### List custom models examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-custommodel/references/arkcli-custommodel-list.md

Various examples demonstrating filtering, pagination, and sorting options for the list command.

```bash
# 列出全部（默认分页）
arkcli models custommodel list

# 只看自己的
arkcli models custommodel list --mine

# 按状态过滤（支持逗号分隔或重复传）
arkcli models custommodel list --statuses ready
arkcli models custommodel list --statuses ready,processing

# 按训练类型过滤
arkcli models custommodel list --customization-types lora,sft

# 按 base 模型名 / ID 过滤
arkcli models custommodel list --base-models doubao-seed-1-6
arkcli models custommodel list --base-model-ids fm-xxxxx

# 模糊搜索（命中 name / id / base 模型 display name）
arkcli models custommodel list --search my-finetune

# 分页与排序
arkcli models custommodel list --page 1,20 --sort-by CreateTime --sort-order Desc

# 自动翻页
arkcli models custommodel list --mine --page-all --page-delay 500
```

--------------------------------

### Prepare Virtual Environment and Install SDK

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-train-finetune/references/ark-finetune-sdk.md

Sets up a Python 3.10+ virtual environment and installs the latest Ark SDK package.

```bash
python3.11 -m venv /tmp/ark-rl     # 必须 ≥ 3.10
source /tmp/ark-rl/bin/activate
pip install https://ark-public-example-cn-beijing.tos-cn-beijing.volces.com/ark-sdk/ark_sdk-latest.tar.gz
```

--------------------------------

### Update custom model examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-custommodel/references/arkcli-custommodel-update.md

Examples showing how to update the name, description, or both for a custom model.

```bash
# 改名
arkcli models custommodel update cm-xxxxx --name new-display-name

# 改描述
arkcli models custommodel update cm-xxxxx --description "updated for v2 release"

# 同时改两个
arkcli models custommodel update cm-xxxxx --name v2 --description "..."
```

--------------------------------

### Endpoint Creation Examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-infer-endpoint/references/arkcli-infer-endpoint-create.md

Examples demonstrating how to create an endpoint with default settings, explicit billing methods, and how to capture the returned ID for future use.

```bash
# 创建 Endpoint
arkcli infer endpoint create \
  --model doubao-seed-2-0-pro-260215 \
  --name my-endpoint

# 显式要求 token 推理方式（当前与不传时的创建 payload 保持一致，但会先校验模型是否支持 token）
arkcli infer endpoint create \
  --model doubao-seed-2-0-pro-260215 \
  --name my-token-endpoint \
  --billing-method token

# 典型返回
{
  "Id": "ep-20260421180049-ngwkm"
}

# 只提取 Id，供后续命令复用
endpoint_id=$(arkcli infer endpoint create \
  --model doubao-seed-2-0-pro-260215 \
  --name my-endpoint \
  --transform Id)
```

--------------------------------

### Purchase plan command examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-plans/references/arkcli-plans-buy.md

Examples covering the protocol agreement flow, dry-run estimation, and direct purchase for individual and team plans.

```bash
arkcli plans buy --plan agent-plan --type small --duration 1
```

```bash
arkcli plans buy --plan agent-plan --type small --duration 1 --dry-run
```

```bash
arkcli plans buy --plan agent-plan --type small --duration 1 --yes
```

```bash
arkcli plans buy --plan agent-plan-team --type medium --duration 2 --quantity 5 --yes
```

```bash
arkcli plans buy --plan coding-plan-team --type pro --quantity 10 --yes
```

--------------------------------

### Generate Model SDK Examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-code-example/references/arkcli-code-example.md

Basic usage examples for generating code samples, including filtering by model version, language, and output format.

```bash
arkcli +code-example --model doubao-seedream-5-0
```

```bash
arkcli +code-example --model doubao-seedream-5-0 --version 260128
arkcli +code-example --model doubao-seedream-5-0-260128
```

```bash
arkcli +code-example --model doubao-seedream-5-0 --lang python
```

```bash
arkcli +code-example --model doubao-seedream-5-0 --lang curl
```

```bash
arkcli +code-example --model doubao-seedream-5-0 --output-dir ./my-examples
```

```bash
arkcli +code-example --model doubao-seedream-5-0 --format json
```

--------------------------------

### Generate model example code

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-code-example/SKILL.md

Retrieve and save example code for a specified foundation model to the local filesystem.

```bash
arkcli +code-example --model doubao-seedream-5-0
```

```bash
arkcli +code-example --model doubao-seedream-5-0 --version 260128 --lang python --format json
```

