> ## Documentation Index
> Fetch the complete documentation index at: https://docs.neus.network/llms.txt
> Use this file to discover all available pages before exploring further.

# Overview

> Give AI agents verifiable identity, permissions, spending limits, and checks before action.

NEUS gives AI agents and assistants a portable identity plus explicit permissions. Connected runtimes check those receipts before granting access or allowing a sensitive action.

| Need                  | How                                                         |
| --------------------- | ----------------------------------------------------------- |
| Identify the agent    | `agent-identity`: the agent account signs its identity      |
| Set what it can do    | `agent-delegation`: an approving account grants permissions |
| Cap what it can spend | `maxSpend` on the delegation (base units)                   |
| Check if it is ready  | `neus_agent_link` → `linked: true`                          |
| Be discoverable       | A2A and EIP-8004 agent cards                                |
| Reuse trust           | Receipt ID (`qHash`) or receipt URL                         |
| Audit later           | Receipt reads and proof pages                               |

Agent frameworks handle reasoning and execution. NEUS supplies verifiable identity, permission limits, and reusable trust receipts.

## Two receipts make an agent ready in NEUS

* **Identity** (`agent-identity`): the agent account signs its name, type, capabilities, and skills. Identity alone grants no permission.
* **Permissions** (`agent-delegation`): an approving account signs the allowed actions, spending cap, and expiration date.

An agent is **ready** only when both are on file.

## Permission limits

The permission receipt defines a boundary:

* **`allowedActions` / `deniedActions`**: explicit allow and deny lists. A denied action takes precedence.
* **`maxSpend`**: a whole-number spending cap in token base units. USDC uses six decimals, so 25 USDC is `"25000000"`.
* **`scope`**: the area where permissions apply, such as `payments:x402` or `global`.
* **`expiresAt`**: expiration time in Unix milliseconds. Set it for payments or other sensitive actions.

Quick setup presets (`full-access`, `payments`, `readonly`, `automation`) are available through `neus_agent_create`. Full field reference: [Agent delegation](./agent-delegation).

## Set it up

<Steps>
  <Step title="Connect">
    ```bash theme={"dark"}
    neus setup
    neus check
    ```

    Configures hosted NEUS MCP for Cursor, Codex, VS Code, or Claude Code.
  </Step>

  <Step title="Check">
    Call **`neus_context`**, then **`neus_agent_link`** with the agent wallet.
  </Step>

  <Step title="Create">
    If identity or permission receipts are missing, call **`neus_agent_create`**. The agent account signs its identity, and the approving account grants permissions.
  </Step>

  <Step title="Confirm">
    Call **`neus_agent_link`** until **`linked: true`**, then reuse receipts.
  </Step>

  <Step title="Connect">
    In your project repo:

    ```bash theme={"dark"}
    neus mount <agentId> --apply cursor
    ```

    Loads the agent's verified identity, permissions, and rules into the project. [Connect agent context](./runtime-mount).
  </Step>

  <Step title="Act">
    Check receipt status before payments, data access, or other sensitive actions.
  </Step>
</Steps>

## Next

<CardGroup cols={2}>
  <Card title="Verification flow" icon="diagram-project" href="./agent-verification-flow">
    Step-by-step order.
  </Card>

  <Card title="Agent identity" icon="id-card" href="./agent-identity">
    Register an agent.
  </Card>

  <Card title="Agent permissions" icon="handshake" href="./agent-delegation">
    Grant allowed actions, spending caps, and expiration dates.
  </Card>

  <Card title="Discover agents" icon="address-card" href="./named-agent-card">
    Public pages, A2A, and EIP-8004 cards.
  </Card>

  <Card title="Pay per call" icon="coins" href="../platform/x402">
    Charge for API requests with HTTP 402.
  </Card>

  <Card title="Connect agent context" icon="layer-group" href="./runtime-mount">
    Load a verified agent's identity and permissions into a project.
  </Card>

  <Card title="MCP setup" icon="plug" href="../mcp/setup">
    Connect supported editors with OAuth or an access key.
  </Card>
</CardGroup>