated disclosure.
Files
eurosp26posters-final4.pdf
Additional details
Related works
- Is part of
- Poster: 10.5281/zenodo.21493141 (DOI)