DJD Agent Score — Wallet Risk API for Paid Agent Workflows

What Is x402?

A simple payment rail for software-to-software calls

x402 lets software pay software per request with USDC on Base. Humans can still use monthly API keys. Agents can pay per call when they need to act on their own.

1

Pay

The route verifies payment

Your paid endpoint checks the x402 payment header before it starts doing expensive work.

2

Screen

AgentScore checks the payer wallet

You get a fast wallet decision before payout, tool calls, or route execution continue.

3

Decide

Allow, review, or block

Start by blocking only obvious bad wallets, then layer in history, forensics, alerts, and policy over time.

What To Expect

New wallets often look thin at first

A lot of Base wallets are still early. If you test a brand-new or lightly used wallet, AgentScore will often return Unverified or Emerging. That usually means thin history, not proven fraud. Confidence and explanation matter as much as the raw score.

1

Cold Start

Low score can mean low data

Thin wallets often lack transaction history, counterparties, and identity markers. Treat them as unknowns first, not automatically malicious actors.

2

Confidence

Use confidence with the score

High-confidence results are much more useful than a raw number alone. The API exposes confidence so you can separate real signal from early sparse data.

3

Best Practice

Start with allow, review, block

For new integrations, allow unknown wallets by default, review the middle, and only block obvious red-zone cases until you have your own workflow data.

See It Work

Watch a paid workflow make a trust decision

A wallet wants work, payment, or access. AgentScore helps your code decide what to do next.

You're building a legal research agent that pays for external services. Three wallets offer the work. Here's what AgentScore returns for each.

LexAgent

$15 / query

78Trusted

6-month history · 43 counterparties

CaseFinder

$8 / query

23Emerging

2-week history · 3 counterparties

LawBot99

$5 / query

9Unverified

Created yesterday · 0 completed jobs

Based on the scores, your agent's logic selects LexAgent. The cheapest option scored a 9 with zero transaction history. Your code caught it before it cost you anything.

You're building a content translation agent that pays for external services. Three wallets offer the work. Here's what AgentScore returns for each.

TranslateAI

$0.02 / word

64Established

3-month history · 18 counterparties

PolyGlot

$0.01 / word

14Unverified

Created 3 days ago · 1 counterparty

LingoPro

$0.03 / word

82Trusted

8-month history · 67 counterparties

Based on the scores, your agent's logic selects LingoPro. PolyGlot was created 3 days ago with a single counterparty. Your integration flagged it automatically.

You're building a market data agent that pays for external services. Three wallets offer the work. Here's what AgentScore returns for each.

DataStream

$0.05 / call

41Emerging

1-month history · 6 counterparties

⚠ sybil_risk

MarketPulse

$0.10 / call

71Established

5-month history · 31 counterparties

AlphaFeed

$0.08 / call

55Established

4-month history · 22 counterparties

Based on the scores, your agent's logic selects MarketPulse. DataStream triggered a Sybil flag, sharing a funding source with 12 other wallets. Your agent never saw it as an option.

Who It's For

Who pays for this

Usually the human team operating the workflow. They need a wallet decision before payout, before a paid route runs, or before routing work between counterparties.

1

Primary ICP

x402 API operators

You run paid Hono or x402 endpoints and need to reject low-trust payers before your handler spends compute, calls tools, or fans out to downstream services.

2

Primary ICP

Agent apps with payouts

Your app or agent sends funds to external wallets for data, research, translation, execution, or fulfillment. You want a fast screening layer before payout.

3

Secondary ICP

Marketplaces and protocols

You rank wallets, gate access, or route work between counterparties. You want a trust signal without building your own scoring, monitoring, and forensics stack.

Start Here

Start with screening. Add the ops layer later.

Most teams only need one trust decision at one workflow boundary. Start there. Add history, forensics, and alerts once that first screen is live.

1

Fastest Test

Score a wallet before payout

Use the free basic score endpoint to decide whether a counterparty gets the job, the funds, or another verification step.

curl "https://djdagentscore.dev/v1/score/basic?wallet=0x..."

Run a free lookup

2

Best Wedge

Gate your x402 route

Screen payer wallets before your Hono handler runs. Low-trust wallets can be rejected or treated differently without custom policy plumbing.

npm i x402-agent-score

Open the x402 example

3

Ops Layer

Turn on alerts

Create monitoring or webhooks when a watched wallet drops below a threshold, gets flagged, or shows up in forensics.

