
Operation ShadowCloud: cloud service–based backdoors targeting government networks in Asia and Africa. Part 1
Security researchers have identified a new intrusion campaign, named Operation ShadowCloud, that abuses a legitimate remote administration tool to drop two brand-new backdoors.
The campaign, active since early 2026, uses a legitimate MeshCentral Agent to install malicious components on compromised systems. This agent, found in two instances within a Taiwanese organization, establishes a command-and-control session with a remote server at wss://27.102.66[.]162:443/agent.ashx.
Once connected, the threat actor drops three items on the endpoint: a legitimate binary for DLL sideloading, a malicious loader DLL, and an encrypted file containing the core payload.
The attackers use trusted executables to sideload malicious loader DLLs. They identified loaders named jli.dll, SbieDll.dll, python39.dll, WINMM.dll, and libcef.dll, which were dropped onto the host and then implicitly loaded by the respective legitimate binaries.
The legitimate binaries used include the OpenJDK platform binary (javaw.exe), Sandboxie Start (start.exe), a Python application (pythonw.exe), and the Sysinternals Screen Magnifier tool (zoomit.exe).
Two loader types with different encryption
We discovered two distinct loaders, each using a different method to decode the encrypted backdoor payload. In both cases, the loader first decrypts the payload and then injects the resulting code into the current process’s memory.
Loader (Type 1) employs a hybrid scheme: first running the data through an RC4 stream cipher and then applying an XOR mask. Loader (Type 2) simplifies the process by using a straight XOR operation alone.
The Type 1 loader reads the file config.ini located next to it using native NtCreateFile/NtReadFile calls. It then Base64-decodes the data with a custom character set, decrypts the first layer of encryption with a 16-byte RC4 key, and decrypts the second layer with a 16-byte XOR key.
The loader transforms the decrypted backdoor bytes by abusing the RtlEthernetAddressToStringA and RtlEthernetStringToAddressA APIs to evade defenses. It converts the decrypted backdoor bytes into an Ethernet MAC-address string, then translates the string into a six-byte binary format.
The resulting bytes are written to memory newly allocated via an AllocADsMem API call. The loader changes the memory protection to PAGE_EXECUTE_READWRITE using VirtualProtect and jumps to the MZ header, which acts as a trampoline into bootstrap code.
The bootstrap code dynamically resolves API addresses and uses CRC32 to compute a hash of each API name. Both the import library name and the associated API name are XOR-decrypted with a hardcoded 7-byte key before resolution.
GitDoor and GraphDoor backdoors
The Type 2 loader, internally named SbieDll_st.dll, performs several checks on the infected machine before attempting to prepare and execute the backdoors. It first searches for a specific byte pattern inside KERNEL32.DLL and KernelBase.dll.
If the pattern is found, the code installs a vectored exception handler and clears the hardware breakpoint by zeroing the CONTEXT.DR7 register. It then verifies that the current process’s name matches a hard-coded value.
After passing these checks, it reads an encrypted blob from config.ini and Base64-decodes the data. The resulting data is a backdoor executable named GitDoor (MD5: 552f0a9cc5fc3755721fa49be05e8035).
The loader allocates memory in the current process, copies the decrypted backdoor code into that region, and starts a new thread at the MZ header. This acts as a trampoline into a bootstrap routine identical to that of the Type 1 loader.
Both GitDoor and GraphDoor incorporate several encryption and obfuscation techniques, including Control Flow Flattening, which obscures their genuine execution flow. They also make heavy use of indirect calls, encapsulating routines in several layers of helper functions.