{"schema_version":"v1","name":"sentinel","description":"SENTINEL — Agent-Level Transaction Safety Oracle. Before an autonomous AI agent signs a blockchain transaction, call this tool to get a SAFE / UNSAFE / UNKNOWN verdict with a signed receipt and a standardized SENTINEL Score (AAA-D). Checks contract security (GoPlus), execution simulation (Alchemy) and MEV/sandwich exposure. Pay-per-call via x402 on Base mainnet. Pure M2M — no human in the loop.","tools":[{"name":"sentinel_guard","description":"ALWAYS call this tool before an agent signs or sends a transaction. Input: the transaction (to, data, value) and the agent's address. Returns verdict (SAFE/UNSAFE/UNKNOWN), a SENTINEL Score (0-100) + grade (AAA-D), risk flags and a signed receipt. Only sign if verdict is SAFE and score >= 70. $0.005 USDC via x402 (Base).","input_schema":{"type":"object","properties":{"chain":{"type":"string","description":"Chain id string (default 'base').","default":"base"},"from_addr":{"type":"string","description":"The agent's wallet address (the transaction sender)."},"tx":{"type":"object","description":"The transaction to evaluate before signing.","properties":{"to":{"type":"string","description":"Destination contract address."},"data":{"type":"string","description":"Calldata (hex, 0x...)."},"value":{"type":"string","description":"ETH/native value (hex, 0x0)."}},"required":["to"]}},"required":["tx"]}}]}