Felix CDP

Felix CDP is a Hyperliquid-native Liquity V2 fork implementing a CDP-based stablecoin and lending suite on Hyperliquid L1, centered on the feUSD stablecoin. Protocol identification

• Name: Felix Protocol / Felix CDP.
• Website (UI): usefelix.xyz (referenced in multiple independent guides and docs).
• Docs: Felix GitBook at usefelix.gitbook.io/docs, covering CDP, feUSD, and Vanilla markets.
• Category: Collateralized Debt Position (CDP) / over‑collateralized stablecoin plus lending money market (“Vanilla” markets built on Morpho’s stack).
• Chains: Hyperliquid L1 / HyperEVM only; all sources describe Felix as Hyperliquid‑native and do not mention deployments elsewhere.
• Native token(s):
• feUSD – protocol’s USD‑pegged, over‑collateralized stablecoin minted via CDPs.
• No separate “Felix” governance token is documented in retrieved sources (collateral assets are HYPE, UBTC, kHYPE, wstHYPE, WBTC, ETH, LSTs, etc.).
• Launch date: Not explicitly stated; ecosystem and analysis pieces describe Felix as an established Hyperliquid protocol by mid‑2025. Precise launch date: Not verifiable as of 2026‑09‑03.
• Main functions: Users deposit approved collateral, mint feUSD at ~40% LTV, choose borrower interest rates, and interact with stability pools and Vanilla lending markets. Contract / address information & verification
• GitHub repo felixprotocol/felix-contracts describes Felix CDP contracts (Trove‑style CDP, feUSD token, etc.), confirming contract‑based implementation but does not list canonical deployed addresses.
• Analytics platforms (DefiLlama Felix / Felix CDP pages) report TVL and fees on Hyperliquid L1, implying tracked addresses, but do not expose them directly in retrieved snippets.
• Explorer‑level contract addresses and verification status (e.g., on a Hyperliquid EVM explorer) are Not verifiable as of 2026‑09‑03 with the data available here.
• On‑chain verification via decoded contract code and Dune‑style queries is Not verifiable as of 2026‑09‑03. Fork lineage & changes vs. upstream
• Multiple independent sources (risk analysis, ecosystem guides, IQ.wiki, Delphi) describe Felix explicitly as a Liquity V2 fork.
• Core inherited features: Trove/ CDP architecture, over‑collateralization, redemptions, stability pools, USD‑pegged stablecoin.
• Key modifications vs Liquity V2:
• Runs natively on Hyperliquid L1 rather than mainline L1s.
• User‑chosen borrowing interest rates instead of purely algorithmic rates; liquidation priority tied to chosen rate.
• Expanded collateral set aligned with Hyperliquid (HYPE, UBTC, Hyperliquid LSTs, etc.).
• Integration with Vanilla money markets built on Morpho’s lending stack under the same protocol brand. Audits & fork‑risk history
• Retrieved content does not include audit reports or auditor‑site links for the Felix CDP contracts. Audit status: Not verifiable as of 2026‑09‑03.
• Hindenrank and other risk write‑ups analyze Felix safety but do not link to formal audits.
• No independent record of malicious modifications or exploit history in Felix or in closely related Hyperliquid Liquity‑fork CDPs appears in the retrieved material. Exploit / malicious‑fork history: Not verifiable as of 2026‑09‑03.

Felix CDP appears to be a real product with published developer docs and code, not just a static landing page: the project GitBook describes a CDP market (feUSD) and the GitHub repo states users can lock collateral and issue feUSD. However, app-level maturity is only partially verifiable from the available sources: there is evidence of documented deposit/withdraw mechanics and APIs in the broader Hyperliquid stack, but live Felix-specific deposit/withdrawal functionality, broken-link rate, and whether the public portal is fully operational are not verifiable as of 2026-09-03. The clearest open API signal is indirect: Felix’s GitBook and GitHub show developer-oriented documentation/repo material, but the search results do not confirm a public Felix-specific API endpoint or SDK beyond the protocol’s documentation surface.

The Hyperliquid ecosystem itself exposes documented REST endpoints for deposits, withdrawals, and info calls, which supports that the underlying chain and adjacent tooling are API-capable, but that does not by itself prove Felix has its own open API. Overall assessment: real protocol presence and non-trivial documentation are confirmed; full product maturity, live UX quality, and open Felix-specific API availability remain not verifiable as of 2026-09-03.

Felix CDP appears to have an active bug bounty program on Immunefi. It started on 02 October 2025 and is triaged by Immunefi. The stated max payout is $100,000.

