Richmond, BC Cryptocurrency Theft: CAD $10M Stolen via Police Impersonation — 2023
BlockedPhysical coercion was applied — the custody structure did not protect against forced transfer.
In 2023, a cryptocurrency holder in Richmond, British Columbia fell victim to an escalated physical attack that demonstrated the vulnerability of self-custody holders to targeted social engineering. Criminals impersonating police officers gained entry to the victim's home by exploiting public trust in law enforcement. Once inside, they employed coercion to force access to approximately CAD $10 million in cryptocurrency holdings. The attackers' method represented a significant evolution in wrench attack tactics: rather than relying solely on technical exploitation or simple burglary, they leveraged social authority and deception to bypass initial physical barriers.
TRM Labs documented this incident in a report on wrench attacks, identifying it as one of the most brazen physical cryptocurrency theft cases in Canada. The case underscores a critical vulnerability in self-custody models: while private keys may be securely stored cryptographically, the human operator and their physical security remain potential failure points. The attacker's success depended entirely on social engineering and the victim's location being known, suggesting either prior reconnaissance or information leakage. This incident occurred during a period of increasing organized crime sophistication in targeting high-value cryptocurrency holders, particularly in jurisdictions with significant crypto adoption.
| Stress condition | Coercion |
| Custody system | Unknown custody system |
| Outcome | Blocked |
| Documentation | Present and interpretable |
| Year observed | 2023 |
| Country | Canada |
What custody structure can and cannot protect against coercion
The relevant structural question is not whether a custody setup can prevent coercion — it typically cannot — but whether it can limit what an attacker can obtain through coercion. A setup where the holder has sole knowledge of all credentials, with no geographic distribution and no multisig threshold, gives an attacker everything they need by controlling one person. A setup where credentials are geographically distributed, where multisig requires coordination with parties in other locations, or where a passphrase-protected decoy wallet exists, limits what any single physical attack can yield.
Observed cases in this archive range from violent home invasions and kidnappings to subtler forms of coercion: legal threats, family pressure, business disputes that escalated. The outcomes depend on whether structural protections existed and whether they held under pressure. Setups with no geographic distribution or threshold requirements produced the worst outcomes.
The legal dimension adds complexity: transactions executed under coercion are technically valid. The blockchain cannot distinguish voluntary from involuntary signatures. Recovery after a coerced transfer depends entirely on legal processes — identifying the attacker, prosecuting, and attempting asset recovery — which is slow, expensive, and uncertain.
The most effective structural protection against coercion is geographic key distribution combined with a signing threshold that cannot be met from one location. An attacker who controls one person in one place cannot force a transaction that requires coordination with key holders in other jurisdictions. This protection requires accepting coordination overhead during normal use.
Translate