DeFi ProtocolDevelopment &Engineering

Solvency modelling is the first deliverable on every DeFi engagement, not a side artefact. Before contract code is written, we document what solvency means for your specific protocol - the mathematical conditions that must hold for the protocol to remain solvent under every adversarial scenario we can simulate. For lending markets, that's collateral coverage under price shocks, liquidation efficiency and bad-debt absorption paths. For AMMs, it's invariant preservation under donation attacks and rounding edge cases. For perpetuals, it's funding rate convergence and insurance fund adequacy under cascading liquidations. These conditions are encoded as on-chain invariants tested via Echidna and Foundry fuzzing - and where the cost of failure justifies it, formally verified via Certora or Halmos. We pair this with adversarial simulation against historical exploit patterns: bZx-style oracle manipulation, Cream-style donation attacks, Mango-style governance capture, Euler-style donation reentrancy. If a published attack pattern works on your protocol, we find it before an auditor or attacker does. We coordinate economic risk parameter modelling with firms like Gauntlet or Chaos Labs when scale and complexity warrant - but the engineering invariants are non-negotiable on every engagement.

AMMs (constant product, stable, concentrated liquidity), lending and perpetuals, vaults and aggregators, liquid staking interfaces, stablecoin cores, oracle aggregation and cross-chain messaging integrations. We ship across Ethereum, Arbitrum, Optimism, Base, Polygon, BNB Chain, Avalanche and Solana where your liquidity and security model allows - with explicit trade-offs documented before build.

We select oracle topology against manipulation cost, staleness tolerance and failover paths: Chainlink, Pyth, RedStone, TWAPs and hybrid aggregation with circuit breakers. Fallback and pause semantics are documented; heartbeat and deviation thresholds are tuned to asset liquidity.

Yes - batch auctions, commit-reveal flows, Flashbots-aware routing and integrations with private RPC or auction venues where your product model supports them. We document residual MEV surface honestly after implementation.

We specify triggers, partial liquidations, auction types, keeper incentives and bad-debt socialisation paths - then fuzz under volatility paths and oracle delay scenarios. Liquidation correctness is a first-class invariant, not an afterthought.

Core practice is contract and protocol engineering with documented parameter rationale. For large programmes we coordinate with economic risk partners (e.g. Gauntlet, Chaos Labs) for simulation-backed recommendations - your counsel and treasury own final numbers.

We ship the audit preparation pack, coordinate with your firm or our partner network, triage every finding, implement root-cause fixes and add regression tests through sign-off. Formal verification is scoped to invariants where proof materially reduces uncertainty.

Yes - we align scope, severity rubrics and disclosure timelines with Immunefi, Code4rena, or your preferred venue, typically starting during adversarial validation so findings flow into a single triage process.

Discovery through audit-ready core often runs 10-22 weeks depending on market count, oracle complexity and formal verification depth. Core team: lead protocol engineer, security reviewer, DevOps for deploy and monitoring - scaled per scope.

Market designs, asset universe, target chains, oracle preferences, liquidation philosophy, governance model, intended listing or integration partners and target launch window. We respond within one business day with a scoped recommendation.