Critical smart-contract bugs are rewarded at 10% of funds directly affected, capped at $100,000, with a minimum critical reward of $20,000. High-severity issues related to theft/permanent freezing of unclaimed yield or royalties are paid $4,000–$10,000, with temporary-freeze rewards increasing by 2x for each additional 24 hours until the high-severity cap. A PoC is required, KYC is required, and payouts are handled by the Felix team in USDC on Ethereum.

No public incident/results page was found in the gathered sources, so results are Not verifiable as of 2026-09-03.

Оценка на 5 сентября 2026 г.

• Основная зависимость: Felix CDP работает только на Hyperliquid L1; DeFiLlama указывает 100% TVL на этой сети. Это концентрирует риск в Hyperliquid L1/HyperEVM: сбой консенсуса, остановка сети, неэффективные ликвидации и ограниченная ликвидность HYPE могут создать bad debt.
• Оракулы и манипуляции: риск высокий. Для HYPE Hyperliquid использует validator-weighted oracle; HYPE может не получать внешние источники до достижения достаточной ликвидности. Felix-аудит отдельно отмечает, что цена HYPE может определяться преимущественно активностью внутри Hyperliquid. Для других активов Felix-код описывает Chainlink, тогда как текущие Felix-документы называют RedStone или эквивалент. > Противоречие: документация/код не дают единой подтверждённой картины oracle stack: Chainlink в коде и RedStone в текущих risk/terms-материалах. Это повышает риск ошибочной оценки stale-price, liveness и fallback-механизмов.
• Внешние протоколы и collateral: подтверждённые CDP-активы включают HYPE/WHYPE и UBTC/feUBTC; публичные сторонние данные также указывают на kHYPE и wstHYPE. LST/wrapped-активы добавляют smart-contract, exchange-rate, depeg и liquidity risk. Точные доли collateral не подтверждены on-chain в этом прогоне.
• Bridges/custodians/CEX/MM: для CDP не подтверждена зависимость от кастодиана, конкретного CEX или внешнего market maker. Bridge-риск остаётся косвенным для любых bridged/wrapped collateral и ликвидности Hyperliquid, но конкретная доля не подтверждена.
• RWA issuer/SPV: Not verifiable as of September 5, 2026. Найденные материалы о HIP-3/TSLA относятся к отдельному продукту/рынку, а не доказывают наличие RWA-зависимости в CDP.
• Stablecoin/depeg scenarios: feUSD — collateral-backed CDP stablecoin. Сценарии потерь: падение HYPE/UBTC/LST, oracle overpricing, задержка ликвидаций, depeg feUSD и недостаточная Stability Pool liquidity. Аудит отмечает риск bad debt при неэффективных ликвидациях или остановке/возобновлении branch. dependency_failure_active: null — Not verifiable as of September 5, 2026. max_exposure_pct: null — Not verifiable as of September 5, 2026; Dune MCP недоступен, поэтому on-chain concentration не рассчитывалась.

Felix CDP is organized as a non-custodial protocol: users retain control of their wallets, private keys, and recovery phrases, and Felix states that the platform/operator does not provide custodial services for tokens. In the CDP flow, users deposit approved collateral into smart-contract vaults/Troves that enforce minting, repayment, and liquidation rules, so the protocol contracts hold and account for collateral on-chain until the position is closed or liquidated. I did not find verifiable evidence that user assets are segregated in the traditional custodial sense; the closest accurate characterization is that positions are isolated per smart contract/vault and managed on-chain, but segregation as a formal custody feature is not verifiable as of 2026-09-05.

I also did not find verifiable evidence that withdrawals are paused as of 2026-09-05.

No confirmed security incident, exploit, or user-loss event was found in the supplied sources for Felix CDP since launch. The only concrete third-party security material is a July 23–25, 2025 Three Sigma audit that reported two medium-severity pricing issues and one informational risk, all of which were remediated in commit #398ca73; this is a code-quality finding, not an incident. The protocol was also later reported to be shutting down its DEX/related markets in June 2026, but that source does not describe a hack or loss event.

For Felix CDP on Hyperliquid L1, key management is not verifiable as of 2026-09-03 from the available sources. The public materials I found describe the protocol’s product mechanics—permissionless collateral deposits, feUSD minting, Troves, Stability Pools, and liquidation/redemption flows—but they do not explain how cryptographic or admin keys are generated, stored, rotated, or controlled. One relevant clue is that Felix has an EOA mode for some products, where users hold assets in their own externally owned wallets rather than an embedded account system, which suggests user-side custody in that context.

