How Does Superfluid Finance Work?
Superfluid Finance is a money streaming protocol that lets you send tokens continuously — every second — rather than in one-time transfers. Instead of sending someone $1,000 at the end of the month, you stream $0.00038 per second so they receive it in real time. This is useful for payroll, DAO grants, subscription payments, and vesting schedules. The protocol has been live since 2021 and is used by major DAOs including ENS and Optimism. It runs on Polygon, Optimism, Arbitrum, Base, and other chains. In February 2022, Superfluid suffered a critical exploit where an attacker stole approximately $13 million by exploiting a flaw in how the protocol tracked who was initiating a transaction — the vulnerability has since been patched and multiple follow-up audits have been conducted. The protocol recently launched its native SUP governance token (February 2025) with a $11M fully diluted valuation, though the token currently trades well below its IDO price. Superfluid does not rely on price oracles for its core streaming functionality, which reduces one common category of DeFi risk. However, its streaming accounts depend on a network of 'sentinel' bots to close insolvent streams — if these bots fail or go offline, users could lose their buffer deposits. The protocol's small TVL relative to larger DeFi protocols means systemic risk is limited, but the historical exploit and ongoing complexity of its novel streaming architecture warrant careful attention.
TVL
$5M
Sector
DeFi
Risk Grade
C
Value Grade
C-
Core Mechanisms
Token Streaming / Continuous Payment
NovelConstant Flow Agreement (CFA) — per-second ERC-20 token streams between two accounts using a virtual balance model updated at each block
Superfluid pioneered programmable real-time token streaming on EVM chains in 2021. The CFA uses a semantic money / token monadic design: actual balances are computed on-demand from a stored flow rate and timestamp rather than stored explicitly, enabling gas-efficient continuous transfers without per-second transactions.
Token Distribution / One-to-Many Streaming
NovelGeneral Distribution Agreement (GDA) — scalable one-to-many pool-based token distribution where members receive proportional shares of inflows in real-time
The GDA contract enables a single streamer to distribute to thousands of recipients simultaneously, with pool units determining allocation. Used by ENS DAO for $5.4M in grants and Optimism for 30M OP token retroPGF distributions. Technically distinct from Merkle-airdrop patterns.
Token Standard Extension / Wrapper
Super Token — ERC-20 + ERC-777 extension with built-in streaming and distribution capabilities, wrapping standard ERC-20s via a 1:1 upgrade/downgrade mechanism
Super Tokens are the unit of value in Superfluid. Users wrap standard ERC-20 tokens (e.g., USDC → USDCx) to gain streaming capabilities. The wrapper is non-custodial and 1:1 redeemable. Deployed across Polygon, Optimism, Arbitrum, Base, and 6+ other EVM chains.
Solvency / Keeper Network
Sentinel Network — decentralized keeper bots that monitor stream solvency and trigger liquidations of insolvent (zero-balance) streaming accounts
Stream senders must lock a buffer deposit when opening a stream. When an account balance hits zero while streaming, any sentinel can call the liquidation function to close the stream and claim the buffer as reward. The TOGA (Transparent Ongoing Auction) mechanism determines the Patrician-in-Charge (PIC) who gets priority liquidation rights.
Governance / DAO
SUP Token Governance — native governance token enabling DAO voting on protocol upgrades, fee structures, and ecosystem incentive distribution
SUP (1B max supply) launched February 2025. Initially non-transferable. Team/investor lockups: 3-year stream with 1-year cliff. 60% community allocation. SUP holders vote on protocol parameters and treasury allocation via Superfluid DAO.
Incentive / Staking
SUP Locker System — stake SUP tokens to earn streaming rewards distributed via GDA pools; the locker system uses a tax mechanism to discourage early exit
The locker system was audited by Sherlock in 2024 and 2025. A medium-severity finding identified that locker owners could deploy SUP liquidity into Uniswap V3 low-liquidity pools to bypass the exit tax mechanism. Team acknowledged but deferred fix. Adds composability risk.
How the Pieces Interact
Context injection vulnerability: the Host contract serializes/deserializes a 'ctx' state object to track msg.sender across agreement calls. In Feb 2022, an attacker crafted malicious calldata that exploited a discrepancy between Host-side serialization and Agreement-side deserialization to spoof account identities, stealing ~$13M. Although patched, the ctx pattern remains a complex trust boundary.
Delayed liquidation risk: if sentinel bots fail to monitor (offline, congestion, low incentive) an insolvent stream, the protocol accrues bad debt. The stream sender's buffer is slashed, but in extreme cases the protocol absorbs losses. Over-reliance on economic incentives for external keepers creates a liveness assumption.
Composability risk amplification: protocols integrating Super Tokens (e.g., Aave, Uniswap wrappers, yield optimizers) inherit Superfluid's full contract risk surface. A vulnerability in the Superfluid Host or CFA agreement contracts could propagate to all integrated protocols simultaneously, creating correlated failure.
Unit saturation in GDA pools: due to technical limitations in the GDA implementation, total pool units cannot exceed the flow rate. In high-membership pools with small individual allocations, unit downscaling can cause rounding errors that result in members receiving less than expected, potentially creating unfair distributions in DAO grant scenarios.
Tax bypass via low-liquidity pool manipulation: locker owners can deploy SUP tokens into a low-liquidity Uniswap V3 pool paired with ETH, skew the pool price so liquidity is ETH-heavy, then withdraw ETH without triggering the SUP exit tax. Acknowledged by team but unfixed as of 2025 Sherlock audit.
What Could Go Wrong
- Historical context-injection exploit (Feb 2022, ~$13M lost) demonstrated critical smart contract vulnerability in Host contract ctx serialization — though patched, the incident reveals inherent complexity risk in the Super Agreement architecture
- Sentinel liquidation network: insolvent streams rely on external keepers to close positions; delayed liquidation can result in protocol bad debt absorbed by stream initiator's buffer deposit
- Super Token wrapper composability risks: any integrated dApp or protocol that wraps ERC-20s into Super Tokens inherits Superfluid's contract risk surface, creating transitive exposure for downstream users
- Emission-heavy SUP token launch with 60% community supply targeted at rewards/incentives creates sustained sell pressure against nascent fee-capture mechanisms
Host Contract Re-entrancy or ctx Bypass Exploit
TailTrigger: A novel attack vector is discovered in the Host contract's agreement dispatch mechanism — such as a new path to inject a forged ctx, bypass isCtxValid() checks, or exploit ERC-777 callback hooks during Super Token operations
- 1.Attacker identifies a new calldata injection or context spoofing path in the Host contract — Ability to impersonate any streaming account and redirect outgoing streams or drain approved Super Token balances
- 2.Attacker drains Super Token balances from targeted accounts before sentinel network detects anomaly — Immediate loss of Super Token assets held by affected wallets; could target high-value DAO treasury streams or yield distribution pools
- 3.Protocol team triggers emergency pause/upgrade of Host contract — All active streams and GDA distributions halt across all integrated chains simultaneously; downstream protocols (payroll tools, vesting contracts) stop functioning
- 4.Confidence in Super Token standard collapses; integrators suspend Super Token support — Mass stream closures as users panic-withdraw wrapped tokens back to base ERC-20; TVL drops to near zero
- 5.SUP token value collapses as protocol utility is suspended — Ecosystem partners (Coinshift, Request Finance) migrate to alternative payment rails; recovery path unclear without substantial re-audit
Risk Profile at a Glance
Overall: C (43/100)
Lower score = safer