Arize 模型提供商集成

作者原文

Creates, reads, updates, and deletes Arize AI integrations that store LLM provider credentials used by evaluators and other Arize features. Supports any LLM provider (e.g. OpenAI, Anthropic, Azure OpenAI, AWS Bedrock, Vertex AI, Gemini, NVIDIA NIM). Use when the user mentions AI integration, LLM provider credentials, create integration, list integrations, update credentials, delete integration, or connecting an LLM provider to Arize.

SkillPi 解读

能完成的任务

Arize 模型提供商集成

适合谁

请按作者原文中的适用场景判断。

输入和输出

以作者完整文档为准；本站未运行验证产物。

软件、账号及付费要求

使用前按完整文档确认软件、账号、费用及权限；由宿主核对并授权实际文件读写和外部服务。文档引用域名（非运行白名单）：app.arize.com, integrate.api.nvidia.com, my-llm-proxy.example.com, my-resource.openai.azure.com, new-endpoint.example.com 会访问 Arize 的数据集、追踪、评估或AI集成，部分命令可批量修改或删除，调用模型可能收费；先确认账号/空间、数据接收方和操作范围。API密钥不写入公开日志，不自动授权新凭据集成或上传敏感数据。 外部工具按官方来源固定版本并单独核验；本站未安装或运行依赖。文件改动、外部发送、密钥和费用仍由使用者在宿主中逐项授权。

限制与区别

原文中的功能、版本状态和效果描述属于作者陈述；中文为对应翻译。仅核对固定版本及静态附件证据，不保证运行效果、外部服务或兼容性。

实测状态

未测试

安全评级与访问范围

caution · 静态证据核验，不代表运行安全保证。

使用前按完整文档确认软件、账号、费用及权限；由宿主核对并授权实际文件读写和外部服务。文档引用域名（非运行白名单）：app.arize.com, integrate.api.nvidia.com, my-llm-proxy.example.com, my-resource.openai.azure.com, new-endpoint.example.com 会访问 Arize 的数据集、追踪、评估或AI集成，部分命令可批量修改或删除，调用模型可能收费；先确认账号/空间、数据接收方和操作范围。API密钥不写入公开日志，不自动授权新凭据集成或上传敏感数据。 外部工具按官方来源固定版本并单独核验；本站未安装或运行依赖。文件改动、外部发送、密钥和费用仍由使用者在宿主中逐项授权。

使用前请由宿主核对并授权实际文件读写、外部服务和联网目的地；本站的核验不会授予任何运行权限。

原始 SKILL.md

---
name: arize-ai-provider-integration
description: Creates, reads, updates, and deletes Arize AI integrations that store LLM provider credentials used by evaluators and other Arize features. Supports any LLM provider (e.g. OpenAI, Anthropic, Azure OpenAI, AWS Bedrock, Vertex AI, Gemini, NVIDIA NIM). Use when the user mentions AI integration, LLM provider credentials, create integration, list integrations, update credentials, delete integration, or connecting an LLM provider to Arize.
metadata:
  author: arize
  version: "1.0"
compatibility: Requires the ax CLI and a configured Arize profile.
---

# Arize AI Integration Skill

> **`SPACE`** — Most `--space` flags and the `ARIZE_SPACE` env var accept a space **name** (e.g., `my-workspace`) or a base64 space **ID** (e.g., `U3BhY2U6...`). Find yours with `ax spaces list`.
> **Note:** `ai-integrations create` does **not** accept `--space` — AI integrations are account-scoped. Use `--space` only with `list`, `get`, `update`, and `delete`.

## Concepts

- **AI Integration** = stored LLM provider credentials registered in Arize; used by evaluators to call a judge model and by other Arize features that need to invoke an LLM on your behalf
- **Provider** = the LLM service backing the integration (e.g., `openAI`, `anthropic`, `awsBedrock`)
- **Integration ID** = a base64-encoded global identifier for an integration (e.g., `TGxtSW50ZWdyYXRpb246MTI6YUJjRA==`); required for evaluator creation and other downstream operations
- **Scoping** = visibility rules controlling which spaces or users can use an integration
- **Auth type** = how Arize authenticates with the provider: `default` (provider API key), `proxy_with_headers` (proxy via custom headers), or `bearer_token` (bearer token auth)

## Prerequisites

Proceed directly with the task — run the `ax` command you need. Do NOT check versions, env vars, or profiles upfront.