However, that does not establish the protocol’s internal key-management model for the CDP system itself. A separate search result for “Felix” was clearly about an unrelated contract management software company, so it is not relevant to this DeFi protocol. Based on the available evidence, the safest institutional answer is: Not verifiable as of 2026-09-03.

No source here confirms whether Felix uses multisig governance, a timelock, HSMs, MPC, privileged admin keys, or a fully immutable deployment for the CDP on Hyperliquid L1.

Assessment (as of September 5, 2026): HIGH admin/upgrade risk. Felix’s official HyperEVM deployment registry now publishes AdminController 0xf42fdd953e68d0010f5fa9d61ef1ba0fc997ef2f, core ProxyAdmin 0x44a600c031dbdf652500f4a6d84b168056f13692, AdminController ProxyAdmin 0xdf1293b46d3d8f6c090ab98094805db68922ce30, CollateralRegistry 0x9de1e57049c475736289cb006212f3e1dce4711b, and feUSD 0x02c6a2fa58cc01a18b8d9e00ea48d65e4df26c70. Four collateral branches are listed: WHYPE, UBTC, KHYPE and WSTHYPE. Architecture: AdminController → ProxyAdmin(s) → Transparent/EIP-1967 proxies → branch contracts (BorrowerOperations, TroveManager, StabilityPool, pools, PriceFeed, TroveNFT) + CollateralRegistry + feUSD. Dedaub confirms Felix uses transparent proxies, with proxy administration centralized under AdminController; upgrades and critical parameters are subject to a documented three-day timelock.

Mutable controls include MCR, CCR, debt caps, PriceFeed, InterestRouter and Stability Pool yield. The report found no critical issues and its single high-severity finding was marked resolved. Privileged actions/risk: AdminController can upgrade implementations, change risk parameters, add collateral/branches, modify oracle/interest-routing configuration, and invoke temporary shutdown. Shutdown blocks new troves and debt/collateral/interest adjustments; closing troves, liquidations, Stability Pool withdrawals and urgent redemptions remain available.

A compromised admin key could therefore deploy malicious logic, manipulate parameters/oracles, freeze borrowing, or potentially drain collateral through an upgraded implementation. Direct current-implementation withdrawal authority is not established. Exit protection: Users appear able to exit without admin during normal operation and documented temporary shutdown, but this is not an immutable guarantee because upgrade authority remains centralized. Timelock existence is documented, but its current on-chain delay, proposer/executor roles, Safe threshold, and role-renunciation status are Not verifiable as of September 5, 2026.

Dune decoded-event verification was unavailable. Contradiction / change from prior review: Deployment addresses are now publicly listed, improving discoverability; however, live proxy-slot verification and deployment-specific audit matching remain Not verifiable as of September 5, 2026. Structured fields: admin_can_drain=true; audited_deployment=null; upgradeable=true; unresolved_critical=0; unresolved_high=0. Bottom line: material rug/freeze risk is concentrated in centralized upgrade and parameter-admin authority, partially mitigated—but not eliminated—by the documented timelock and user exit paths.

Price Feed and Interest Router Audit — Felix-specific assessment listed by the bug-bounty disclosure.

Coinspect Core Protocol Audit I — Felix-specific report listed by Felix’s bug-bounty disclosure.

Coinspect Core Protocol Audit II — Felix-specific report listed by Felix’s bug-bounty disclosure.

Dedaub — Felix Smart Contract Security Assessment (first Felix audit).

Dedaub — Felix Smart Contract Security Assessment (second audit).

Three Sigma — Felix Protocol price-feed audit.

Based on currently available information, Felix CDP’s *founders and team are effectively anonymous to the public*, with no clear, verifiable real‑world identities or corporate footprint disclosed as of 2026‑09‑03. ### Founders & team