POST /v1/monitor

See monitoring plans

Try It

Score any wallet

Paste any Base wallet address. Free basic lookups daily. No account needed.

Free basic lookup. Full dimension breakdowns, integrity flags, and real-time refresh available via paid tiers.

Expect many brand-new wallets to come back as Unverified or Emerging at first. In early Base data, that usually means not enough history yet, not confirmed fraud.

—

This score is algorithmically generated from public blockchain data and unverified third-party submissions. It is not verified, audited, or guaranteed. Not a consumer report under the FCRA. Terms

What You Get

Screen first. Add operating depth later.

The first sale is wallet screening. History, forensics, monitoring, and supporting signals become useful after you decide a wallet matters to your workflow.

Screen

"Should I trust this wallet right now?"

Basic and full scores for payout checks, route gating, and simple allow/review/block policies.

History

"Is this wallet getting better or worse?"

Score history, trend views, and refresh tools when a one-time score is no longer enough.

Forensics

"What happened here?"

Fraud reports, disputes, watchlists, and timelines for wallets you need to investigate.

Alerts

"Tell me when something changes."

Webhook and monitoring subscriptions for score changes, anomalies, and forensics events.

Signals

"What else supports the decision?"

Counterparty context and optional identity metadata when you need more evidence around a wallet, not a separate product story.

Want the scoring internals? Read the methodology after you understand the screening workflow.

Integration

Start with the x402 gate

This is the cleanest wedge: verify payment, screen the payer wallet, and block obvious bad counterparties before your paid Hono route does any work.

Hono + x402

import { agentScoreGate } from 'x402-agent-score'

app.use(
  '/premium/*',
  agentScoreGate({
    minScore: 25,
    onUnknown: 'allow',
  }),
)

app.post('/premium/search', async (c) => {
  return c.json({ ok: true })
})

Install x402-agent-score Open docs See reference example

Endpoints

The core paths

Most teams only need a few paths to get started: one for screening, one for deeper context, and one for workflow ops. The rest lives behind monthly plans when you need more depth.

Endpoint

Price

Method

/v1/score/basic?wallet=

Composite score, tier, confidence, recommendation

Free tier

GET

/v1/score/full?wallet=

Full dimension breakdown, integrity flags, data quality

$0.10

GET

/v1/score/history?wallet=

Historical scores with trend and trajectory analysis

$0.15

GET

POST /v1/score/batch

Batch score up to 20 wallets in one request

$0.50

POST

/v1/forensics/summary?wallet=

Wallet risk summary, incidents, and penalties

$0.10

GET

POST /v1/monitor

Create monitoring subscriptions for watched wallets

Paid plan

POST

POST /v1/webhooks

Create wallet-scoped alert webhooks (API key required)

Paid plan

POST

Need live service health instead of raw JSON? Use the public status page.

Integrate

SDKs & middleware

Most teams can start with the REST API. If you want less plumbing, the middleware and SDKs are already published.

Package

Install

Type

x402-agent-score

Hono middleware to screen payer wallets before a paid x402 route runs

npm i x402-agent-score

x402

djd-agent-score

TypeScript/JavaScript client for the screening and forensics API

npm i djd-agent-score

SDK

djd-agent-score

Python client for API access, with agent-framework examples in the repo

pip install djd-agent-score

SDK

djd-agent-score-mcp

MCP server for Claude, GPT, or any MCP-compatible AI

npx -y djd-agent-score-mcp

MCP

Need a design-partner setup path?

If you run paid x402 routes, wallet payouts, or an agent marketplace, we can help you start with screening and add forensics, history, or monitoring only where the workflow needs it. Monthly plans include a small grace window so your workflow is less likely to stop exactly at quota.

Email for a pilot See pricing

Screen wallets before you pay them or your paid route does work.

The route verifies payment

AgentScore checks the payer wallet

Allow, review, or block

Low score can mean low data

Use confidence with the score

Start with allow, review, block

x402 API operators

Agent apps with payouts

Marketplaces and protocols

Score a wallet before payout

Gate your x402 route

Turn on alerts

Screen

History

Forensics

Alerts

Signals

Need a design-partner setup path?