Imagine a situation where two people need to exchange sensitive details, such as a meeting location or a critical update, while knowing that outside parties may already possess the means to read their usual messages. In such circumstances, the question arises whether any form of truly private exchange can still occur through the same tools. Research into techniques that operate within established secure messaging systems offers a path forward by demonstrating how additional information can travel undetected alongside regular traffic.[1]
Secure messaging applications like Signal rely on end-to-end encryption to keep conversations private. This approach functions much like placing a letter inside a box that locks automatically, with only copies of the key held by the sender and the intended recipient. The messaging service itself, along with any network operators or potential interceptors, sees only an opaque package that reveals nothing about the contents without the proper key. Signal enhances this protection through its Double Ratchet mechanism, which generates fresh encryption material for ongoing conversations. Each new message or exchange incorporates updated keys, ensuring that access to one set of keys does not automatically expose earlier exchanges or allow prediction of future ones. This design supports both forward secrecy, which limits damage from a single compromise, and post-compromise security, which helps restore protection after a breach.
Governments in various countries have increased pressure on such systems by demanding what is euphemistically called "lawful access". These requirements often take the form of mandates compelling companies to enable the decryption of user communications for law enforcement purposes. Methods under discussion include the introduction of backdoors into encryption processes, the maintenance of key escrow arrangements where authorities hold duplicate decryption capabilities, or provisions for direct device searches that extract stored keys. Legislation such as the European Union’s Digital Services Act and the United Kingdom’s Online Safety Act illustrates this trend in democratic settings, where platforms face obligations related to content oversight and potential access that can intersect with encryption protections. Comparable frameworks appear in other contexts as well, including systems such as Russia’s SORM laws, which enforce extensive monitoring of communications infrastructure. While presented as tools for addressing serious crimes or public safety, these measures carry risks because the technical capabilities created for one purpose can extend to broader surveillance, and distinctions between legitimate and overreaching use become difficult to enforce once the infrastructure exists.
Against this backdrop, a recent paper examines whether users can maintain an additional layer of secrecy even when authorities obtain the ability to read ordinary messages. The work, titled “Crypto Wars in Secure Messaging: Covert Channels in Signal Despite Leaked Keys” by Rosario Giustolisi, Gabriele Lenzini, Chuanwei Lin, Mohammadamin Rakeei, and Andy Rupp, focuses on this scenario. It demonstrates that hidden information can be incorporated directly into the normal operation of Signal in ways that remain indistinguishable from standard activity.
The approach employs anamorphic encryption, a method that allows one party to embed a secondary message within the outputs of an encryption process. Unlike traditional steganography, which typically conceals data by making subtle alterations to an unrelated carrier such as a photograph or sound recording, anamorphic techniques modify the cryptographic values themselves. The resulting outputs continue to appear random and valid under normal inspection. An observer who possesses the standard decryption key can recover the visible message but gains no indication that additional content exists. Recovery of the hidden portion requires a separate piece of information, often called a double key in this context, or depends on the observer lacking certain intermediate secrets. This creates a channel where the hidden message travels with the regular traffic yet remains protected by the cryptographic structure.
The researchers applied these ideas to Signal’s Double Ratchet by targeting two specific elements that arise naturally during normal use. One construction embeds the hidden information inside the public keys exchanged as part of the ratchet process. These keys help refresh the encryption material and maintain security properties over time. By generating the keys through a modified procedure that incorporates the extra secret, the sender produces values that function identically to ordinary keys from an external perspective. The receiver, equipped with the matching extra secret, can isolate the embedded portion through a straightforward computation. This method remains effective even under conditions of complete key disclosure, where authorities hold every standard secret used for ordinary decryption. Users can establish the necessary extra secret through direct means, such as an in-person exchange of a passphrase or QR code, or through a one-sided version where the recipient publishes a corresponding public value that any sender can reference without prior coordination.
The second construction places hidden information inside the authentication tags that accompany each message. These tags serve as a verification step to confirm that the message arrived unaltered. Signal derives the keys for these tags from multiple independent sources within the ratchet state. When an adversary obtains only partial access to that state, as occurs in many practical surveillance setups where full ongoing capture of every ephemeral value proves costly or incomplete, the authentication tag key stays out of reach. The sender can then encode the hidden message into the tag in a manner that preserves its validity for ordinary verification while allowing extraction by the intended recipient. This approach requires no pre-established extra secret and aligns with situations where monitoring captures the visible message content but leaves gaps in the supporting key material.
Both constructions were implemented in a Signal-compatible library and an Android client application. Testing showed that the authentication tag method introduces negligible additional processing time or resource use compared with unmodified Signal. The public key method, after an optimization that replaces exhaustive search with a table lookup during extraction, adds only a small delay measured in milliseconds for short hidden payloads. Both operate without any alterations to the central servers that relay Signal messages, preserving compatibility with existing deployments and delivery behaviors, including handling of out-of-order or delayed messages.
The hidden payloads remain limited in size, supporting roughly 32 bits per key-exchange step in the optimized public-key approach and up to 64 bits per message in the tag approach. Such capacity suits brief items like identifiers, coordinates, status indicators, or a fresh key that could bootstrap an entirely separate secure session. It does not extend to bulk data transfer. The techniques focus on content confidentiality and undetectability rather than metadata protection, so details such as communication partners or timing patterns stay observable through other means. In public-key construction, exposure of the extra secret allows reconstruction of previously hidden content from stored transcripts. In tag construction, the acquisition of the complete key material removes the protection. The base versions address passive observation; extensions that bind messages to the transcript history provide additional resistance against active interference such as injection or modification.
These outcomes are directly relevant to individuals who rely on secure messaging under elevated risk. Journalists maintaining contact with sources, activists organizing activities, legal professionals handling sensitive client matters, and others in vulnerable positions gain options that remain effective even after standard encryption protections have been bypassed via disclosure mechanisms. The work illustrates that weakening or circumventing primary encryption does not automatically eliminate all avenues for private exchange when the underlying protocol structure itself can carry additional protected information.
In real-world terms, the findings suggest that technical responses to surveillance demands can incorporate layered protections without requiring entirely new applications or infrastructure. Users facing partial monitoring scenarios can apply the tag-based method without prior setup, while those able to establish an additional secret beforehand can employ the key-based method to improve resilience against full access. The research also outlines boundaries, such as the absence of forward secrecy for the hidden channel in its current form and the need for separate handling of active adversaries, which points toward targeted extensions in subsequent development.
Looking ahead, the constructions open avenues for integration into privacy-focused tools and for further refinement, including mechanisms that evolve the extra secret over time to limit long-term exposure and explorations of increased capacity while retaining low overhead. As legislative and technical pressures on encryption continue across different governance models, the demonstrated ability to embed undetectable channels inside widely used protocols underscores the ongoing interplay between access requirements and methods that preserve user options. This body of work equips technically aware users and developers with concrete building blocks for maintaining communication privacy in environments where conventional guarantees are challenged.
This article is shared at no charge for educational and informational purposes only.
Red Sky Alliance is a Cyber Threat Analysis and Intelligence Service organization. We provide indicators of compromise information (CTI) via a notification/Tier I analysis service (RedXray) or an analysis service (CTAC). For questions, comments, or assistance, please contact the office directly at 1-844-492-7225 or feedback@redskyalliance.com
- Reporting: https://www.redskyalliance.org/
- Website: https://www.redskyalliance.com/
- LinkedIn: https://www.linkedin.com/company/64265941
Weekly Cyber Intelligence Briefings:
REDSHORTS - Weekly Cyber Intelligence Briefings
https://attendee.gotowebinar.com/register/7855487668891299929
[1] https://six3ro.substack.com/p/hidden-signals-within-signals-how
Comments