• Publicly available profiles (risk write‑ups, ecosystem guides, protocol docs) describe Felix CDP/Felix Protocol mainly as a *Liquity V2 fork on Hyperliquid L1* with CDP and vanilla markets, but do not name any founders, executives, or core devs.
• External analyses (e.g., risk ratings and ecosystem guides) likewise discuss design and usage, not people; no linked LinkedIn / legal names / prior projects are cited.
• There is no independent evidence tying Felix CDP’s operators to prior major DeFi projects, nor to known teams involved in past hacks or failures. Given this, you should treat the team as pseudonymous/anon for risk purposes. Any claimed identities found only in social channels without independent corroboration are Not verifiable as of 2026‑09‑03. ### Jurisdiction, entity, and office reality check
• Neither the docs nor third‑party writeups identify a registered company, jurisdiction, or regulatory status for Felix CDP/Felix Protocol.
• No credible independent source (auditor report, regulator filing, or corporate registry) links Felix CDP to a named legal entity, onshore or offshore.
• There is no evidence of a physical “real office” (address, photos, staff, or local corporate filings). For institutional risk, assume remote, globally distributed, and non‑domiciled from your perspective. Thus, all of the following are Not verifiable as of 2026‑09‑03:
• Exact founding team identities and backgrounds
• Location of incorporation and governing law
• Existence of a staffed physical office ### Credibility & “real business vs web front”
• Felix is widely covered as a major Hyperliquid‑native CDP/lending protocol with substantial TVL and ecosystem integration, and is treated as a key DeFi building block on Hyperliquid by independent commentators and analytics platforms.
• That ecosystem footprint supports that this is a real, functioning protocol, not a pure “web front.” However, operational control, treasury control, and governance power remain opaque without a known entity or named principals. From an institutional risk lens, Felix CDP scores strong on *usage traction* but weak on *transparency and accountable legal/personnel structure* — the latter should be treated as a material governance and recourse risk.

Felix CDP currently has a moderately positive but still emerging reputation, with recognized technical lineage and some audit coverage, but also non‑trivial smart‑contract and market‑structure risks. No public allegations of fraud, rug pull, or insolvency have surfaced as of 2026‑09‑03. Protocol & technical reputation

• Felix CDP is described as a Liquity V2‑style / Liquity V2 fork CDP protocol on Hyperliquid L1, minting the feUSD stablecoin and offering additional lending products (Felix Vanilla). This links it to a well‑known, battle‑tested architecture, which is generally seen as a positive signal.
• Independent risk platforms (e.g., Hindenrank) characterize Felix CDP as “mostly safe” with grades around Risk B / Value D, reflecting perceived solid core design but more limited value accrual or conservative upside. Audits and security review
• Felix relies on the Liquity V2 codebase, which has undergone audits by firms such as Dedaub, ChainSecurity, Recon, Coinspect, Three Sigma, and formal verification by Certora; this is noted in Felix documentation and is therefore an *unverified marketing claim* unless cross‑checked directly with auditors.
• A specific Three Sigma security review of Felix Lending Protocol identified two medium‑severity issues (price feed deviation thresholds and implicit reliance on the USDC/USD peg) plus market‑structure risk around kHYPE withdrawals, with remediation recommendations. This shows at least one independent audit engagement and transparent issue disclosure. Sentiment, usage, and ecosystem standing
• Coverage from DeFi media and analytics (e.g., Defiant, Delphidigital, Hyperliquid ecosystem guides, DefiLlama) highlights Felix as a leading native CDP/lending protocol on Hyperliquid with high TVL and loan volume, indicating strong user adoption and neutral‑to‑positive sentiment.
• Risk‑focused write‑ups (Hindenrank and others) emphasize CDP mechanics, liquidation and oracle dependencies, and interest‑rate choice features—framing risks as technical/market rather than governance or integrity concerns. Founders, investors, and governance
• Publicly accessible profiles provide limited detail on founders and investors; no major venture names or public corporate backers are prominently cited. *Not verifiable as of 2026‑09‑03.*
• Governance structure, legal entity, and jurisdiction are not clearly documented in independent sources. *Not verifiable as of 2026‑09‑03.* Fraud, rug, regulatory or sanctions issues
• No credible reports of rug pulls, fraud, insolvency, sanctions, or enforcement actions specifically targeting Felix CDP were identified in independent media, analytics, or ecosystem resources. *Not verifiable as of 2026‑09‑03* for exhaustive checks, but nothing surfaced in mainstream or specialist coverage. Unresolved concerns for institutional users
• Reliance on Hyperliquid L1 infrastructure and price feeds (including dependencies on USDC and LST markets), plus relatively concentrated chain exposure, are recurring risk themes.
• Limited transparency on legal structure, compliance posture, and key personnel is a reputational gap by institutional standards.
• Some security assurances derive from Liquity V2’s audit history rather than Felix‑specific full‑scope audits, which should be treated with caution absent direct auditor confirmations.

Economic model (Felix CDP; Hyperliquid L1): Liquity-style, overcollateralized CDP. Users deposit HYPE/WHYPE and UBTC, mint feUSD, and remain economically long the collateral; this is directional, not market-neutral. feUSD can be swapped into other assets, creating user-driven leverage/looping, but no embedded leverage, restaking, or external yield strategy is documented. Yield source: Stability Pool depositors receive borrower interest and collateral from liquidations; interest is protocol-minted feUSD, with documented routing primarily to the Stability Pool and potentially DEX-LP incentives. Liquidation gains are compensation for absorbing bad debt, not risk-free yield.

No independent subsidy or emissions program was identified; the organic share cannot be quantified. Assets, collateral and mechanics: Collateral is held in Felix contracts and priced through oracle feeds. Borrowers face MCR/CCR requirements, borrower/upfront interest, dynamic redemption fees, minimum-debt constraints, liquidation, and possible collateral-branch shutdown. feUSD redemptions exchange feUSD for oracle-valued collateral, less a dynamic fee. No fixed maturity or lock-up is documented; withdrawals/repayments are available through Trove operations, but must preserve solvency constraints and may be affected by shutdown conditions. Fees and revenue: Documented fee sources are borrower interest, upfront borrowing fees, redemption fees, and liquidation/gas-compensation mechanics.

DeFiLlama currently reports approximately $35.09m TVL, all on Hyperliquid L1, $85.9k fees over 30 days, and $0 protocol revenue; its methodology indicates fees largely flow to Stability Pools/borrowers rather than Felix treasury. TVL/product/trend: DeFiLlama’s CDP page tracks only the CDP product and one chain: Hyperliquid L1 = 100% of reported TVL. A separate ranking page shows $36.81m, versus $35.09m on the protocol page, indicating an aggregator snapshot discrepancy. Dune TVL, product decomposition, and trend: Not verifiable as of September 5, 2026. APY: DeFiLlama shows an average tracked APY of 3.95%; historical APY, volatility, and sustainability are Not verifiable as of September 5, 2026.

Returns are inherently variable with borrower utilization, interest parameters, liquidations, collateral prices, and feUSD demand.

Assessment — as of September 5, 2026

• Liquid reserves / treasury size: Not verifiable as of 2026-09-05. Dune MCP was unavailable, so no on-chain balance, valuation, latest block, or execution-ID verification can be provided. DefiLlama reports $35.59 million TVL, 100% on Hyperliquid L1, but TVL is not a disclosed treasury or liquid-reserve balance.
• Liabilities: Not verifiable as of 2026-09-05. Felix documentation identifies protocol read functions for total collateral, total feUSD debt, and Stability Pool deposits, but no current values were available from a verifiable on-chain query.
• Known protocol addresses: Felix publishes branch contracts for the WHYPE and UBTC markets, including Active Pool, Default Pool, Stability Pool, Trove Manager, feUSD, and collateral-registry addresses. These are CDP accounting/collateral contracts—not evidence of a separate corporate treasury.
• Composition: The documented CDP collateral branches include WHYPE and UBTC, with feUSD debt and Stability Pool mechanics. This describes user collateral backing, not independently held reserve assets.
• Custody and control: Felix’s terms state that the CDP uses immutable smart contracts governed by a Platform Operator-controlled multisig; the operator manages risk parameters but is described as unable to seize or redirect user funds. This is a protocol-control disclosure, not a reserve attestation.
• Reserve policy / attestations: Not verifiable as of 2026-09-05. The audit materials concern smart-contract security and modified Liquity code, not proof-of-reserves, treasury statements, custody attestations, or liabilities attestations. Contradiction / data-quality callout: A Hyperion DeFi release dated October 29, 2025 claimed Felix CDP had over $250 million deposited collateral and tens of millions of feUSD debt, while the latest retrieved DefiLlama page reports $35.59 million TVL. These figures use different dates and may use different scope/methodologies; neither establishes treasury reserves. The on-chain figure is unavailable in this run. A separate Hyperion–Felix arrangement involved 500,000 HYPE in a jointly controlled multisignature wallet for HIP-3 market support; this is not demonstrated to be Felix CDP reserve capital.

Felix CDP currently appears to have no live, tradable native token on Hyperliquid L1, and its tokenomics are Not verifiable as of 2026-09-03 under the constraints of this run. Key points:

• Native token / contract address Web search for “Felix CDP token”, “FLX Felix CDP Hyperliquid”, “felix-cdp hyperliquid token address” returns no credible references to a deployed ERC‑20/chain‑native token, no contract address, and no listings on major aggregators (CoinGecko, DefiLlama, Token Terminal) or Hyperliquid ecosystem overviews. In the absence of on-chain tools and any independent confirmation, the existence of a Felix CDP token is Not verifiable as of 2026-09-03.
• Supply, market cap, FDV Because no token contract or listing can be confirmed, total supply, circulating supply, market capitalization, and fully diluted valuation cannot be established. Not verifiable as of 2026-09-03.
• Token utility, governance, revenue share No independent documentation (audits, Hyperliquid ecosystem docs, GitHub, governance records) describing a Felix CDP governance or utility token, staking, revenue share, buybacks, or burns can be found. Any such roles are Not verifiable as of 2026-09-03 and any claims from protocol-controlled channels would be “unverified marketing claims”.
• Emissions and unlock schedules; allocations There is no verifiable token, so emission schedule, cliff/vesting unlocks, team/investor/treasury/community allocations, and whether announced unlocks occurred on-chain are all Not verifiable as of 2026-09-03.
• Top-holder concentration / insider wallets Without a confirmed token contract, holder distributions and insider concentration cannot be analyzed. Not verifiable as of 2026-09-03.
• Mint / blacklist / fee-switch controls No verified Felix CDP token contract means no assessment is possible of privileged functions (minting, pausing, blacklisting, fee-switch, parameter changes) or who controls them (EOA vs multisig vs DAO). Not verifiable as of 2026-09-03.
• Listings, liquidity depth No independent evidence of a Felix CDP token pair on major DEXs or orderbooks (Hyperliquid spot, other exchanges) shows up in search. DEX liquidity depth and main listings are Not verifiable as of 2026-09-03. Given current evidence, the working risk-analyst stance should be: treat Felix CDP as a protocol with no confirmed native token, and disregard any token-related metrics unless and until a specific contract address and independent listings can be verified.

For Felix CDP on Hyperliquid L1, a Bitcoin drop below $10,000 is not verifiable as of 2026-09-03 from the available web sources, because I do not have on-chain checks and the retrieved material does not provide Felix’s exact collateral composition, liquidation parameters, or insolvency waterfall. The only Felix-specific source indicates the protocol is exposed to HYPE collateral reflexivity—i.e., a sharp HYPE decline can trigger mass liquidations that may further दब pressure on HYPE. What can be said with confidence is that a BTC crash to sub-$10k would likely be a system-wide risk-off shock, not a protocol-specific parameter trigger unless Felix has direct BTC exposure.

In that case, the main transmission channels would be: reduced collateral values, higher liquidation frequency, wider oracle/market spreads, and potential redemption stress if users rush to exit positions. Felix’s design also appears to rely on redemption logic that can create a floor for feUSD under normal conditions, but that does not by itself demonstrate resilience under an extreme BTC tail event. The key unanswered questions are: whether BTC is accepted as collateral on Felix, what the liquidation LTVs and oracle thresholds are, and whether the protocol has backstop liquidity or an insurance module.

Those details are Not verifiable as of 2026-09-03 from the retrieved sources. If you want a risk view in one line: BTC < $10,000 is a severe macro stress scenario, but the material Felix failure mode cannot be confirmed without protocol-specific collateral and liquidation data.

A 20% depeg of the largest collateral asset is a severe stress event for Felix CDP, but the exact loss, liquidation volume, and bad-debt outcome are Not verifiable as of 2026-09-03 from the available sources because the current collateral mix, the largest collateral by value, and the system’s live per-asset exposure are not on-chain verified here. Felix’s CDP liquidates positions when collateral health falls below asset-specific thresholds, and a stability pool first absorbs liquidated debt before any residual is redistributed under the protocol’s liquidation logic. What can be said from the sources is that Felix’s CDP supports multiple collaterals, including HYPE and other assets, and liquidation is triggered by falling collateral value relative to debt.

Independent risk analysis flags a reflexive HYPE-driven cascade as a major failure mode, where collateral selling can amplify price declines and force more liquidations. That means a 20% depeg in the largest collateral could meaningfully raise liquidation pressure, but whether it causes a contained liquidation wave or a systemic shortfall depends on that collateral’s share of total CDP backing and the stability pool’s size, neither of which is verifiable here. The only TVL figure in the results is an aggregator estimate for Felix CDP of about $201.84m.

A separate social post claims $283m CDP TVL, but that conflicts with the aggregator figure and is not independently verifiable from the available evidence.

For Felix CDP on Hyperliquid L1, a “top counterparty insolvent” stress is not verifiable as of 2026-09-03 from the available web results because I could not confirm the protocol’s liquidation engine, keeper set, debt auction path, or any formal loss waterfall from independent sources. What *is* supported is that Felix has at least one identified severe failure mode tied to HYPE reflexivity: a sharp HYPE drop can trigger mass liquidations, and those liquidations can further depress HYPE via forced selling. In that path, the loss first lands on undercollateralized vaults/CDP positions, then on liquidation counterparties/keepers only to the extent they cannot execute or absorb slippage, and finally on any remaining protocol backstops—*but the exact smart-contract compensation path is not verifiable as of 2026-09-03*.

