ScatterChat was an encrypted instant messaging client that the Hacktivismo group released in July 2006. It was discontinued long ago, and nobody maintains it or fixes its security problems, so it should not be relied on for anything sensitive. It tried to give activists and dissidents encrypted chat with optional routing through Tor. Today most people meet that need with Signal, which uses the independently analysed Signal Protocol, and they add Tor Browser from the Tor Project when they need to hide which sites they visit.
The advisory at a glance
ScatterChat Advisory 2006-01 was a non-technical report about a weakness in the program’s cryptography, published on 11 August 2006. The flaw is tracked as CVE-2006-4021. The National Vulnerability Database gives it a CVSS 2.0 base score of 2.6, which counts as low severity.
| Detail | Value |
|---|---|
| Advisory | ScatterChat Advisory 2006-01 |
| Published | 11 August 2006 |
| CVE | CVE-2006-4021 |
| Affected versions | ScatterChat 1.0.x |
| Credited researcher | Steven Murdoch, University of Cambridge |
| Weakness | Birthday attack on custom padding in ECB-mode message encryption |
| CVSS 2.0 score | 2.6 (low) |
| Secure file transfer | Not affected |
What Steven Murdoch found
Steven Murdoch, a security researcher at the University of Cambridge, found a theoretical weakness in how ScatterChat encrypted messages. An eavesdropper who collected a very large number of messages from one conversation had a small chance of spotting a repeated pattern in the encrypted traffic.
The non-technical report described this in simplified terms. It said an eavesdropper had a 1% chance of finding a small 12-character pattern once more than 1,100 messages had been exchanged in a single conversation. The technical advisory gave the exact figures: two 16-byte segments could be matched with 1% probability, but only after at least 1,188 instant messages in a single, uninterrupted session.
How the weakness worked
The problem came from the encryption mode. ScatterChat encrypted messages in ECB (Electronic Codebook) mode with a custom padding scheme. ECB encrypts each block on its own with the same key, so two identical blocks of plaintext produce two identical blocks of ciphertext. ScatterChat’s padding was meant to stop that from happening. The birthday paradox, though, means that collisions between blocks become likely far sooner than intuition suggests once enough blocks pile up. That is the kind of match the advisory described.
Modern designs avoid this whole class of problem. Cryptographers now advise against ECB entirely. They use authenticated modes such as AES-GCM, or the per-message keys that ratcheting protocols like the Signal Protocol create.
Why patterns matter in encrypted chat
Finding patterns in encrypted conversations can let an eavesdropper link messages together and work out sensitive information without breaking the encryption itself. In ScatterChat’s case the exposure was very narrow. Only small matching segments could be found, with low probability, in extreme conditions.
The project concluded that the flaw posed no practical risk to ordinary users. Given the 1,188-message threshold and the 1% probability, that judgement was reasonable for 2006.
What the flaw did not allow
The flaw did not let an eavesdropper decrypt messages. It also did not reveal a user’s identity when the anonymity feature, which routed traffic through Tor, was switched on. Secure file transfers were not affected.
On that basis, the project said in 2006 that ScatterChat v1.0.x remained safe to use in the vast majority of cases. It advised extra caution for high-risk, non-technical users, especially people behind national firewalls. That 2006 assessment is history now. The software has been abandoned for many years, its cryptography was never brought up to modern standards in a maintained release, and nobody would fix any newly found flaw. The NVD record noted that the vendor planned to address the issue in a future ScatterChat 2.0.
The project’s disclosure policy
The ScatterChat developers published the advisory even though they judged the issue to have no practical impact. Their stated policy was to tell the public about any defect, theoretical or practical. Disclosing low-severity flaws openly is still considered good practice, and it is one sign of a trustworthy security project.
The project announced new versions and advisories through a Scatterchat-announce mailing list. That list and the project behind it are no longer active.
