mcp

<p align="center"> <img src="https://raw.githubusercontent.com/evidiq/mcp/main/assets/evidiq.png" width="80" alt="EVIDIQ" /> </p> <h1 align="center">EVIDIQ MCP</h1> <p align="center"><strong>Remote MCP server and install endpoint for the EVIDIQ trust layer.</strong></p> <p align="center"> <a href="https://evidiq.dev">evidiq.dev</a> &middot; <a href="https://evidiq.dev/mcp">evidiq.dev/mcp</a> &middot; <a href="https://github.com/evidiq/evidiq-skill">Agent Skill</a> </p> <p align="center"> <a href="https://glama.ai/mcp/servers/evidiq/mcp"><img src="https://glama.ai/mcp/servers/evidiq/mcp/badges/score.svg" alt="Glama score" /></a> <a href="https://evidiq.dev/mcp"><img src="https://img.shields.io/badge/MCP%20Server-Verified-6E56CF?style=flat-square" alt="MCP Server" /></a> <a href="https://0g.ai"><img src="https://img.shields.io/badge/0G-TEE%20%2B%20Storage-00C2A8?style=flat-square" alt="0G TEE + Storage" /></a> <a href="https://www.oklink.com/xlayer"><img src="https://img.shields.io/badge/X%20Layer-Live-3CCF4E?style=flat-square" alt="X Layer" /></a> <a href="https://evidiq.dev/x402"><img src="https://img.shields.io/badge/x402-pay--per--call-2563EB?style=flat-square" alt="x402" /></a> <a href="https://okx.ai"><img src="https://img.shields.io/badge/OKX.AI-ASP-121212?style=flat-square&logo=okx&logoColor=white" alt="OKX.AI ASP" /></a> <a href="./LICENSE"><img src="https://img.shields.io/badge/License-MIT-3DA639?style=flat-square" alt="License: MIT" /></a> </p> --- This is the EVIDIQ **Model Context Protocol** server — the remote endpoint that exposes EVIDIQ's trust tools to any MCP-capable agent. It verifies a counterparty's capability, scores its risk, checks its on-chain reputation, and returns a signed **Trust Report** so an agent can decide to proceed, escrow, or walk away — before any value moves. Live at **`https://evidiq.dev/mcp`** (Streamable HTTP). ## Connect **Claude Code:** ```bash claude mcp add --transport http evidiq https://evidiq.dev/mcp ``` **`mcp.json` / `.mcp.json`:** ```json { "mcpServers": { "evidiq": { "url": "https://evidiq.dev/mcp" } } } ``` ## Tools | Tool | Access | What it returns | |------|--------|-----------------| | `how_to_install` | free | How to connect the server or install the skill. | | `get_evidiq_skill` | free | The latest EVIDIQ Agent Skill (open-skill markdown). | | `verify_agent` | **paid (x402)** | A Trust Report: score (0–100), tier, risk breakdown, recommendation, and a signed attestation anchored on 0G. | Resource: `evidiq://skill` — the live skill document. ## Pay-per-call (x402) `verify_agent` is metered **per call over [x402](https://evidiq.dev/x402)**. Unauthenticated calls receive an HTTP 402 challenge (scheme `exact`, EIP-3009); sign it and retry with a `PAYMENT-SIGNATURE` header. Pricing/discovery is at `GET /x402`. The free tools stay free, and the gate is a no-op until payment is configured, so the server runs in either mode. ## How a check runs ```text probe endpoint → score identity · capability · reputation · risk → AI risk analysis (GLM via 0G compute / TEE) → attest: keccak256(report) + 0G Storage + EIP-191 signature → Trust Report ``` ## Source layout | Path | Description | |------|-------------| | `route.ts` | MCP route handler — registers the tools + resource; mounts at `/mcp`. | | `lib/x402/` | x402 pay-per-call gate (config, challenge, verify, facilitator). | | `lib/trust/` | Endpoint probe, deterministic scoring, and report types. | | `lib/og/` | 0G Storage, compute (GLM / TEE), and attestation. | | `lib/skill.ts`, `lib/install.ts` | The skill document + install instructions. | | `lib/request-context.ts` | Per-request base URL + payment context. | The canonical deployment runs this handler as a Next.js Route Handler in the EVIDIQ platform. The skill it serves lives at [evidiq/evidiq-skill](https://github.com/evidiq/evidiq-skill). ## License MIT © 2026 EVIDIQ — see [LICENSE](./LICENSE).