--------------------------------

### Generation Command Examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-gen/references/arkcli-gen.md

Various examples for image and video generation tasks, including multi-input, frame control, and advanced configuration flags.

```bash
# 1) 文生图 (T2I) — seedream 模型
arkcli +gen --model doubao-seedream-5-0-260128 "简约商务笔记本电脑，深蓝色，白色背景"

# 2) 图生图 / image-edit (I2I) — 用 --input 艾特参考图
arkcli +gen --model doubao-seedream-5-0-260128 \
  --input @ref.jpg \
  "保留构图，把背景换成黄昏的海滩"

# 多张参考图（多张同时传 --input）
arkcli +gen --model doubao-seedream-5-0-260128 \
  --input @style1.jpg --input @style2.jpg \
  "融合两张图的风格，主体是一只柴犬"

# 一次出多张候选（image-count > 1 自动启用 sequential）
arkcli +gen --model doubao-seedream-5-0-260128 --image-count 4 "未来城市天际线 4 个风格变体"

# 3) 文生视频 (T2V)
arkcli +gen --model doubao-seedance-1-5-pro-251215 "一只柴犬在樱花树下奔跑，慢镜头"

# 4) 图生视频 / 首帧到视频 (I2V) — 第 1 张图默认作为首帧
arkcli +gen --model doubao-seedance-1-5-pro-251215 \
  --input @first.jpg \
  "镜头缓慢拉远，主体保持不动"

# 显式指定 first / last 帧
arkcli +gen --model doubao-seedance-1-5-pro-251215 \
  --input first:@start.jpg --input last:@end.jpg \
  "从这两帧之间生成补间动画"

# 5) 参考视频 (R2V) — 把视频当作参考素材
arkcli +gen --model doubao-seedance-2-0-r2v-260128 \
  --input ref:@reference.mp4 \
  "保持参考视频的运动轨迹，替换主角为机器人"

# 6) 参考音频 — 让视频节奏与音频同步
arkcli +gen --model doubao-seedance-2-0-260128 \
  --input ref:@beat.mp3 \
  "随节拍切换的城市夜景蒙太奇"

# 进阶 flag — 图片任务
arkcli +gen --model doubao-seedream-5-0-260128 \
  --guidance-scale 7.5 --optimize-prompt --output-format png \
  "产品摄影风格"

# 进阶 flag — 视频任务
arkcli +gen --model doubao-seedance-2-0-260128 \
  --frames 96 --camera-fixed --return-last-frame --draft \
  "城市夜景航拍"

# 输出完整 JSON
arkcli +gen --model doubao-seedance-1-5-pro-251215 --format json "产品广告视频"

# 自定义推理接入点 (endpoint ID) — 必须显式 --modality
arkcli +gen --model ep-20260416234150-zsd4v --modality image "简约商务笔记本"
arkcli +gen --model ep-20260416234150-zsd4v --modality video "一只柴犬奔跑"
```

--------------------------------

### Custom model retrieval examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-custommodel/references/arkcli-custommodel-get.md

Examples demonstrating full detail retrieval, field filtering, and status polling.

```bash
# 完整详情
arkcli models custommodel get cm-xxxxx

# 只取关键字段（避免大块 JSON）
arkcli models custommodel get cm-xxxxx --transform id,name,active_endpoints,artifact_types

# 仅查 status（轮询用）
arkcli models custommodel get cm-xxxxx --transform status
```

--------------------------------

### JSON output for keys use

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-profile/references/arkcli-profile-keys.md

Example output format after switching the default key.

```json
{
  "profile": "platform_cn-beijing_default",
  "new_default": "abc****1234"
}
```

--------------------------------

### Generating Code Examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-infer-endpoint/references/arkcli-infer-endpoint-create.md

Generate multi-language API call examples for a specific base model.

```bash
arkcli +code-example --model doubao-seed-2-0-pro --lang python
```

--------------------------------

### Team Plan Output Example

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-helper/references/arkcli-helper.md

Example output when injecting MCP servers for a team agent plan, which excludes OpenViking services.

```text
团队版 Agent Plan 不含 OpenViking,已跳过数据面/控制面两台 MCP。
✓ 已为 claude-code 注入 MCP → [mcp-server-askecho-search-infinity dataPro-search]  (plan: agent-plan-team_cn-beijing_team)
提示:Claude Code 需重启后才会加载新注入的 MCP。
```

--------------------------------

### Example JSON response for plans get

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-plans/references/arkcli-plans-get.md

The command returns a JSON object containing an array of active plans, including their status and scope.

```json
{
  "plans": [
    {
      "key": "agent-plan",
      "name": "Agent Plan",
      "scope": "personal",
      "tier": "small",
      "status": "Effective"
    },
    {
      "key": "agent-plan-team",
      "name": "Agent Plan Team",
      "scope": "team",
      "tier": "medium",
      "status": "Running",
      "seat_id": "seat-..."
    }
  ]
}
```

