diff --git a/public/index.html b/public/index.html index 53c7aa9..5f4ecd3 100644 --- a/public/index.html +++ b/public/index.html @@ -257,9 +257,11 @@

See it in a minute

by · all tools and live demos at fizzl.eu

diff --git a/public/openapi.json b/public/openapi.json index 2e7bd72..533e7ee 100644 --- a/public/openapi.json +++ b/public/openapi.json @@ -3,6 +3,7 @@ "info": { "title": "PlainText Wallet Approval Checker", "version": "1.0.0", + "termsOfService": "https://fizzl.eu/terms/", "description": "Looks up a wallet's live token/NFT approvals via the GoPlus Security API and explains them in plain language for people with no crypto background.", "contact": { "email": "Fizzl13@protonmail.com" }, "x-guidance": "Use POST /api/check-wallet when you have an EVM wallet address and want to know whether its current token/NFT approvals are safe. Example: {\"address\":\"0x6B0F4651eD42893ab58139938175E4a69f175F25\",\"chain\":\"base\",\"kind\":\"token\"} — chain defaults to \"ethereum\" and kind defaults to \"token\" if omitted, so a call can be as small as {\"address\":\"0x...\"}. Use POST /api/explain instead when you already have a specific approval/permission JSON payload (from a scanner, a dapp prompt, etc.) and just want it translated to plain language, without an on-chain lookup. Both return {verdict: SAFE|CAUTION|RISK, explanation: string} so downstream agents can branch on verdict without parsing prose."