Silo Finance Review

How Silo Finance scores in crypto lending

This entry rates Silo v3, the current version of Silo's isolated two-asset lending markets, in our lending ranking. The earlier entry described Silo v2; the documentation now covers v3, and this pass re-reads every criterion against it. The basis is desk research of Silo's documentation and the silo-contracts-v3 repository on GitHub, checked on November 1, 2026. No position was opened. The category uses five criteria, each marked 0 to 10 and weighted equally, as set out in our ratings methodology.

The overall score is the plain average: (6 + 3 + 7 + 5 + 9) ÷ 5 = 30 ÷ 5 = 6.0, down from 6.2. One mark moved. Audit disclosure fell from 10 to 9 because the v3 reports, which cover the code now in use, are listed without dates. The other four marks were checked against v3 sources and hold.

Strongest criteria: audits and liquidation terms

Audit disclosure (9)

Silo's audits page lists a v3 audit report from each of Certora, Enigma Dark and Spearbit Cantina, plus a Certora formal verification report. Below that it keeps the v2 record. Certora's v2 audit is dated November 2024 and May 2025, Enigma Dark's May 2025, and the Cantina contest ran from January 13 to February 10, 2025. Sigma Prime reviewed v2 as well. Silo Vaults went through four rounds between March and April 27, 2025, with Sigma Prime, Certora and Code4rena. A Certora leverage audit and formal verification date from July 2025, and the rate model has Spearbit and Certora reports. The rubric asks for reports with firms named and dated. The v3 entries name the firms but give no dates, and that gap is why the mark fell to 9. Audits do not insure deposits. Silo's own leverage contract was exploited on June 25, 2025; Certora's incident report says the funds lost were Silo's, not users'.

Liquidation terms (7)

The liquidation documentation says the fee is "set per market by the deployer and cannot be changed afterward". A borrower in an existing market cannot have the penalty raised on them, and the factory caps it at 30% for new markets. Silo v3 adds a second engine, the collateral-debt swap (CDS). Above a market's CDS threshold, debt is written off and lenders receive the collateral instead of a DEX sale. The docs publish the split: lenders get the full debt plus 80% of the liquidation fee, the liquidator 20%. With a 20% fee on 1,000 USDC of debt, that is $1,160 to lenders and $40 to the liquidator. Collateral-sale liquidations are partial, with full liquidation allowed where the remainder would be dust. The mark stays at 7 because the DEX and CDS thresholds, and each market's actual fee, are read from the SiloConfig contract rather than published as figures.

Weakest criteria: licence and governance

Licence & code openness (3): the longest delay here

The LICENSE file names Distributed Coders Inc. as licensor of Silo V3 Core and grants production rights only to SiloDAO, "for the operation and governance of the Silo protocol". Anyone else may copy and modify the code for non-production use only. It converts to GPL-2.0-or-later on January 31, 2031, or four years after a version's first release if that comes first. A delayed-open licence scores mid-table under the rubric. This one sits below Aave and Kamino at 4 because its conversion date is more than four years away.

Governance & timelock (5): immutability without a stated owner

The architecture page says silos "are immutable to change by default after deployments except when the deployer uses upgradable settings". That protects borrowers in default markets, but a reader has to check each market's configuration to know which kind it is. The SiloFactory contract inherits Ownable2Step, and its owner can change the fee limits for future markets up to the 50% cap. The documentation does not name who holds that ownership, and no timelock is described. The mark is unchanged.

Fee disclosure (6): caps, not values

The factory hard-codes a 50% ceiling on any fee and sets these limits at deployment: DAO fee 5% to 50%, deployer fee up to 15%, flash-loan fee up to 15% and liquidation fee up to 30%. These are published ceilings. The rubric scores them above undocumented fields but below fees published as figures, and each market's actual values have to be read from its configuration. Our guide to reading a DeFi lending market shows where to find them.

What the score does not cover

• Rates and yields. Silo's interest rate models move with utilization and were not measured.
• Silo Vaults and curators. Vault allocation and curator fees are outside the five criteria.
• Oracles and hooks. Each market chooses its own oracles and optional hooks; their quality is not scored.
• Protected deposits. Silo's non-borrowable deposit type, which can always be withdrawn, is a design feature the rubric does not measure.

Silo Finance vs Spark and Aave

Spark scores 6.2, two tenths above Silo. It marks higher on licence (6 against 3), liquidation terms (8 against 7) and governance (6 against 5), and lower on fees (4 against 6) and audits (7 against 9). Aave scores 6.6. It matches Silo closely on audits (10 against 9) and leads on governance (9 against 5), while Silo leads on fees (6 against 4) and liquidation terms (7 against 6).

Who this rating suits

The score suits borrowers who value a liquidation fee that cannot be raised once they are in a market, and lenders who want to know how a collateral-debt swap would pay them. Readers who need an open-source licence, a named owner with a stated delay, or per-market thresholds published in the docs should weigh the lower marks. Borrowing against volatile collateral can end in liquidation; this rating describes published documentation and is not advice to lend or borrow.