x402 Trust Score
Trust scoring for MCP servers and x402 services. Know before you connect. Know before you pay.
x402 Protocol
Base Mainnet
USDC
No API Keys
No Accounts
What you get
87
PROCEED
Strong trust signals. Health endpoint OK (45ms). 3 governance invariants. 7 tools, well-documented. Live audit endpoint.
52
CAUTION
Gaps exist. No governance declarations. Tools found but poorly documented. Suggest $1 max spend.
18
DO NOT TRANSACT
Server unreachable. No agent.json. No tools found. No audit endpoint.
Endpoints
/tools/evaluate
$0.005 USDC
Full trust evaluation for any MCP server or x402 service. Returns 0-100 score with recommendation. Checks reachability, governance declarations, tool quality, and audit endpoints.
GET https://x402.asalvocreative.com/tools/evaluate?server_url=https://target-server.com/mcp
/tools/lint
$0.001 USDC
The only MCP tool definition linter. Validates naming, descriptions, parameter schemas, error handling, auth docs. 14 checks with specific fixes.
GET https://x402.asalvocreative.com/tools/lint?tools_json=[{"name":"my_tool",...}]
/tools/health
$0.002 USDC
Governance scoring for agent system prompts. Checks for authority leaks, silent inference, missing audit trails, scope violations. 0-100 score.
GET https://x402.asalvocreative.com/tools/health?system_prompt=You+are+a+trading+agent...
What It Checks
Reachability 15%
Health endpoint response, latency, HTTP status codes
Service Declaration 25%
agent.json presence, schema version, payment methods, governance invariants
Tool Quality 30%
MCP session init, tool count, description quality, parameter documentation
Governance 30%
Declared invariants, live audit endpoint, coupling bonus for both present
How It Works
1
Agent calls endpoint
GET /tools/evaluate with target server URL. x402 middleware returns 402 with payment instructions.
2
Agent pays $0.005 USDC
Signs USDC transfer on Base. Resends with payment proof. Settled in ~200ms via CDP facilitator.
3
Service is evaluated
4 live checks: health endpoint, agent.json, MCP tool listing, audit endpoint. 5-15 seconds.
4
Score returned
0-100 trust score with PROCEED/CAUTION/HIGH RISK/DO NOT TRANSACT recommendation. Machine-readable JSON.
Bazaar Trust Report
Every service on the x402 Bazaar, evaluated. Updated daily.
| Service |
Score |
Recommendation |
Tools |
Governance |
| Report loading... Run evaluate-bazaar.py to generate. |
Integration
// Using @x402/axios -- auto-handles 402 payment flow
import { withPaymentInterceptor } from "@x402/axios";
const client = withPaymentInterceptor(axios.create(), wallet);
const { data } = await client.get(
"https://x402.asalvocreative.com/tools/evaluate",
{ params: { server_url: "https://some-mcp-server.com/mcp" } }
);
if (data.result.includes("DO NOT TRANSACT")) {
console.log("Skipping untrusted server");
} else {
// Safe to connect
}
Discovery
Same tools available free via MCP at api.asalvocreative.com/mcp. The x402 endpoints are for agents that prefer pay-per-call without registration.
Governance: no silent inference, auditability at decision level, explicit authority transfer.