Design a Risk Framework and Monitoring Dashboard for XAI Collateral

Design a Risk Framework and Monitoring Dashboard for XAI Collateral

Date: 02. Jan, 2023
Status: Request for Comment
Authors: @Lavi & @dabar90
Reviewed by: @aiham.eth & @Tenzent


After initial discussions with the Silo team and community, we’re proposing to establish and maintain a custom risk framework for Silo Finance. Specifically for the evaluation of collateral for XAI credit lines.

The risk framework has two main objectives:

  1. Assess whether an asset should be enabled as collateral for XAI (i.e. opening a new credit line).
  2. Propose the initial parameter setting for the credit line (credit size, LTV, LT for the asset that can be used as collateral)

In a second step, the framework shall also be expanded to monitor the risks related to each open credit line and lending market. The monitoring system will be derived from the initial parameter setting. It will be developed, customized, and further improved over the coming months.

This proposal will define the (1) process to evaluate eligibility for new credit lines and respective parameter setting, (2) propose the initial risk framework and its grading system (3) explain the grading system and how it relates to collateral parameterization, (4) formulate the process and roadmap to establish a continuous monitoring system, (5) the estimated effort and (6) the proposed reward system.

This is a request for comment. Feedback will be implemented and the proposal adjusted before moving to a Snapshot vote.


The main objective of the newly launched stablecoin XAI is to serve as a second bridge asset (alongside ETH). This makes XAI unique. It’s designed to facilitate bridging between siloes and generate revenue for the DAO. At the same time, XAI requires this feature to ensure over-collateralization and thus stability. In other words, the extension of new credit lines is also the process of enabling new collateral for XAI. But every collateral comes with risks. Risks that can affect XAI’s price stability, user deposits, or user loans.

Hence, we propose to establish a thorough risk assessment and monitoring service, to ensure that the underlying collateral is fit for purpose. Some examples of risks to consider are an asset’s volatility, available liquidity, and slippage. But also the token’s distribution, inflation schedule, or the protocol’s technical safety and governance. Essentially, the framework should reduce the risks of bad debt and liquidations, and it will help to minimize the risk of XAI losing its peg.

There are other measures to strengthen XAI’s peg, which are outside of the scope of this proposal (e.g. interest rate setting, liquidity incentivization, efficient liquidations, and emergency mechanisms). However, a comprehensive collateral risk framework and market risk monitoring can build a very strong basis for XAI to thrive. In a later stage, additional aspects related to holistic risk management (mentioned above) shall be included.

The process of adding new collateral

To systematically evaluate the risks of XAI collateral, we propose to establish a streamlined process. The process would look like this:

It starts with an initial application to trigger the process. This can be done through an application form. The form will collect some basic information about the protocol (e.g. purpose, market cap, liquidity, link to the website, socials, Github, etc). It can also start with another DAO approaching Silo and applying to get its token added as collateral. Essentially, the process can be triggered through what the community deems most efficient.

In the second step, a Snapshot vote shall decide whether a token gets assessed or not. This step is to gather community sentiment and it serves as a first filtering mechanism. If an asset is deemed very reputable, the community can also suggest forgoing an in-depth assessment. These assets would directly jump to the parameterization step (see next section).

In the third step, we will conduct a detailed evaluation of the asset and publish our findings in the forum. The format of the evaluation is defined in this document further down. The assessment will inform the initial parameter setting. Including the size of the credit line (XAI debt ceiling), the LTV, LP, and LT.

After the initial assessment and parameter suggestions, the community and the team’s feedback is collected and implemented. For this, we propose to set a time limit (e.g. 5 days of community feedback), before moving the final executive vote that accepts or rejects the new credit line.

Once the new credit line is deployed, the monitoring of the underlying assets will start. A significant part of the evaluation will be looking at quantitative metrics. Thus, changes in the market or modifications to the underlying asset - that can impact its risks - will impact the overall health metrics of the asset. This will inform regular parameter updates. Therefore, a dashboard shall be built in close collaboration with the Silo team (more in it further down).

The Evaluation Framework Overview

The risk framework is separated into two major steps. First, a fundamental and qualitative assessment of the asset. The first objective is to understand the “intrinsic risks” related to an asset. The findings will be used to provide a suggestion on whether it can be used as collateral to borrow XAI. As mentioned above, the community can decide to forgo this in-depth assessment (via Snapshot vote), if an asset is deemed reputable. However, for most “long-tail” assets it is recommended to conduct this evaluation. On a high level, the assessment can be separated into a quantitative and a qualitative part. The following categories will be assessed:

  • Fundamentals - Does the protocol show signs of PMF, is it built by a reliable and experienced team, is there demand to use the token as collateral
  • Governance - How the protocol governed, who has access privileges, potential governance risks
  • Asset Details - Tokenomics, token distribution and supply, token functionality
  • Technical security - Audits, active bug bounty program, any severe hacks or exploits, number of days in operation
  • Composability risks - Is the token backed by other collateral (e.g. stables, LP tokens, yield-bearing tokens), does it rely on other protocols, is it a rebase token
  • Market risks - Is there enough liquidity, how volatile is the token, what’s the slippage or liquidity depth, and how many CEX listings and DEX pools are available

The second step takes the above findings as basis, to propose the initial setting of the credit line parameters:

  • XAI amount (debt ceiling)
  • Loan-to-value (LTV)
  • Liquidation threshold (LT)
  • Lquidation penalty (LP)

Essentially, the parameter setting is mostly informed by an assets related market risks (liquidity & volatility). In other words, the market risk part of the assessment (i.e. the quantitative part) is mandatory and needs to be conducted for every single asset.

Risk Assessment - Categories and Questionnaire

The following section lists all risk categories and details the qualitative questionnaire and quantitative assessment. This will be used as a basis to assess every token, which requires a complete evaluation process.

The framework is a combination of industry best practices (see references) and our own fundamental and technical research framework (as applied by PrimeRating). This initial version will be adjusted and improved over time.


  • Name - Project and token name
  • Resources - Link to website, social media, docs, blog, Github, Coingecko, and other relevant links
  • Introduction - History Brief description of the project and the token
  • Rational for Proposal - Expected value-add and borrow volume is for Silo. Other strategic benefits to consider.


  • Value Proposition - What does the protocol offer? Is it the market leader, innovator/ pioneer, or close competitor?
  • Product-Market-Fit - Number of active users (DAU/MAU), TVL, trading volume, market share, or other relevant metrics
  • Protocol Revenue - Revenue streams, supply side vs protocol revenue, revenue from integrations or other strategies
  • Team and Investors - Is the founding team well-known or anon, do they have relevant experience? Were they able to attract relevant resources and investors?


  • Governance - system Governance mechanism design, voting system (i.e.1 token = 1 vote)
  • Decisions - (DAO, multi-sig, EOA) Governance structure, who has control (community, team or individuals), can gov. decisions affect the token price?
  • Governance Attack Vectors - Governance take-over risk, interest alignment, buying influence
  • Governance Metrics - Voting Turnover, Governance participation, any concerning disputes that are absorbing the community
  • Other Risks - Other risks related to Governance

Asset Details & Tokenomics

  • Overview & Purpose - Detailed description of the token and its tokenomics. What does the asset do? Who uses it?
  • Token Inflation/ Distribution - How was the token initially distributed? What are the ongoing distribution mechanisms? How decentralized is the token (e.g. total-% of top 10 holders)
  • Extrinsic Use Cases - Is the token enabled as collateral elsewhere? What other use cases exist?
  • Access Privilege - List all of the privileged roles in the token contract. This can include whitelisted EOAs, Multi-sigs, or DAO governance
  • Other risks - Is the contract plausible? Does it have a blacklist? Can it depeg?
  • Yield generating assets - Does the token have the privilege to a revenue share?

Technical Security

  • Activity - Age of token in days/developer activity
  • Oracle - Availability of oracle price feed
  • Testing - Was the code tested? Provide test coverage
  • Audits - What audits, if any, was performed? Provide links to the reports if they exist.*
  • Bug Bounty - Does the project have an active bug bounty program?*
  • Formal Verification - List additional security and formal verification tools used in development
  • Hack or Exploits - Did the protocol suffer from any severe hacks or exploits?
  • Maturity - Number of days in operation and number of transactions
  • Permissions - All permissions across all smart contracts that are not direct on-chain voting
  • Custom Contracts - How much does the token contract deviate from a standard implementation of ERC-20?
  • Burn mechanism - Is it burnable?
  • Supply - Does it have a fixed supply? If not, who can mint?
  • Delegatecalls - Is the contract performing arbitrary delegatecalls? If the answer is yes, indicate who can make these calls and to what contracts.
  • Flash minting - Is it flash mintable? If yes, provide more information on this feature
  • Flash Loans - Is it flash loanable? If yes, indicate who offers the service
  • Immutability - Is the token contract immutable? If not, who is authorized to make upgrades, can an upgrade happen instantaneously or is there a time-lock delay? How does the upgradeability design work? Who manages it and how are upgrades performed?

Market Risks

  • Market Size - Market cap of the token and FDV (Mcap/FDV ratio)
  • Exchanges & Liquidity - The largest exchanges where the token is listed and its respective liquidity, +/- depth (define minimum 2% depth)
  • Volatility - Indicate the volatility of the token, defined as the standard deviation of log-returns for specific time frames (daily volatility + average 30-day volatility)
  • Token Inflation - Emission schedule and unlocks
  • Supply an Activity Distribution - To remove dust wallets and define holders/active users

Parameter Settings

Depending on the outcome of the above assessment, the parameters will be set. Please bear in mind that this is still under development and will be further detailed over the first quarter of 2023. In summary, the following quantitative inputs will be used to grade an asset. The grade is then used to determine its parameters:

  1. Overall liquidity on all pools and exchanges
  2. Deepest liquidity pools
  3. Market capitalization or fully dilution valuation (FDV).
  4. Volatility is measured as a 30-day average. To get an understanding of the token’s average price movements.

