Many established lending protocols, like Aave, rely on a network of external liquidators—specialized actors who monitor loans and compete to liquidate undercollateralized positions in exchange for a fee, ensuring the protocol remains solvent. In the long run, we aim to adopt a similar decentralized liquidation model, leveraging Kaspa’s high-throughput blockDAG to enable efficient, on-chain monitoring and liquidation.
For now, as we build traction, our v1 protocol handles liquidations by itself to mitigate bad debt swiftly. This approach has no impact on how users interact with or perceive their collateral—it’s simply a transparent, pragmatic mechanism to maintain stability while we scale, ensuring users’ assets are managed with the same care and security they expect. Here is how we will ensure liquidations :
- Ensuring the security of loans and assuming risks: When a user deposits collateral, it’s held off-exchange to align with our decentralization ethos. Storing collateral on centralized exchanges exposes users to risks like exchange hacks or outages—something we categorically reject. Instead, our oracle monitors collateral prices in real time. If a loan’s LTV breaches the liquidation threshold, the protocol automatically flags the collateral for liquidation. The collateral is then sent to a carefully selected exchange (chosen based on liquidity, price competitiveness, and transaction speed) to be sold swiftly, minimizing bad debt and protecting the protocol from further price drops.
- Why This Approach? This method ensures that the user funds are never compromised by defaulting to self custody. Kaskad will take the risk of slower liquidation processes in order to guarantee the safety of healthy loans. The protocol handles the sale of collateral transparently, and users are informed of the process upfront to avoid surprises.
- A Robust v1 Oracle: Our v1 oracle aggregates price feeds from major exchanges, using a weighted median algorithm—similar to industry leaders like Chainlink, Pyth, and Flare. The weight on exchanges will be determined based on trading volume, liquidity, and reliability (criteria we’re finalizing). This allows us to filter out outliers and reduce the risk of manipulation. In another posts we will detail the limitations of such models in high-scale trestles environments.