If the “top counterparty” means the largest liquidation counterparty becoming insolvent, the expected loss path is: failed liquidation execution → bad debt left in the CDP system → collateral sale at worse prices → additional pressure on HYPE price → more liquidations. The absorption layer would normally be the protocol’s liquidation mechanism and any reserve/fund if one exists, but Not verifiable as of 2026-09-03. The user compensation path for affected users would depend on whether Felix socializes losses, re-prices debt, or pauses markets; none of those behaviors are independently confirmed here.

No on-chain contract-level waterfall, insurance module, or bad-debt settlement procedure was independently verifiable from the provided results, so the exact impact path through the smart contracts is Not verifiable as of 2026-09-03.

For a fraud by the DAO or owners stress scenario, there is no direct evidence in the supplied sources that Felix CDP’s DAO or owners have committed fraud. The available material points instead to ordinary protocol risk, oracle-design issues, and liquidation/redemption mechanics, not governance fraud or misappropriation. What *is* verifiable is that Felix has had security-review findings around oracle pricing: Three Sigma reported incorrect deviation thresholds in the KHYPE and WSTHYPE feeds and an implicit USDC/USD peg assumption, and noted the fixes were applied in the price path.

Those are design-risk findings, not fraud allegations. The closest relevant stress interpretation is therefore trusted-party / governance failure, not proven fraud: if owners or governance were to act maliciously, the likely damage path would be parameter changes, oracle manipulation, or contract/admin abuse causing bad debt, forced liquidations, or feUSD depeg. However, a specific fraud event by DAO or owners is Not verifiable as of 2026-09-03 based on the provided sources.

Independent risk sources classify Felix as having meaningful market and concentration risk, especially reliance on HYPE collateral and rapid growth, but they do not document owner fraud. The Immunefi program indicates the protocol maintains an active bug bounty, which is a control signal but not evidence about misconduct.

For Felix CDP on Hyperliquid L1, a 30d negative primary yield source stress should be treated as a borrowing-demand / collateral-quality shock rather than a generic “yield” issue. Felix’s CDP design is over-collateralized, and borrower positions face liquidation risk when collateral value falls below thresholds; stability-pool depositors absorb liquidations in feUSD and receive collateral instead. In this protocol, the most plausible negative-yield transmission is: if the main source of borrower demand to mint feUSD weakens or becomes negative over 30 days, users have less incentive to keep debt open, fee income can compress, and the system may see lower utilization and more rapid deleveraging.

That can matter more if the collateral base is concentrated in HYPE, because Felix is described as heavily dependent on HYPE as collateral, and a HYPE selloff can trigger liquidations that further pressure HYPE in a reflexive loop. A useful stress framing is:

• Primary effect: lower borrowing demand and weaker protocol revenue from CDP activity.
• Secondary effect: higher sensitivity to forced closures/liquidations if borrowers migrate to thinner safety margins or if collateral prices weaken at the same time.
• Tail risk: a collateral reflexivity spiral if HYPE drops sharply and liquidations amplify the move. What I can verify from the provided sources is limited: Felix’s documentation confirms liquidation/redemption mechanics and over-collateralization, but there is no verified on-chain evidence here for the actual 30d yield series, current utilization, or chain-specific exposure. Therefore, the actual magnitude of a 30d negative-yield stress is Not verifiable as of 2026-09-03. For an institutional risk memo, the key conclusion is that negative 30d primary yield is a medium severity stress for feUSD economics, but not necessarily a solvency stress unless paired with collateral-price weakness or a peg event.

Assessment — as of September 13, 2026. Felix CDP appears centrally controlled, not governed by a functioning DAO. Felix’s repository identifies AdminController as owner of system contracts and proxy administration. It can control critical parameters, route interest, change collateral/risk settings, shut down/resume branches, and upgrade transparent-proxy implementations.

This creates a credible admin-controlled path to impair or redirect user funds; therefore admin_can_drain = true under the requested definition. Governance/proposals. The documented process is an AdminController proposal/apply mechanism with a delay, but no verified token-holder Governor or self-executing DAO was found. Felix’s Terms explicitly state that feUSD confers no governance rights and that market creation/parameter changes are controlled by the Platform Operator with no DAO governance. The DAO is therefore symbolic/absent.

