Safety Analysis
Safety means no two honest nodes commit different blocks at the same height. We test for equivocation, quorum bypass, double-signing, and fork-choice manipulation under adversarial voting power.
Service
Consensus Security
Consensus is where blockchains live or die. We model validator behaviour, leader selection, quorum mechanics, and fault tolerance to identify safety and liveness violations before mainnet.
Liveness & Finality
A protocol that cannot make progress is unusable. We stress timeout parameters, leader rotation, network partitions, and censorship scenarios to find stalls, liveness leaks, and finality delays.
Leader Selection
Leader-driven protocols are vulnerable to targeted attacks on the proposer. We review randomness sources, lookahead predictability, rotation fairness, and recovery from leader failure or eclipse.
Fault Tolerance Bound
We verify that the claimed fault tolerance matches the actual quorum math. A gap between claimed and actual tolerance is a critical finding because it changes the entire adversarial cost model.