// the take
A one-shot incident-response tool built during the May 2017 WannaCry outbreak: it scans the memory of a still-running WannaCry process to recover the RSA private key before it's overwritten, then decrypts the victim's files. It's for someone sitting in front of an actively infected, not-yet-rebooted Windows XP/2003/7 machine, not a general-purpose ransomware recovery tool.
Exploits an actual implementation bug (CryptReleaseContext not zeroing the prime numbers in memory) rather than attempting to brute-force RSA, which is why it works at all. It extends Adrien Guinet's original XP-only technique to Windows 7, meaningfully widening who it could save during the real outbreak. It never touches the original .WNCRY files, only writes decrypted copies plus the recovered key, so there's no way for it to make things worse. Single-purpose C code with no network calls and only OpenSSL as a dependency, which makes it easy to audit — worth something when the tool's job is poking around a malware-infected process.
The core requirement is fragile and binary: if the machine has been rebooted or the WannaCry process killed, which is most people's first instinct, the key is gone for good and the tool can do nothing. x64 compatibility is listed as an unresolved '?' in the README itself, so on a lot of real Windows 7 boxes it may simply not work. Zero activity since June 2017 — it only ever covered the original campaign's binaries (wcry.exe, tasksche.exe, etc.) and was never extended to later variants or other ransomware families, so its usefulness today is essentially historical. No tests or CI, and no evidence it's been re-validated against modern Windows patches, AV/EDR behavior, or ASLR changes since 2017, so you're trusting a nine-year-old binary technique to still hold on current systems.
We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →