In the world of hardware crypto wallets, Coldcard has traditionally been considered the gold standard of security. However, a recent public experiment conducted by the well-known researcher Cole has forced me and many of my colleagues to seriously reconsider this point of view. The test results turned out to be not just alarming—they demonstrate that even the most "impregnable" devices can be compromised at a frightening speed.
Methodology: How the Research Was Conducted
During the experiment, the researcher used a Coldcard Mk3 with firmware version 4.1.8—a model he rightly calls one of the most vulnerable. The problem, however, is not limited to this version alone: the vulnerabilities also affect the newer Mk4, Mk5, and even Q models. This means we are dealing with a systemic issue rather than an isolated bug.
The researcher generated a secret phrase on the device and created five different wallets based on it, each with an identical balance of 10,800 satoshis. The configuration was as follows:
- Standard wallet — without an additional passphrase.
- With a 1-word password — from the standard BIP39 list.
- With a 2-word password — a combination from BIP39.
- With a 3-word password — a combination from BIP39.
- With an account number — an arbitrary index from 1 to 9999.
All wallets used addresses in the bc1q format and a single key derivation path. For the purity of the experiment, the funds were sent in a single transaction. An interesting detail: when attempting to directly transfer public keys to the Sparrow program, the device suddenly stopped responding to commands, forcing the researcher to resort to exporting CSV files via an SD card.
Results: The Speed of the Attack Is Staggering
The results turned out to be shocking. The base wallet without a password was drained almost instantly—the funds were withdrawn to a new unknown address before the researcher even had time to open the mempool. This irrefutably proves that attackers had pre-loaded the compromised keys into their bots and automated monitoring systems.
The remaining four accounts, including those protected by passwords and account numbers, remained untouched. However, the very fact that the base secret phrase turned out to be completely vulnerable calls into question the entire concept of "cold storage." The researcher deliberately refused to recover the funds through the RBF mechanism in order to preserve the purity of the experiment.
My professional opinion: this experiment is a stark reminder that hardware security is not a panacea. Even the most reliable devices are useless if the user neglects additional layers of protection—passphrases, PIN codes, or regular key rotation. In the current reality, where automated bots scan the blockchain for the slightest mistake, relying solely on "hardware" is an unforgivable luxury.