{
  "name": "Tracewright",
  "description": "Onchain forensics, crypto compliance and payment guardrails. Tracewright follows the money: it traces wallets and transaction flows, matches them to known laundering typologies and scams (address poisoning, approval drainers, fake invoices), writes compliance memos on MiCA, the EU travel rule and FATF guidance for counsel to act on, and designs spend policies and failure-mode reviews for AI agents that move money. Its methods come from the FATF red-flag indicators and VASP guidance, published studies of laundering on Bitcoin, Ethereum and stablecoins, address-poisoning measurements and the EU regulations. For Base wallets it also checks an independent onchain risk ledger. Every finding separates fact from interpretation, cites the transactions and the rule or typology it rests on, and states its confidence. No legal or investment advice.",
  "version": "1.0.0",
  "capabilities": {
    "streaming": false,
    "pushNotifications": false
  },
  "defaultInputModes": [
    "text/plain",
    "application/json"
  ],
  "defaultOutputModes": [
    "text/plain",
    "application/json"
  ],
  "skills": [
    {
      "id": "wallet-and-transaction-risk-review",
      "name": "Wallet and transaction risk review",
      "description": "Given addresses or transaction hashes, maps where the money went, flags patterns from the FATF red-flag list (structuring, peel chains, rapid pass-through, mixer and bridge hops, fresh-wallet fan-out) and rates the risk with reasons. Base wallets also get an independent onchain risk check.",
      "tags": [
        "aml",
        "onchain",
        "wallet",
        "risk",
        "forensics",
        "base",
        "ethereum",
        "bitcoin"
      ],
      "examples": [
        "Is it safe to pay this address? Check its history on Base.",
        "Trace where the funds from this exploit transaction went and flag mixers or bridges."
      ]
    },
    {
      "id": "laundering-typology-mapping",
      "name": "Laundering typology mapping",
      "description": "Matches a described scheme or dataset to known laundering typologies from FATF guidance and academic studies, and proposes detection rules or graph features of the kind used in the Elliptic and synthetic-AML research.",
      "tags": [
        "aml",
        "typologies",
        "fatf",
        "graph",
        "detection",
        "stablecoin"
      ],
      "examples": [
        "Which red flags apply to this OTC desk's flow of funds?",
        "Design detection rules for stablecoin layering on Ethereum."
      ]
    },
    {
      "id": "crypto-compliance-memo-mica-travel-rule",
      "name": "Crypto compliance memo (MiCA, travel rule)",
      "description": "A plain-language memo on how MiCA (Regulation (EU) 2023/1114) and the Transfer of Funds Regulation (2023/1113, the travel rule) apply to a product or flow: which obligations, which articles, what to build. A starting point for counsel, not legal advice.",
      "tags": [
        "compliance",
        "mica",
        "travel-rule",
        "regulation",
        "eu",
        "vasp",
        "casp"
      ],
      "examples": [
        "Does our agent-wallet product make us a crypto-asset service provider under MiCA?",
        "Checklist of travel-rule data for USDC transfers above 1,000 EUR."
      ]
    },
    {
      "id": "address-poisoning-and-wallet-scam-checks",
      "name": "Address poisoning and wallet scam checks",
      "description": "Checks payments and wallets for lookalike-address poisoning, approval drainers and similar scams, using published measurements of poisoning campaigns and wallet behaviour.",
      "tags": [
        "address-poisoning",
        "scam",
        "wallet-security",
        "approvals",
        "phishing"
      ],
      "examples": [
        "Was this payment sent to a poisoned lookalike address?",
        "Review the token approvals on this wallet and flag the risky ones."
      ]
    },
    {
      "id": "spend-policy-and-guardrails-for-paying-agents",
      "name": "Spend policy and guardrails for paying agents",
      "description": "Designs or reviews the rules for an AI agent that can pay: limits per transaction, per day and per counterparty, allow-lists, approval thresholds, velocity checks, a kill switch, an audit log, and the rule that stops text the agent reads from changing any of it. Ends with a go, go-with-conditions or no-go verdict and the tests to pass first.",
      "tags": [
        "agent-payments",
        "spend-policy",
        "guardrails",
        "agentic-finance",
        "kill-switch",
        "audit-log",
        "prompt-injection",
        "stablecoin"
      ],
      "examples": [
        "Write a spend policy for our agent that pays suppliers in USDC on Base.",
        "Review our payment agent's guardrails before we give it a real wallet."
      ]
    },
    {
      "id": "payment-flow-failure-modes-review",
      "name": "Payment flow failure-modes review",
      "description": "Lists how a money-moving agent or payment flow can fail (wrong decision, manipulated input, tool timeouts, partial success, stale data, double payments), rates likelihood and impact, says how each would be noticed within minutes, and proposes prevention, recovery, idempotency and reconciliation rules.",
      "tags": [
        "failure-modes",
        "payments",
        "reliability",
        "idempotency",
        "reconciliation",
        "agentic-finance",
        "risk"
      ],
      "examples": [
        "What can go wrong when our agent settles invoices automatically?",
        "Design retries so a timeout never pays twice."
      ]
    }
  ],
  "documentationUrl": "https://zeroclickz.com/agents/zc-sentinel"
}