# Xerberus Risk Methodology
Xerberus rates DeFi objects by mapping what they are made of.
For each Asset, Protocol, Pool, or Organisation, the methodology builds a **dendrogram**: a tree of the mechanisms that make the object work. Each mechanism is then mapped to the risk factors that are unique to that mechanism and scored against evidence-backed safeguards.
---
## The Method in 60 Seconds
1. **Choose the object.** Xerberus rates Assets, Protocols, Pools, and Organisations as separate object classes.
2. **Map its mechanisms.** The dendrogram breaks the object into the mechanisms that make it function: governance, minting, liquidation, strategy execution, redemption, legal structure, and so on.
3. **Activate only what exists.** If a mechanism is present, that branch is evaluated. If it is absent, the branch is pruned. A pool is not punished for missing a mechanism it does not use.
4. **Score the mechanism's risk factors.** Each activated mechanism contains subscores for its unique failure modes. A timelock has timelock risks; a receipt token has receipt-token risks; an organisation has legal and leadership risks.
5. **Show the decomposition.** The final rating is only the summary. The useful output is the traceable path: object -> mechanisms -> active risk factors -> evidence -> score.
The important idea is simple: **risk is not a single opaque number; it is a map of mechanisms and their weak points.**
The methodology is open and transparent. The ratings are free, and the data behind them is verifiable in the frontend. Xerberus is never paid by the entities it rates. See [[overview/Independence]] for why this matters.
---
## Start Here
> [!tip] **Investors and Allocators**
> *"What is this score telling me?"*
>
> 1. [[overview/What Xerberus Measures]] — The object -> mechanism -> risk-factor model
> 2. [[methodology/The Dendrogram]] — How the tree exposes the rating
> 3. [[overview/Rating Scale]] — How to read the number
> 4. [[limitations/Known Limitations]] — What the rating does not claim
> [!example] **Protocol Teams and DAOs**
> *"Where is the score coming from?"*
>
> 1. [[overview/Object Classes]] — Which object class applies
> 2. [[domains/index]] — Which mechanisms are in scope
> 3. [[methodology/Subscores]] — Which risk factors are scored
> 4. [[governance/Dispute Process]] — How to challenge a gate, score, or source
> [!abstract] **Domain Experts and Subscore Authors**
> *"How do I add or improve a risk factor?"*
>
> 1. [[methodology/The Dendrogram]] — Where mechanisms live
> 2. [[methodology/Subscores]] — How risk factors become subscores
> 3. [[design/Design Principles]] — The constraints for new methodology
> 4. [[governance/Subscore Author Program]] — How contributions are reviewed
> [!info] **Platforms and Integrators**
> *"How do I consume the data?"*
>
> 1. [[methodology/Scoring]] — Aggregation from subscores to ratings
> 2. [[methodology/Deterministic Hashing]] — Stable IDs for templates and subscores
> 3. [[reference/Build Status]] — Current coverage
> [!quote] **Risk Researchers**
> *"I want to understand the full methodology."*
>
> Follow the [[methodology/index|Methodology reading order]]: Dendrogram -> Subscores -> Conditional Logic -> Scoring -> Hashing. Then read [[design/Design Principles]] and [[design/Why K-Only]].
---
## Vault Structure
| Section | What It Contains |
|---------|-----------------|
| [[overview/index\|Overview]] | The object classes, mechanism model, score meaning, and independence |
| [[methodology/index\|Methodology]] | The dendrogram workflow: mechanisms, gates, subscores, scoring, hashing |
| [[domains/index\|Domains]] | The mechanism templates for each object class |
| [[design/index\|Design]] | Architectural constraints and key design decisions |
| [[examples/index\|Examples]] | Full rating walk-throughs for real objects |
| [[governance/index\|Governance]] | Evaluation process, disputes, methodology updates, author program |
| [[reference/Build Status\|Reference]] | Glossary, subscore catalog, build status |
| [[limitations/Known Limitations\|Limitations]] | Honest assessment of what the methodology cannot do |
---
## Quick Links
- [[overview/Rating Scale]] — What AAA through D means
- [[methodology/The Dendrogram]] — How the mechanism tree works
- [[overview/Object Classes]] — Asset, Protocol, Pool, and Organisation scope
- [[design/Design Principles]] — The rules governing methodology design
- [[limitations/Known Limitations]] — What the methodology cannot do
- [[reference/Glossary]] — Key terms
- [[reference/Build Status]] — Current coverage
- [[changelog]] — Version history
---
**Methodology version:** 1.1.0
**Effective date:** 2026-04-30
**Maintained by:** Xerberus DAO LLC