Cross-Chain Messaging
Conditional: bridge trust model, sequencer risk, fraud proof maturity, cross-chain replay risk.
How This Score Is Built
Conditional: bridge trust model, sequencer risk, fraud proof maturity, cross-chain replay risk.
Scoring Tree
Sub-Score Breakdown
Score Composition
L1Write is FIRE-AND-FORGET: emits events, never reverts, returns nothing — the EVM contract cannot verify the L1 op actually executed (no read-after-write reconciliation). This is the core cross-chain weakness.
Scored 42 (weak): a real, exploitable, no-acknowledgement bridge surface with multiple passing-PoC desync findings. Activating D10 LOWERS the composite vs leaving it out — the honest direction.
ACTIVATED in this pass. Kinetiq has a genuine EVM<->HyperCore (Hyperliquid L1) read/write bridge: L1Write precompile at 0x3333...3333 (delegate/undelegate/spot-transfer writes) and L1Read precompiles (0x2222/0x1111 endpoints) for reading L1 staking/spot state. This is a real cross-chain-style messaging surface, weight ~0.07.
Burn-before-L1-ack (#2248, passing PoC): confirmWithdrawal burns kHYPE and pays from a shared address(this).balance pool with no per-request L1 settlement proof — cross-user drain / undercollateralization risk during L1 desync
L1Write silent-drop (#37, passing PoC): HyperCore's per-address pending-op queue limit silently drops ops beyond the limit while the EVM updates accounting as if successful — EVM<->L1 state desync
Trusted L1 precompile endpoints (#2249); no independent verification layer between EVM accounting and L1 truth
Evidence Chain (1 files)
Score History
No dimension-level score changes recorded yet.