Exact timelock duration is Not verifiable as of September 13, 2026; a separate September 2026 governance document claiming a 3-day timelock was not treated as reliable evidence of deployed control. Multisig/admin. Previously recorded evidence indicates a 3-of-6 multisig. Signer identities, signer independence, current membership, and whether all upgrade/admin paths use that multisig are Not verifiable as of September 13, 2026. Dune verification of voting concentration and top holders was unavailable; Not verifiable as of September 13, 2026. Company/front end. The Terms name Felix Labs Inc, stated to be a Delaware company, with a Newark, Delaware address.

Registration number and directors are Not verifiable as of September 13, 2026. The front end is described as an interface rather than the executor of transactions, but the operator controls deployment, maintenance, and protocol parameters. Contradiction: published materials describe timelocked administrative governance, while the Terms simultaneously state there is no DAO governance; the latter is consistent with AdminController/company control.

As of September 4, 2026, Felix CDP’s stated contractual entity is Felix Labs Inc., described as a Delaware corporation with a business address at 131 Continental Drive, Newark, Delaware, U.S. The Terms select Cayman Islands law, LCIA arbitration, and Cayman courts for non-arbitrable claims. Independent verification of incorporation, ownership, beneficial owners, licensing, or regulatory registrations was not found: Not verifiable as of September 4, 2026.

ToS/restrictions: The Terms prohibit U.S. users, sanctioned persons/jurisdictions, FATF high-risk jurisdictions, and locations where access is illegal; VPN/proxy circumvention is prohibited. View-only access may apply. The protocol disclaims financial, investment, legal, tax, and securities advice and limits operator liability to US$100.

KYC/AML/sanctions: The platform generally states that it does not conduct customer due diligence, except where required for products such as the Perp DEX or fiat-backed tokens. Users bear AML/CTF and sanctions compliance responsibility; limited screening/geoblocking may apply. This is not evidence of a comprehensive AML program or regulated status.

Classification/legal risk: Felix presents feUSD as an overcollateralized stablecoin and the product as a non-custodial CDP/lending protocol. Legal classification is jurisdiction-dependent and unresolved; lending, stablecoin issuance, yield products, and any derivatives/synthetic products could trigger money-transmission, securities, commodities, consumer-protection, or AML regimes. The Terms’ labels do not determine legal classification. Data protection: The privacy policy describes collection of wallet addresses, transaction history, usage data, and potentially personal data; disclosures to service providers, regulators, courts, and law enforcement are contemplated.

On-chain data is permanent. The policy does not establish a clearly identified GDPR/CCPA controller, supervisory authority, retention schedule for all data, or user-rights process: Not verifiable as of September 4, 2026. Contradiction / actual-risk finding: “Decentralized/non-custodial” positioning coexists with operator-controlled multisig governance, operator-managed risk parameters, and 25% of CDP fees routed to the operator.

This creates identifiable control, governance, and regulatory-perimeter risk despite the contractual decentralization language. No protocol-specific regulator warning, enforcement action, court case, or sanctions designation was identified in the reviewed sources: Not verifiable as of September 4, 2026.

GLEIF LEI registry unavailable at scan time. OFAC SDN screening of 'Felix Labs Inc', 'Felix CDP': no match. SEC litigation and administrative release feeds: no mention.

Felix CDP issues its own native stablecoin, feUSD, and the available web evidence supports that feUSD is intended to remain dollar-pegged via redemptions and overcollateralized CDPs. A specific historical depeg was verifiable: feUSD traded above peg on 2026-08-10, peaking at about +5.02% versus $1.00; no independent web source in this pass verified any below-peg depeg, so the number of depeg events below $1 is not verifiable as of 2026-09-05.

Felix CDP’s principal risks are concentrated in privileged administration, oracle integrity, collateral liquidity, shutdown mechanics, and dependency on the relatively novel Hyperliquid environment. Dune/on-chain exposure, TVL concentration, and current collateral shares are Not verifiable as of September 5, 2026 because the Dune MCP is unavailable; no on-chain figures are inferred.

Felix CDP’s top strengths are: native Hyperliquid integration, which makes it a purpose-built borrowing layer for Hyperliquid L1 and HyperEVM users; Liquity V2-based architecture, which gives it a proven CDP design rather than a novel, untested model; user-set borrowing rates, a differentiating mechanism that lets borrowers choose their own interest rate instead of accepting a purely market-determined rate; strong capital efficiency and utility, because users can mint feUSD against collateral while retaining upside exposure and can route collateral/deposits into the broader Hyperliquid ecosystem; and audited, risk-controlled operations, with multiple sources citing audits and added controls such as mint caps, admin parameter controls, and emergency pause functionality.