Notable Exploits in History
Exploit Chains & Real-World Case Studies · Exploit Development track
8 min read
In this module
Exploits That Changed the World
A handful of exploits have done more than compromise their targets — they've reshaped how the whole field thinks about security. Some opened up entirely new attack classes. Some proved a thing was possible that everyone assumed wasn't. Others did so much damage that the industry had no choice but to rewrite its assumptions. These are the case studies worth knowing.
Stuxnet: The First Cyber Weapon
Discovered in 2010, Stuxnet was the first publicly known piece of malware built to break physical machinery. The target was Iran's uranium enrichment facility at Natanz, specifically the Siemens STEP 7 PLCs (Programmable Logic Controllers) running the gas centrifuges.
Technical sophistication: Stuxnet shipped with four separate Windows zero-days. Four. In a single sample. Among them: a Windows Shell LNK bug that fired the moment a USB drive was viewed in Explorer, a Print Spooler vulnerability, and a pair of local privilege escalations.
Propagation: it spread via USB drives (to clear the air gap), Windows network shares, and the Siemens STEP 7 project file format. The authors capped how far it would walk — propagation throttling was baked in to keep it from waking the neighbours.
Payload: when Stuxnet spotted a very specific arrangement — Siemens S7-315 and S7-417 PLCs driving frequency converter drives at frequencies consistent with centrifuge operation — it tampered with the PLC code. It periodically pushed the centrifuges past their rated speeds, then dropped them. While it sabotaged the equipment, it replayed normal telemetry to the monitoring stations, so operators saw nothing wrong.
Impact on security: Stuxnet proved cyber attacks could cause real-world destruction. The conversation pivoted from data theft to physical safety. Industrial control system security got a serious budget boost, and the term "cyber weapon" entered public vocabulary for good.
An air gap is the practice of physically cutting a system off from the internet and any other untrusted network. No network link, no remote attack — that's the theory. Stuxnet showed how the gap closes anyway: USB drives, infected vendor laptops, contractors carrying files in. Physical isolation makes attacks harder, not impossible.
EternalBlue: The Vulnerability That Launched WannaCry
EternalBlue (CVE-2017-0144) targeted a bug in Windows' SMBv1 (Server Message Block version 1). The NSA developed it. The Shadow Brokers leaked it in April 2017.
The vulnerability: a buffer overflow in how the Windows SMB server handled transaction requests. The kernel-mode driver srv.sys mishandled certain malformed packet sequences, opening up a pool overflow — the kernel-heap cousin of a userland heap overflow — and from there, kernel code execution on any reachable Windows box.
What made it devastating: SMB is a network file-sharing protocol, and Windows machines listen for it by default. EternalBlue was fully remote, no user interaction required, and any machine with port 445 open was reachable. Because it landed in the kernel, the moment it succeeded the attacker had SYSTEM.
WannaCry (May 2017): a month after the leak, WannaCry strapped EternalBlue to a ransomware payload and sent it round the world. Files encrypted, ransom demanded. Over 200,000 machines in 150 countries hit within days. Hospitals stopped working. Factories halted. Government agencies froze.
NotPetya (June 2017): round two with EternalBlue, this time aimed at Ukraine via a compromised accounting software update. It looked like ransomware but was a wiper — no real recovery path, the encryption was a smokescreen for destruction. Over $10 billion in damage worldwide, the most expensive cyber attack on record.
The wider lesson is about stockpiling. The NSA found the bug and sat on it for offensive use. Microsoft had shipped the patch (MS17-010) in March 2017, a month before the leak, and two months before WannaCry. The damage came not from a missing patch but from a missing rollout — and from an offensive program that lost control of its own toolchain.
Lay the dates out and the gap is plain to see. Every machine WannaCry and NotPetya hit was one that had gone at least eight weeks without applying an available update:
The Morris Worm: The Internet's First Wake-Up Call
The Morris Worm (November 1988) was one of the first worms to chew through the open internet. Robert Tappan Morris, a graduate student at Cornell, wrote it. It used three bugs:
- a buffer overflow in the Unix
fingerddaemon - a debug backdoor in
sendmail - weak-password guessing against the
rsh(remote shell) service
It wasn't meant to be destructive. A flaw in the spreading logic made it reinfect the same hosts over and over, gradually starving them of CPU and memory until they fell over. Around 6,000 machines went down — roughly 10% of the internet at the time, which gives you a sense of how small the network was.