If an `ax` command fails, troubleshoot based on the error:
- `command not found` or version error → see references/ax-setup.md
- `401 Unauthorized` / missing API key → run `ax profiles show` to inspect the current profile. If the profile is missing or the API key is wrong, follow references/ax-profiles.md to create/update it. If the user doesn't have their key, direct them to https://app.arize.com/admin > API Keys
- Space unknown → run `ax spaces list` to pick by name, or ask the user
- LLM provider call fails (missing OPENAI_API_KEY / ANTHROPIC_API_KEY) → run `ax ai-integrations list --space SPACE` to check for platform-managed credentials. If none exist, ask the user to provide the key or create an integration via the **arize-ai-provider-integration** skill
- **Security:** Never read `.env` files or search the filesystem for credentials. Use `ax profiles` for Arize credentials and `ax ai-integrations` for LLM provider keys. If credentials are not available through these channels, ask the user.

---

## List AI Integrations

List all integrations accessible in a space:

```bash
ax ai-integrations list --space SPACE
```

Filter by name (case-insensitive substring match):

```bash
ax ai-integrations list --space SPACE --name "openai"
```

Paginate large result sets:

```bash
# Get first page
ax ai-integrations list --space SPACE --limit 20 -o json

# Get next page using cursor from previous response
ax ai-integrations list --space SPACE --limit 20 --cursor CURSOR_TOKEN -o json
```

**Key flags:**

| Flag | Description |
|------|-------------|
| `--space` | Space name or ID to filter integrations |
| `--name` | Case-insensitive substring filter on integration name |
| `--limit` | Max results (1–100, default 15) |
| `--cursor` | Pagination token from a previous response |
| `-o, --output` | Output format: `table` (default) or `json` |

**Response fields:**

| Field | Description |
|-------|-------------|
| `id` | Base64 integration ID — copy this for downstream commands |
| `name` | Human-readable name |
| `provider` | LLM provider enum (see Supported Providers below) |
| `has_api_key` | `true` if credentials are stored |
| `model_names` | Allowed model list, or `null` if all models are enabled |
| `enable_default_models` | Whether default models for this provider are allowed |
| `function_calling_enabled` | Whether tool/function calling is enabled |
| `auth_type` | Authentication method: `default`, `proxy_with_headers`, or `bearer_token` |

---

## Get a Specific Integration

```bash
ax ai-integrations get NAME_OR_ID
ax ai-integrations get NAME_OR_ID -o json
ax ai-integrations get NAME_OR_ID --space SPACE   # required when using name instead of ID
```

Use this to inspect an integration's full configuration or to confirm its ID after creation.

---

## Create an AI Integration

Before creating, always list integrations first — the user may already have a suitable one:

```bash
ax ai-integrations list --space SPACE
```

If no suitable integration exists, create one. The required flags depend on the provider.

### OpenAI

```bash
ax ai-integrations create \
  --name "My OpenAI Integration" \
  --provider openAI \
  --api-key $OPENAI_API_KEY
```

### Anthropic

```bash
ax ai-integrations create \
  --name "My Anthropic Integration" \
  --provider anthropic \
  --api-key $ANTHROPIC_API_KEY
```

### Azure OpenAI

```bash
ax ai-integrations create \
  --name "My Azure OpenAI Integration" \
  --provider azureOpenAI \
  --api-key $AZURE_OPENAI_API_KEY \
  --base-url "https://my-resource.openai.azure.com/"
```

### AWS Bedrock

AWS Bedrock uses IAM role-based auth. Provide the ARN of the role Arize should assume via `--provider-metadata`:

```bash
ax ai-integrations create \
  --name "My Bedrock Integration" \
  --provider awsBedrock \
  --provider-metadata '{"role_arn": "arn:aws:iam::123456789012:role/ArizeBedrockRole"}'
```

### Vertex AI

Vertex AI uses GCP service account credentials. Provide the GCP project and region via `--provider-metadata`:

```bash
ax ai-integrations create \
  --name "My Vertex AI Integration" \
  --provider vertexAI \
  --provider-metadata '{"project_id": "my-gcp-project", "location": "us-central1"}'
```

### Gemini

```bash
ax ai-integrations create \
  --name "My Gemini Integration" \
  --provider gemini \
  --api-key $GEMINI_API_KEY
```

