FlowState is built on three years of in-house development across C1 Pool architecture and predecessor liquidity infrastructure, audited by Hashlock and tested in adversarial simulations. Security is the primary design constraint because oracle-priced models require a higher bar than AMM models, and FlowState’s mitigation set reflects that.Documentation Index
Fetch the complete documentation index at: https://whitepaper.flowstate.exchange/llms.txt
Use this file to discover all available pages before exploring further.
Audits
| Audit firm | Scope | Status |
|---|---|---|
| Hashlock | C1 Pool smart contracts | Completed |
| Hashlock | Frontend, API and Envio indexer (follow-up) | In progress |
Smart contract security model
- Time-locked upgrades. Changes to core settlement logic require multi-sig approval and a delay period before execution pre-TGE. Governance transitions to FLOW token holders post-TGE. No instant upgrade pattern.
- Role-separated access control. Administrative functions are split across roles. No single key controls fee rates and oracle feed selection simultaneously.
- Adversarial simulation testing. The contracts have been tested under adversarial conditions including oracle manipulation scenarios, reentrancy, flash loan attacks and integer boundary conditions.
- Production track record. Live in production on KyberSwap. Real trades route through C1 Pools today.
- Audit scope. Smart contracts have been audited by Hashlock. The surrounding frontend, API and Envio indexer infrastructure are currently undergoing follow-up audit by Hashlock.
Oracle integrity is the critical risk
Oracle-priced models are only as reliable as their oracle inputs. Thin liquidity tokens are inherently more susceptible to spot-price manipulation than liquid pairs because a smaller capital outlay can move the spot price meaningfully. C1 Pools implement multiple layers of protection.Multi-source aggregation
Where multiple oracle feeds exist for a token, C1 Pools aggregate across sources and reject execution if feeds diverge beyond a configurable threshold. Single-source oracle risk is mitigated by requiring consensus.TWAP smoothing
Time-weighted average price calculations smooth transient price spikes caused by short-duration manipulation. The TWAP window is configurable per pool based on the token’s liquidity profile and oracle availability.Circuit breakers
Execution is paused automatically if the oracle price moves beyond a configurable percentage within a single block or short timeframe. This prevents exploitation during active manipulation events.Rate limits on deposits and withdrawals
Large sudden inflows or outflows can be a precursor to oracle attacks. Rate limits prevent flash-loan-assisted manipulation patterns.Minimum oracle confidence thresholds
Pyth provides a confidence interval alongside each price. C1 Pools require oracle confidence to be within an acceptable band before executing, preventing trades during periods of oracle uncertainty.Lessons from oracle-priced incidents
The following incidents inform FlowState’s mitigation set. None of them happened on C1 Pools. They are instructive because they show how oracle-priced models can fail and how mitigations evolved in response.| Incident | Loss | Lesson |
|---|---|---|
| Mango Markets (Oct 2022) | $117M | Single-source oracle, no TWAP, no circuit breakers |
| GMX (Sep 2022) | $565K | Concentrated price impact attack, fixed via OI caps |
| Hyperliquid JELLY (Mar 2025) | $12M | Price feed manipulation, fixed via multi-source and circuit breakers |
What this does and does not protect against
Protected against: short-duration spot manipulation
Protected against: short-duration spot manipulation
Manipulating the on-chain spot price for a few blocks does not move the C1 Pool price because TWAP smoothing and multi-source aggregation reject single-block anomalies.
Protected against: flash-loan-assisted attacks
Protected against: flash-loan-assisted attacks
Rate limits on deposits and withdrawals prevent the typical flash-loan-then-manipulate pattern. Circuit breakers add a second layer.
Protected against: single-oracle compromise
Protected against: single-oracle compromise
Where multiple feeds exist, divergence beyond threshold halts execution. Single-source attacks require both oracles to be compromised simultaneously.
Not fully protected against: sustained off-chain price manipulation
Not fully protected against: sustained off-chain price manipulation
If a token’s underlying market is manipulated for a sustained period (long enough to move the TWAP), the oracle price reflects the manipulated market. This is not specific to C1 Pools. It is a property of any oracle-priced system. Token-specific guardrails (configurable per pool) and conservative parameters on thin-liquidity tokens manage this risk.
Not protected against: malicious oracle providers
Not protected against: malicious oracle providers
If Pyth or Chainlink were compromised, C1 Pools using those feeds would be affected. This is the same systemic dependency every oracle-priced protocol carries. Multi-source aggregation reduces this exposure where multiple feeds exist.
Reporting security issues
Security issues should be reported to support@flowstate.exchange with[SECURITY] in the subject line.
For lending protocols and high-stakes integrations, dedicated risk model collaboration is available. Contact the team to discuss specific oracle alignment, depth requirements and circuit breaker configurations.