The fallout was structural. The first CERT (Computer Emergency Response Team) was set up in response, and Morris himself was convicted under the Computer Fraud and Abuse Act, setting an early legal precedent that unauthorised access was a crime.
Technically, the worm showed something the field is still digesting: a gets() buffer overflow, the textbook example you saw in the stack overflow module, scaled up into a self-propagating network attack. Three decades on, that bug class is still putting out new variants.
Heartbleed: One Missing Check
Heartbleed (CVE-2014-0160, dissected in the TLS module) earns a slot here on impact alone. A missing bounds check in OpenSSL's heartbeat implementation let any client read up to 64KB of server memory per request.
What made Heartbleed memorable wasn't sophistication. The bug was almost embarrassingly simple. It was the blast radius. Netcraft's survey at disclosure found the vulnerable heartbeat extension enabled on about 17.5% of SSL sites, around half a million sites with trusted certificates, all exposed at the same instant. (The often-quoted "two-thirds of the web" is the share of servers running Apache or nginx, which usually use OpenSSL, not the share that was actually vulnerable.) Private keys, session cookies, login credentials — all leaking out, all at once.
Several trends accelerated in its wake:
- serious investment in OpenSSL auditing, plus the rise of LibreSSL and BoringSSL as forks
- the cert ecosystem moved towards automation (Let's Encrypt was announced not long after)
- memory safety in systems programming got promoted from niche concern to industry conversation
- fuzzing for similar one-line bugs became a much hotter research area
Pegasus: Surveillance at Scale
Pegasus, developed by the Israeli company NSO Group, represents the state of the art in mobile exploitation. It is a commercial surveillance platform that has been used by government clients worldwide to monitor journalists, activists, lawyers, and political figures.
Pegasus has evolved through multiple generations of exploit chains:
Early versions (2016): Used one-click attacks via SMS links. The chain exploited Safari WebKit bugs, then kernel vulnerabilities, to install a persistent implant.
Later versions (2019-2021): Shifted to zero-click attacks via iMessage. The FORCEDENTRY chain (described in the previous lesson) is one example. These attacks required no user interaction and left minimal forensic traces.
Technical capabilities: Once installed, Pegasus can read all messages (including end-to-end encrypted messengers, since it reads messages after decryption on the device), record phone calls, activate the camera and microphone, track GPS location, and exfiltrate files. It can do all of this silently, with the user seeing no indication of compromise.
Pegasus demonstrates the full exploit chain concept at production scale: initial access (zero-click iMessage exploit), sandbox escape, kernel exploitation and payload deployment - all automated and reliable enough to be sold as a commercial product.
Lessons from Real-World Exploits
These case studies reveal several themes:
Complexity breeds vulnerability: The most complex code (browsers, kernels, protocol implementations) has the most bugs. Stuxnet exploited four zero-days; Pegasus chains multiple bugs per target. Complex systems provide ample attack surface.
Patching matters enormously: EternalBlue's patch was available two months before WannaCry. Prompt patching would have prevented the catastrophe. The gap between vulnerability disclosure and universal patching remains one of the largest security challenges.
Defence in depth works: The progression from single-exploit compromise (Morris Worm era) to multi-stage exploit chains (modern era) shows that layered defences genuinely increase attacker costs. Modern attacks require million-dollar chains precisely because the defences are effective.
The offence-defence cycle is continuous: Each major exploit triggers defensive improvements, which require new offensive techniques, which prompt new defences. This cycle drives the evolution of both attack and defence capabilities.
Understanding real-world exploit chains provides context for all the individual techniques studied in earlier modules. Buffer overflows, use-after-free, ROP chains, heap exploitation, and kernel bugs are components that combine into the devastating attack capabilities that shape modern security.