---
name: orchardcore-ai-a2a-client
description: Skill for configuring the CrestApps A2A Client feature in Orchard Core. Covers adding remote A2A hosts, authentication, assigning agent connections, and how remote agents become tools. Use this skill when requests mention Orchard Core A2A Client, Configure an A2A Client, Feature Overview, Enable the A2A Client Feature, Add a Connection via Admin UI, Supported Authentication Types, or closely related Orchard Core implementation, setup, extension, or troubleshooting work. Strong matches include work with CrestApps.OrchardCore.AI.A2A, CrestApps.OrchardCore.AI, ApiKey, OAuth2ClientCredentials, OAuth2PrivateKeyJwt, OAuth2MutualTls, CustomHeaders, Profile, A2AAgentProxyTool, listAvailableAgents, findAgentForTask, findToolsForTask. It also helps with Add a Connection via Admin UI, Supported Authentication Types, Assign Agent Connections to AI Profiles, plus the code patterns, admin flows, recipe steps, and referenced examples captured in this skill.
license: Apache-2.0
metadata:
  author: CrestApps Team
  version: "1.0"
---

# Orchard Core A2A Client - Prompt Templates

## Configure an A2A Client

You are an Orchard Core expert. Generate configuration and admin guidance for connecting Orchard Core to remote A2A hosts through the CrestApps A2A Client feature.

### Guidelines
- Use feature ID `CrestApps.OrchardCore.AI.A2A`.
- A2A client connections are managed under **Artificial Intelligence → Agent to Agent Hosts**.
- Each connection points to a remote A2A host whose agent card is discovered at `/.well-known/agent-card.json`.
- Remote agent cards are cached for 15 minutes per connection and invalidated when the connection changes.
- Connected remote agents become AI tools available to the model.
- Sensitive authentication values are encrypted at rest with ASP.NET Core Data Protection.

### Feature Overview

| Feature | Feature ID | Description |
|---------|-----------|-------------|
| A2A Client | `CrestApps.OrchardCore.AI.A2A` | Connect to remote A2A hosts and use their agents |

### Enable the A2A Client Feature

```json
{
  "steps": [
    {
      "name": "Feature",
      "enable": [
        "CrestApps.OrchardCore.AI",
        "CrestApps.OrchardCore.AI.A2A"
      ],
      "disable": []
    }
  ]
}
```

### Add a Connection via Admin UI

1. Navigate to **Artificial Intelligence → Agent to Agent Hosts**.
2. Click **Add Connection**.
3. Enter:
   - **Display Text**: A friendly name such as `Production Agent Hub`.
   - **Endpoint**: The base A2A host URL, such as `https://agents.example.com`.
   - **Authentication**: Choose the authentication method required by the remote host.
4. Save the connection.

The system resolves the agent card automatically from `/.well-known/agent-card.json` on the configured host.

### Supported Authentication Types

| Type | Description |
|------|-------------|
| `Anonymous` | No authentication |
| `ApiKey` | Sends an API key in a configurable header with an optional prefix |
| `Basic` | HTTP basic authentication with username and password |
| `OAuth2ClientCredentials` | OAuth 2.0 client credentials flow |
| `OAuth2PrivateKeyJwt` | OAuth 2.0 client credentials with private key JWT assertion |
| `OAuth2MutualTls` | OAuth 2.0 client credentials with mutual TLS |
| `CustomHeaders` | Arbitrary HTTP headers defined as JSON |

### Assign Agent Connections to AI Profiles

1. Open the AI Profile editor.
2. Go to the **Capabilities** tab.
3. Under **Agent Connections**, check the desired remote A2A hosts.
4. Save the profile.

### Assign Agent Connections to AI Profile Templates

1. Open the AI Template editor for a template whose source is `Profile`.
2. Go to the **Capabilities** tab.
3. Under **Agent Connections**, check the desired connections.
4. Save the template.

### Assign Agent Connections to Chat Interactions

1. Open the Chat Interaction editor.
2. Go to the **Parameters** tab under **Capabilities**.
3. Under **Agent Connections**, select the desired connections.
4. Save the interaction.

### How Remote Agents Work

When a profile or interaction has A2A connections configured:

1. Orchard fetches and caches the remote agent card.
2. Each remote skill is registered as an AI tool.
3. The AI model can invoke those tools through the `A2AAgentProxyTool`.
4. The remote agent response is returned as tool output.

### Built-In A2A Discovery Functions

When the A2A client feature is enabled, these system functions are available:

| Function | Description |
|----------|-------------|
| `listAvailableAgents` | Lists local and remote agents |
| `findAgentForTask` | Finds the best-matching agent for a task |
| `findToolsForTask` | Finds the best-matching AI tools for a task |

### Permissions

| Permission | Description |
|------------|-------------|
| `Manage A2A Connections` | Required to create, edit, and delete A2A connections |

### Best Practices

- Use descriptive connection names so editors can distinguish environments.
- Use authenticated connections for production hosts.
- Assign only the connections a profile actually needs.
- Keep remote host agent descriptions current because they affect tool selection.