The table below provides a high-level example of the risk framework.

Overall liquidity Deepest liquidity +/- 5% depth Market Cap Grade
$10M+ $2M+ $50k 1B+ A
$5M+ $1M+ $25k 250M+ B
$2.5M+ $250k+ $10k 100M+ C

The full table can be seen here. Essentially, the grade that results from this assessment helps to determine the respective parameter settings for each underlying collateral. The table below displays the current version.

Grade Loan to Value (LTV) Liquidation Threshold (LT) Liquidation Penalty (LP) Size of Credit Line
Stables 90% 95% 5% $20M
A 85% 90% 5% $10M
B 80% 90% 10% $5M
C 65% 85% 15% $2M
D 50% 75% 15% $500k
E No credit line

This is still under construction and will be finalized over the coming weeks.

Risk Team Scope & Working Agreement

As eluded at the beginning, the aim of this proposal is kick-start the formation of a risk- evaluation and monitoring-group. Currently, these efforts are spearheaded by Lavi and Dabar, in close collaboration with the Silo team. In this section, we want to outline the scope and a rough roadmap for this group. We’ll start phase one (first quarter) with the proposal if accepted. The following milestones were defined for this first phase.

Phase 1

  • Support and define the collateral onboarding process (initiated in this proposal).
  • Establish a customized risk assessment, incl. parameterization for XAI collateral (initiated in this proposal). Especially parameterization based on market risks requires additional research, data management, and the creation of dedicated formulas. In a later stage, simulations can be considered as well.
  • Conduct the initial risk assessments for up to 15 potential collateral assets. Asset priority is provided by the Silo team.
  • Orchestrate and support the creation of a dedicated monitoring dashboard. To monitor the health of XAI collateral and its lending markets.

These milestones will build a strong basis for sophisticated risk-adjusted collateral management. After successfully establishing the dashboard and onboarding new collateral, phase two will start in Q2. This will be initiated with a second proposal. However, as of today, we foresee the following responsibilities for this risk working group.

Phase 2

  • Operating the health dashboard and ensuring risk monitoring related to XAI collateral.
  • Regular reviews of LTVs and LTs of silos. Specifically the ones with active XAI credit lines. Informed suggestions to adjust parameters.
  • Support of proposals to expand or contract existing credit lines.
  • Evaluate liquidity risks for ETH or XAI providers to silos where collateral is enabled, and compare liquidation risks and available liquidity (CRV squeeze situation). Current pools: USDC/XAI on Uniswap and XAI-FRAXBP on Curve
  • Continue risk assessments for new collateral.
  • Support ad-hoc projects, support governance, and act as a contact for risk-related requests.

This list is indicative and not binding. The team will first and foremost focus on building the necessary infrastructure and processes. Therefore, close collaboration with and support from the Silo core team is required.

Reward System

The reward framework should incentivize long-term alignment and sustainable growth of Silo, while minimizing the related risks. We propose to separate the reward system into two phases as well.

Phase 1

We propose a fixed monthly stipend of $20k (50/50 USD/SILO). This is for the first quarter, to build and deploy the framework, its process, and steer the building of the monitoring dashboard. This also includes the assessment of up to 15 silos.

Phase 2 (Outlook)

For the second phase, we propose a combination of fixed and variable rewards. For instance, based on the number of risk assessments performed by the team ($2k/protocol). And/or a quarterly success fee, based on the average borrowing volume (e.g. 10bps of the 30-day average borrowing volume). However, this will be detailed further in a second proposal.

Next Steps

We hope that this forum post is received well by the Silo community. We’d appreciate your input or questions. The plan is to move this proposal to a Snaphot vote, after we gather feedback and implement respective changes.

About the Authors

Lavi and Dabar are DeFi analysts, builders, and freelance writers. Previously working for PrimeDAO, Index Coop, Curve Risk Team, and other DAOs has led them down the risk rabbit hole. They have written articles related to collateral risk onboarding, performed risk assessments (ex-1 & ex-2), and written multiple fundamental rating reports for PrimeRating (ex-3 & ex-4). Most recently, they built a stablecoin risk framework for a hackathon. With a combined experience of more than 5 years in DeFi, and over 8 years in TradFi, they are very well suited to spearhead this endeavor to build a custom risk framework for Silo.


1 Like

NAY, too early for this.

After the initial Discord discussion, it became clear that the service and support offered in this proposal are not welcomed. Therefore, we’ll be withdrawing the proposal. Thanks to all who supported us. We wish Silo all the best.