A risk rating protocol for DeFi Vaults — independent by design, transparent by default, and governed by the investors it serves.
Single scores hide critical failure points. Our Intrinsic Risk framework maps the complete structural hierarchy of a vault to reveal where fragility truly hides.
Uncovering opaque dependencies across nested yield protocols.
High returns often mask unsustainable economic loops and structural debt.
Measuring the specific fragility of modular asset and liquidity stacks.
Providing a standard benchmark in a market currently built on trust.
Every calculation is public and documented. No black-box algorithms or hidden biases. Full intellectual transparency.
Because the math is open, the data must be too. We pull directly from on-chain truth to feed our models. Trust but verify.
Raw data is objective, but its importance is subjective. We use staked consensus to weight risk factors—shifting power from committees to the market.
The majority of ratings are funded by the very issuers they assess, embedding a structural conflict of interest that undermines market integrity.
Xerberus is funded by you, the investor. We answer only to the truth of the data. This radical independence creates a standard of transparency that is as humbly honest as it is useful.
We never accept payment from asset issuers or DeFi protocols to be rated.
Our revenue is tied directly to providing the most accurate risk assessments for capital allocators via our investor-funded ratings.
Intrinsic risk cannot be measured by algorithms alone. It requires the context of domain experts who understand the subtle nuances of decentralized finance. We bridge code and economics.
Xerberus rates three object classes independently: Assets (tokens like WETH, stETH, USDC), Protocols (technical systems like Aave v3, Morpho, Lido), and Organisations (the entities behind them). Each class has its own dendrogram template with dedicated risk domains. When we say 'Safer DeFi Vaults,' we mean the full stack — the underlying assets, the protocols they interact with, and the organisations operating them — rated separately so you can see exactly where risk lives.
Every subscore in the dendrogram is required to reference a real historical incident — no subscore exists without evidence that its failure mode has actually occurred (Terra/Luna, Euler, Mango Markets, FTX contagion). The framework is designed so that the mechanisms flagged by low scores correspond directly to the mechanisms that have failed in practice. We publish case studies showing how ratings decompose known incidents after the fact, and as our live rating history grows, we will publish real-time tracking of rating signals against market events.
Risk ratings will always remain open and accessible. The Xerberus Protocol is designed as public infrastructure — its ratings are available to everyone. However, the protocol also offers convenience services such as compute for running models and operates optimisation markets that generate fees. Xerberus is a non-profit, and all fees raised are reinvested into the maintenance and development of the protocol.
