50 lines (39 loc) · 1.71 KB

Features

Flashix combines sealed inference, private mempool ingestion, and atomic execution into a single audit-friendly arbitrage stack.

Core Features

TEE-sealed decision making through 0G Compute and the compute verification chain.
Private mempool ingestion with live or simulation modes.
Opportunity ranking using spread, cost, freshness, and risk checks.
Mandatory approval gates before execution.
Flashloan-based atomic execution with no partial settlement state.
On-chain replay protection and proof emission.
Settlement exports and post-trade audit trails.
Human override and circuit breaker controls.
Replayable demo state for judges and operators.

Operator Experience

Dashboard exposes health, actions, and demo controls.
Opportunities shows the queue and its decision trail.
Risk shows breaker state, exposure, and overrides.
Execution shows simulation, gas, broadcast, and proof links.
Settlement shows realized PnL and outstanding obligations.
Compute shows the sealed proof chain and signature validation.

Safety Features

Simulation-before-broadcast.
Stale-signal rejection.
Replay protection on signal IDs.
Router allowlisting.
Profit threshold checks.
Rate, gas, and freshness thresholds.
Explicit user approval when the workflow requires it.

Evidence And Auditability

Append-only logs for decisions and post-trade outcomes.
Exportable ledger records.
On-chain events for execution proof.
Persisted demo artifacts for replay and verification.

Where The Code Lives

Frontend: frontend/src/.
Agent: agent/ and agent/prompts/.
Compute: compute/.
Mempool pipeline: mempool-listener/.
Contracts: contracts/.
Supporting docs: docs/.