> For the complete documentation index, see [llms.txt](https://docs.djzs.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.djzs.ai/for-agents/mcp.md).

# MCP server

```
https://mcp.djzs.ai/mcp
```

Streamable HTTP, stateless, no sign-in. It answers MCP clients only: a browser opening the address gets `406`. Source: `djzs-trust-mcp/src/index.ts`.

|                            |                                                                                                                 |
| -------------------------- | --------------------------------------------------------------------------------------------------------------- |
| Server                     | `djzs-trust-mcp` 1.0.0, title "DJZS Trade Audit"                                                                |
| Protocol versions accepted | 2025-11-25, 2025-06-18, 2025-03-26, 2024-11-05 (probed 2026-11-20; an unknown version negotiates to 2025-11-25) |
| Registry name              | `ai.djzs/trust-mcp` in the official MCP Registry                                                                |

## Connect

```bash
claude mcp add --transport http djzs-trust https://mcp.djzs.ai/mcp
```

```json
{ "mcpServers": { "djzs-trust": { "type": "http", "url": "https://mcp.djzs.ai/mcp" } } }
```

For stdio-only frameworks, the published bridge `@sifr0-dev/djzs-ai` (0.1.0) runs `mcp-remote` against the same address (`packages/djzs-ai-shim/`):

```bash
npx -y @sifr0-dev/djzs-ai
```

## Tools

| Tool                     | Cost      | Use it to                                                                                                                                |
| ------------------------ | --------- | ---------------------------------------------------------------------------------------------------------------------------------------- |
| `verify_pm_trade`        | 2.00 USDC | Check a prediction-market bet before placing it. [Reference](https://docs.djzs.ai/reference/verify-pm-trade.md)                          |
| `verify_perp_trade`      | 2.00 USDC | Check a perp or spot trade before placing it. [Reference](https://docs.djzs.ai/reference/verify-perp-trade.md)                           |
| `query_agent_trust`      | Free      | Read an agent's on-chain audit record before you delegate to it or pay it. [Reference](https://docs.djzs.ai/reference/registry-tools.md) |
| `query_pol_certificates` | Free      | Read individual audit certificates on Irys. [Reference](https://docs.djzs.ai/reference/registry-tools.md)                                |

The server's own instructions tell a model to confirm with the user before a paid call, and to send the user's own words as `intent`. An idea sent to the wrong tool is refused and not charged.

## Paying from an MCP client

The paid tools use x402 on Base mainnet: 2.00 USDC per call, exact scheme, paid to `0xc1923748669dFC3a79497d0403A90a275161eCCA` (the address in the live payment challenge; check it there, never from a transaction history). The tested client is `withX402Client` from `agents/x402`.

* **Raise the cap.** The client's default limit is 0.10 USDC and it refuses this price before signing anything. Set `maxPaymentValue: 2000000n` (2.00 USDC in 6-decimal units) or higher.
* **Fund first.** The wallet must hold at least 2.00 USDC on Base before the call: the facilitator simulates the transfer when it verifies, so an underfunded wallet fails with `INVALID_PAYMENT`. The facilitator pays gas.
* **Find the settlement.** Over MCP the settlement receipt is on the tool result, at `result._meta["x402/payment-response"]` (`success`, `transaction`, `network`, `payer`). It is not an HTTP header.
* **Ask first.** A `confirmationCallback` lets the agent ask a person before each payment.

Full walkthrough with code: the [Connect guide](https://djzs.ai/guide#pay).

## What a paid MCP call writes

Both writes are permanent and public. Neither can change the verdict; a failed write is reported in the result and the verdict stands.

* **A Proof-of-Logic certificate on Irys mainnet**, for every in-scope check. The result's `pol_certificate` says `anchored` (with the Irys id), `disabled` or `error`. `query_pol_certificates` reads them back.
* **The agent's on-chain trust score on Base**, only when you pass `agent_address`. The result's `trust_score` reports the write. `query_agent_trust` reads the record. Leave `agent_address` out if you do not want a public record tied to that wallet.

An optional `target_system` label is written on the certificate only with `target_system_subject` and a valid `target_system_signature` from that address; otherwise it is discarded.

The paid HTTP route writes a different record, an EAS attestation: [Receipts and certificates](https://docs.djzs.ai/how-a-verdict-is-made/receipts.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.djzs.ai/for-agents/mcp.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
