How to use the OnlyFence CLI (`fence`) to execute DeFi actions on-chain with safety guardrails. Use this skill whenever the user asks to swap tokens, lend/borrow/supply/withdraw tokens, check wallet balances, query token prices, query activity history, analyze trading patterns, manage wallets, configure trading limits or token allowlists, trade perpetual futures, open/close perp positions, place limit/market orders, check funding rates, manage margin deposits, start/stop the daemon, or perform any on-chain DeFi action through OnlyFence. Also use when the user mentions "fence", "OnlyFence", trading guardrails, DeFi safety rules, activity analysis, lending, AlphaLend, Jupiter, Aave, Paraswap, Hyperliquid, Bluefin, perpetual, perp, leverage, margin, funding rate, or wants an AI agent to interact with Sui, Solana, or Ethereum from a single tool. If the user asks to "trade", "swap", "lend", "borrow", "supply", "repay", "long", "short", "check balance", "check price", "open position", "close position", "place order
Red team security testing for OnlyFence. Executes real attack vectors against the running system to validate security hardening — process list credential leaks, IPC password brute-force, NODE_OPTIONS code injection, file permission bypasses, unauthenticated trade execution, symlink attacks, and Docker exposure. Use this skill whenever the user asks to 'red team', 'attack', 'pen test', 'security test', 'try to hack', 'test security', 'validate hardening', or 'find vulnerabilities' in OnlyFence. Also trigger when the user asks to 'retry attack', 'attack again', or 'test the fix'.