Private key compromise
If an attacker gets a private key, they can do anything its owner can: move funds, upgrade contracts, change parameters. Keys are compromised through malware on signers' machines, phishing and fake job offers, insecure storage, insiders, and weak key generation. Several of the largest crypto thefts, including Alphapo and Stake.com in 2023, were linked to stolen keys.
How keys are compromised
- Malware on the machines of developers, operators and signers, often delivered by phishing or fake recruiters.
- Insecure storage in code repositories, environment variables, cloud storage or chat.
- Insiders with access to key material or backups.
- Weak generation, such as poor randomness or vanity address tools.
- Physical access to hardware wallets, seed backups and signing devices.
- Compromised signing interfaces that show one transaction and sign another, as in the Bybit theft.
Limiting the damage
- Multisig or MPC so no single key is enough.
- Timelocks on large transfers and contract changes.
- Out-of-band verification for high-value transactions.
- Hardware security modules and hardware wallets for key storage.
- Strict separation between signing machines and everyday work.
- Testing of the whole signing path: wallet penetration testing, physical testing and red teaming.
Questions
Frequently asked questions
Does a hardware wallet solve this?
It protects the key from being copied, but not from being used to sign a malicious transaction. Signers must be able to see and verify what they are signing.
What is the single most useful control?
Making sure no single key, person or machine can move significant funds or change contract logic. Multisig or MPC, timelocks and out-of-band verification together achieve that.
Published Updated