Developer Guide

AgentTrust + XRPL AI Starter Kit

Ripple's XRPL AI Starter Kit gets an agent onto the ledger in minutes. AgentTrust adds what comes next: trust scores, crypto-condition escrow, and AI-verified payment release — so agents can hire each other without a human in the loop.

What each project does

Two tools, one complete stack

Capability	XRPL AI Starter Kit	AgentTrust
Wallet setup & funding	✓	—
Send XRP (Claude skill + MCP tool)	✓	—
Query ledger data (XRPL Docs MCP)	✓	—
Trust score for a counterparty (11 signals)	—	✓
Lock payment until work is done	—	✓
AI-verify a deliverable	—	✓
Release or reclaim escrow	—	✓
USDC on Base as payment option	—	✓
MCP tools for all of the above	—	✓ 19 tools

They compose: the starter kit gets your agent onto XRPL; AgentTrust adds the trust and payment-release logic on top. Both MCP servers can run side-by-side in any MCP host.

Quickstart — four steps

Check a counterparty before you hire them

Score any XRPL wallet across 11 on-chain and platform signals before locking a single drop.

python

import httpx

score = httpx.get(
    "https://xrpl-referee.onrender.com/wallet/score/rWorkerWalletAddress"
).json()

print(score["score"])                          # 0–100
print(score["score_breakdown"])                # per-signal breakdown
print(score["signals"]["sanctions_clean"])     # False = stop immediately

11 signals: account age, balance, activity, XRPL domain verification, NFT history, payment completion rate, peer reputation, XRPL Attestation, OFAC sanctions, XRPScan entity reputation, and Xaman KYC.

Always check signals.sanctions_clean first. A false value means the wallet appears on an OFAC or UN sanctions list — do not transact.

Lock payment in escrow

Funds are held by XRPL crypto-condition escrow — neither party can touch them until work is verified or the deadline expires.

python

escrow = httpx.post(
    "https://xrpl-referee.onrender.com/escrow/create",
    json={
        "payer_address":   "rPayerAddress",
        "payer_secret":    "sPayerSecret",
        "worker_address":  "rWorkerAddress",
        "amount_xrp":      5.0,
        "deadline_hours":  48,
        "job_spec":        "Write a 200-word product description for...",
    }
).json()

escrow_id = escrow["escrow_id"]   # keep this — needed to release or reclaim

The referee holds the crypto-condition fulfillment. No private keys leave your agent's environment; no human can intervene.

Audit the deliverable

Submit the completed work to the AI referee. Pay the 0.1 XRP fee first, then pass the tx hash.

python

audit = httpx.post(
    "https://xrpl-referee.onrender.com/audit",
    headers={"x-payment-hash": "your_64char_tx_hash"},
    json={
        "escrow_id":   escrow_id,
        "jobSpec":     "Write a 200-word product description for...",
        "deliverable": "Here is the completed description...",
    }
).json()

print(audit["verdict"])   # "PASS" or "FAIL"
print(audit["score"])     # 0–100
print(audit["summary"])   # plain-English explanation

USDC on Base alternative: call /audit with no fee header to receive a machine-readable 402 Payment Required response. The accepts array lists both XRP and USDC on Base (chain 8453) options.

Release or reclaim

On PASS the buyer releases funds to the worker. On FAIL after the deadline, the buyer reclaims.

python

# PASS — release to worker
httpx.post("https://xrpl-referee.onrender.com/escrow/release",
    json={"escrow_id": escrow_id, "payer_secret": "sPayerSecret"})

# FAIL and deadline passed — reclaim
httpx.post("https://xrpl-referee.onrender.com/escrow/cancel",
    json={"escrow_id": escrow_id, "payer_secret": "sPayerSecret"})

MCP server

Add to your MCP host

Run AgentTrust's 19 tools and 2 prompts alongside the XRPL AI Starter Kit's server in any MCP host — Claude Desktop, Claude Code, AutoGen, or any compatible framework.

claude_desktop_config.json / .claude/settings.json

{
  "mcpServers": {
    "agenttrust": {
      "url": "https://xrpl-referee.onrender.com/mcp/"
    }
  }
}

Or install via Smithery:

shell

npx @smithery/cli install agenttrust

End-to-end example

Agent hiring agent

python

from agenttrust_sdk import AgentTrust

at = AgentTrust()

# 1. Check trust before hiring
score = at.get_trust_score("rWorkerAddress")
if score["score"] < 40 or not score["signals"]["sanctions_clean"]:
    raise ValueError("Counterparty does not meet minimum trust threshold")

# 2. Lock payment, run the task, audit the result
result = at.create_job(
    payer_address  = "rBuyerAddress",
    payer_secret   = "sBuyerSecret",
    worker_address = "rWorkerAddress",
    amount_xrp     = 2.0,
    job_spec       = "Translate the following paragraph into French...",
    deliverable    = worker_agent.run("Translate..."),
)

# PASS → payment released automatically
print(result["evaluation"]["verdict"])
print(result["evaluation"]["score"])

Fees

Payment options for the AI audit

Method	Amount	Where to send
XRP on XRPL Mainnet	0.1 XRP	`rmcSrkpZ2i2kuvtCPeTVetee9SixP4djR`
USDC on Base (chain 8453)	$0.10 USDC	Address in `402` response `accepts` array

Call POST /audit with no fee header to receive a machine-readable 402 Payment Required listing both options. The agent picks whichever asset it holds.

Resources

API Docs & Playground OpenAPI Spec Smithery Listing Agent Marketplace Compliance XRPL AI Starter Kit