An AI agent has rapidly uncovered serious cryptographic weaknesses in several commercial USB encryption and copy-protection products, including backdoors that allow file decryption with fixed passwords such as 123456 and SECZURE.COM.
The investigation began with the YutianDun drive from SecZure. The product hides encrypted files from Windows Explorer and opens them only in its built-in viewers. Analysis revealed two hidden accounts: Buffer with password 123456 and System with password SECZURE.COM. Both accounts decrypted the same content identifier used to derive the AES file key, bypassing the administrator password entirely.
The cryptographic flow relied on MD5 and HMAC-MD5 to derive a 128-bit wrap key from the credential, followed by AES-128-ECB decryption of the identifier. While the design avoided obvious brute-force attacks, the presence of fixed backdoor accounts rendered the protection ineffective.
NetacFileLock
The second target was NetacFileLock. A single prompt to the AI agent directed it to analyze the crypto system using Binary Ninja MCP. Within seven minutes the agent discovered that the application initialized AES-256 with a hardcoded 32-byte key embedded in the executable. The user password was verified separately and never influenced the encryption key. Additional issues included passwords written to diagnostic logs, ECB-mode encryption without IVs, and MD5-based file signatures.
KakaSoft USB Copy Protection
KakaSoft USB Copy Protection 5.6.0.299 was examined next. The Highest protection level claimed 256-bit Blowfish encryption. The AI agent located protected files hidden in an exFAT directory tree using reserved device names and unusual Unicode characters. It then reverse-engineered a modified MD5 implementation inside the Delphi client, extracted a fixed internal password, and decrypted files without any user-supplied password.
Nexcopy Copy Secure
The final product, Nexcopy Copy Secure prepared with Drive Wizard 1.0.81, stored its content key inside an encrypted configuration file. The configuration was decrypted with a hardcoded 32-byte certificate present in both Windows and macOS viewers. The recovered key and IV allowed full decryption of protected files, while the user password served only as a UI check.
The researcher noted that the AI completed in minutes what previously required hours of manual time-travel debugging and static analysis, demonstrating both the power and the risks of automated vulnerability discovery in closed-source security products.