### NVIDIA NIM

```bash
ax ai-integrations create \
  --name "My NVIDIA NIM Integration" \
  --provider nvidiaNim \
  --api-key $NVIDIA_API_KEY \
  --base-url "https://integrate.api.nvidia.com/v1"
```

### Custom (OpenAI-compatible endpoint)

```bash
ax ai-integrations create \
  --name "My Custom Integration" \
  --provider custom \
  --base-url "https://my-llm-proxy.example.com/v1" \
  --api-key $CUSTOM_LLM_API_KEY
```

### Supported Providers

| Provider | Required extra flags |
|----------|---------------------|
| `openAI` | `--api-key <key>` |
| `anthropic` | `--api-key <key>` |
| `azureOpenAI` | `--api-key <key>`, `--base-url <azure-endpoint>` |
| `awsBedrock` | `--provider-metadata '{"role_arn": "<arn>"}'` |
| `vertexAI` | `--provider-metadata '{"project_id": "<gcp-project>", "location": "<region>"}'` |
| `gemini` | `--api-key <key>` |
| `nvidiaNim` | `--api-key <key>`, `--base-url <nim-endpoint>` |
| `custom` | `--base-url <endpoint>` |

### Optional flags for any provider

| Flag | Description |
|------|-------------|
| `--model-name` | Allowed model name (repeat for multiple, e.g. `--model-name gpt-4o --model-name gpt-4o-mini`); omit to allow all models |
| `--enable-default-models` | Enable the provider's default model list |
| `--function-calling-enabled` | Enable tool/function calling support |
| `--auth-type` | Authentication type: `default`, `proxy_with_headers`, or `bearer_token` |
| `--headers` | Custom headers as JSON object or file path (for proxy auth) |
| `--provider-metadata` | Provider-specific metadata as JSON object or file path |

### After creation

Capture the returned integration ID (e.g., `TGxtSW50ZWdyYXRpb246MTI6YUJjRA==`) — it is needed for evaluator creation and other downstream commands. If you missed it, retrieve it:

```bash
ax ai-integrations list --space SPACE -o json
# or by name/ID directly:
ax ai-integrations get NAME_OR_ID
```

---

## Update an AI Integration

`update` is a partial update — only the flags you provide are changed. Omitted fields stay as-is.

```bash
# Rename
ax ai-integrations update NAME_OR_ID --name "New Name"

# Rotate the API key
ax ai-integrations update NAME_OR_ID --api-key $OPENAI_API_KEY

# Change the model list (replaces all existing model names)
ax ai-integrations update NAME_OR_ID --model-name gpt-4o --model-name gpt-4o-mini

# Update base URL (for Azure, custom, or NIM)
ax ai-integrations update NAME_OR_ID --base-url "https://new-endpoint.example.com/v1"
```

Add `--space SPACE` when using a name instead of ID. Any flag accepted by `create` can be passed to `update`.

---

## Delete an AI Integration

**Warning:** Deletion is permanent. Evaluators that reference this integration will no longer be able to run.

```bash
ax ai-integrations delete NAME_OR_ID --force
ax ai-integrations delete NAME_OR_ID --space SPACE --force   # required when using name instead of ID
```

Omit `--force` to get a confirmation prompt instead of deleting immediately.

---

## Troubleshooting

| Problem | Solution |
|---------|----------|
| `ax: command not found` | See references/ax-setup.md |
| `401 Unauthorized` | API key may not have access to this space. Verify key and space ID at https://app.arize.com/admin > API Keys |
| `No profile found` | Run `ax profiles show --expand`; set `ARIZE_API_KEY` env var or write `~/.arize/config.toml` |
| `Integration not found` | Verify with `ax ai-integrations list --space SPACE` |
| `has_api_key: false` after create | Credentials were not saved — re-run `update` with the correct `--api-key` or `--provider-metadata` |
| Evaluator runs fail with LLM errors | Check integration credentials with `ax ai-integrations get INT_ID`; rotate the API key if needed |
| `provider` mismatch | Cannot change provider after creation — delete and recreate with the correct provider |

---

## Related Skills

- **arize-evaluator**: Create LLM-as-judge evaluators that use an AI integration → use `arize-evaluator`
- **arize-experiment**: Run experiments that use evaluators backed by an AI integration → use `arize-experiment`

---

## Save Credentials for Future Use

See references/ax-profiles.md § Save Credentials for Future Use.

版权与许可声明

MIT License

Copyright GitHub, Inc.

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.