<http://www.scatterchat.com/mailinglists.html>
The project also published a more technical version of this advisory, which went out on the Bugtraq security mailing list in August 2006:
<http://www.scatterchat.com/advisories/2006-01_tech.html>
How ScatterChat was received
ScatterChat was written by J. Salvatore Testa II. It was announced at the H.O.P.E. Number Six conference in New York City in July 2006. It was a fork of the Gaim messenger (later renamed Pidgin) and added encrypted chat, key management and encrypted file transfer, with optional Tor routing to resist traffic analysis.
Security commentators were wary from the start. LWN pointed out that Off-the-Record Messaging (OTR) plugged into Gaim as a module instead of forking the codebase, and that OTR had been audited much more carefully than the brand-new ScatterChat. Advisory 2006-01 arrived within weeks of the launch.
What people use for the same need today
For private messaging, Signal is the usual recommendation from the EFF’s Surveillance Self-Defense guides. It encrypts messages and calls end to end by default. Its protocol has been formally analysed by academic researchers. Signal added post-quantum protection to its first key exchange with PQXDH in 2023, and in October 2025 it added the Sparse Post-Quantum Ratchet (SPQR), creating what it calls a triple ratchet. Cryspen helped verify parts of the core SPQR code, though that work covers only part of the codebase.
For hiding which sites you visit and where you connect from, Tor Browser from the Tor Project routes your traffic through three volunteer relays.
| Need | ScatterChat (2006, abandoned) | Today’s maintained option |
|---|---|---|
| Message encryption | Custom ECB-based scheme | Signal Protocol with forward secrecy and post-quantum ratchet |
| Maintenance | None for many years | Active development and public security research |
| Hiding network location | Optional Tor routing | Tor Browser from the Tor Project |
| Independent analysis | Little before release | Repeated academic and formal analyses |
None of these tools makes anyone fully anonymous. Signal still needs a phone number to register, although usernames let you avoid sharing it with contacts. Tor does not protect you if you log into accounts that identify you, and neither tool helps if your device itself is compromised.
FAQ
Is ScatterChat still safe to use?
No. ScatterChat was abandoned many years ago and gets no security fixes. Its 2006 encryption design, including ECB mode, falls short of modern standards. Use a maintained messenger such as Signal instead.
Did CVE-2006-4021 let attackers read ScatterChat messages?
No. The flaw only let an eavesdropper match small 16-byte segments of encrypted traffic, with 1% probability, after at least 1,188 messages in one uninterrupted session. It did not allow decryption, and it did not reveal users who had the Tor anonymity feature switched on.
What is ECB mode and why is it considered weak?
ECB, or Electronic Codebook, encrypts each block of data on its own with the same key. Identical plaintext blocks therefore come out as identical ciphertext blocks, which can reveal patterns. Modern software uses authenticated modes such as AES-GCM or protocols that create a fresh key for every message.
Who was behind ScatterChat?
Hacktivismo released it, a group connected to the Cult of the Dead Cow. J. Salvatore Testa II wrote it, and it was announced at the H.O.P.E. Number Six conference in New York in July 2006. It was aimed at human rights activists and political dissidents.
Is scatterchat.com still the ScatterChat project’s website?
No. The domain is now an independent guide to privacy, anonymity and secure messaging under new, unrelated ownership. It has no connection to Hacktivismo or the original developers, and it does not distribute the old software.
What should high-risk users use instead of ScatterChat?
The EFF’s Surveillance Self-Defense guides point to Signal for encrypted messaging, and the Tor Project’s Tor Browser for anonymous browsing. Before relying on either one, read what it does not protect, such as phone number registration, contact metadata and a compromised device.
Sources
- Bugtraq archive: ScatterChat Advisory 2006-01 (seclists.org)
- Bugtraq archive: ScatterChat Advisory 2006-01 (Openwall)
- CVE-2006-4021 (CIRCL vulnerability database)
- CVE-2006-4021 (Aqua Vulnerability Database)
- LWN.net coverage of ScatterChat, 2006
- Hacktivismo (Wikipedia)
- Cryspen: Helping Secure Signal’s Post-Quantum Transition
- Heise: Signal secures its chats against quantum computers
- EFF Surveillance Self-Defense: Tool Guides