Fuel Credit development plan
Goal
Build an open-source Fuel-native credit facility where FUEL holders lock FUEL and borrow USDC on fixed terms without mark-to-market liquidation.
Loan product
| Term | Advance rate | Interest |
|---|---|---|
| 6 months | 20% | 10% APY on borrowed USDC |
| 9 months | 22.5% | 10% APY on borrowed USDC |
| 12 months | 25% | 10% APY on borrowed USDC |
Rules:
- User deposits FUEL collateral and receives USDC immediately if pool liquidity exists.
- FUEL stays locked until maturity.
- There are no price-triggered liquidations.
- At maturity, user has 2 weeks to repay in full or partially.
- Partial repayment returns a pro-rata share of locked FUEL.
- Interest can be paid in USDC or FUEL.
- At maturity, unpaid positions become delinquent and are handled by keeper/governance policy.
Phases
Phase 0 — planning and skeleton
- Keep repo structure minimal:
contracts/,app/,workers/stats/,scripts/,docs/. - Document addresses, architecture, UX, risks, and tickets.
- Confirm Fuel asset IDs and Sway standards to use.
Phase 1 — local/testnet MVP
- Single Sway lending contract.
- Fake signed FUEL/USD oracle Worker for local/testnet only.
- React app with Fuel wallet connector.
- Deposit, borrow, view loans, repay, and claim collateral.
- Stats Worker indexes contract events from deployment block.
- Local/testnet deploy scripts write artifacts into
deployments/. - Testnet faucet helper links users to the Fuel faucet after wallet connect.
Phase 2 — production hardening
- Replace fake oracle with Stork Fuel oracle integration.
- Add contract audit checklist and invariant tests.
- Add keeper runbook for expired/delinquent loans.
- Add Cloudflare Pages production deploy on
fuel.credit. - Add mainnet guarded deploy flow with balance warnings and address confirmations.
Phase 3 — bridge and growth UX
- Add LayerSwap bridge option for Ethereum-held FUEL.
- Detect EVM FUEL balance where possible.
- Add richer historical stats export and confidence metrics.
- Add docs site polish and public launch material.
Deliberate MVP cuts
- No automated LayerSwap bridge in Phase 1.
- No Stork production oracle in Phase 1.
- No multi-contract architecture unless tests prove the single-contract path is unsafe.
- No complex admin UI; keeper actions can be CLI scripts first.