--------------------------------

### Personal Plan Output Example

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-helper/references/arkcli-helper.md

Example output when injecting MCP servers for a personal agent plan, including all four services.

```text
✓ 已为 claude-code 注入 MCP → [mcp-server-askecho-search-infinity dataPro-search openviking-dataplane openviking-controlplane]  (plan: agent-plan_cn-beijing)
提示:Claude Code 需重启后才会加载新注入的 MCP。
```

--------------------------------

### Quantization Examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-custommodel/references/arkcli-custommodel-quantize.md

Standard usage examples for quantizing a model, including an optional description flag.

```bash
# 标准用法（必须先跑 available-quantizations 确认 mode 合法，并查看该 mode 预期支持的部署形态）
arkcli models custommodel quantize cm-xxxxx --quantization int8

# 同时附描述
arkcli models custommodel quantize cm-xxxxx \
  --quantization int8 \
  --description "int8 quantized for low-latency inference"
```

--------------------------------

### Install Ark CLI

Source: https://github.com/volcengine/ark-cli/blob/main/README.md

Install the latest version of the Ark CLI globally using npm.

```shell
npm i -g @volcengine/ark-cli@latest
```

--------------------------------

### Common Usage Patterns

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-api-explorer/references/arkcli-api.md

Examples demonstrating minimal calls, pagination, and field extraction using --transform.

```bash
# 1) 最小调用（先跑通再逐步加字段）
arkcli api model.list_foundation_models --params '{}'

# 2) 加分页
arkcli api model.list_foundation_models --params '{"PageSize":10,"PageNumber":1}'

# 3) 只提取关键字段（示例 path，按实际输出结构调整）
arkcli api model.list_foundation_models --params '{"PageSize":10,"PageNumber":1}' --transform 'Result.Items.#.Name'
```

--------------------------------

### Install arkcli CLI

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-shared/SKILL.md

Global installation command for the public version of the arkcli package.

```bash
npm i @volcengine/ark-cli -g
```

--------------------------------

### Execute Metric Queries

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-doctor/references/scope-metrics.md

Examples of querying metrics with parameters, time windows, filters, and output formatting.

```bash
# 拿过去 1h Endpoint X 的 RPM, 按 api 拆
arkcli doctor metrics request.rpm.by_api --param endpoint=ep-202602xxx

# 模型 Y 过去 6h 的 P99 时延, 按 api+le 维度
arkcli doctor metrics request.duration.quantile \
  --param model=doubao-pro-32k \
  --param quantile=0.99 \
  --window 6h

# 错误码分布: 把 endpoint 限定在 ep-x, 用 --filter 过出仅 429
arkcli doctor metrics request.error.distribution \
  --param endpoint=ep-x \
  --filter code=429

# scalar 形态 (给 agent 直接拿数值, 单位与聚合策略由 catalog.aggregation_hint 决定)
arkcli doctor metrics request.error.rate \
  --param endpoint=ep-x \
  --format scalar
# → {"query_id":"request.error.rate","value":0.0234,"unit":"ratio","aggregation":"mean", ...}

# 表格 (面向人, 每条 series 一行: latest / mean / max)
arkcli doctor metrics request.rpm.by_api --format table
```

--------------------------------

### Initialize and Configure Ark CLI

Source: https://github.com/volcengine/ark-cli/blob/main/README.md

Commands to install, verify, authenticate, and connect Ark Skills to a local agent.

```shell
# 1. 安装 CLI
npm i -g @volcengine/ark-cli@latest
arkcli --version

# 2. 登录火山方舟
arkcli auth login volc-sso
arkcli auth status

# 3. 同步 Ark Skills 到本地 Agent
arkcli +connect
```

--------------------------------

### Understand Command Examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-understand/SKILL.md

A collection of common usage patterns for the +understand command, covering various media types and specialized sub-skills.

```bash
arkcli +understand image-caption --input @photo.jpg "描述这张图"
```

```bash
arkcli +understand image-caption --input @doc.png "OCR，原样保留文字"
```

```bash
arkcli +understand image-grounding --input @scene.jpg "框出所有红色car"
```

```bash
arkcli +understand doc-extract --input @invoice.pdf "抽取 发票号/金额/日期"
```

```bash
arkcli +understand video-summary --input @clip.mp4 "按 chapter 总结"
```

```bash
arkcli +understand video-qa --input @clip.mp4 "视频里出现了几辆车？"
```

```bash
arkcli +understand asr --input @speech.mp3
```

```bash
arkcli +understand asr-align --input @speech.mp3
```

```bash
arkcli +understand asr-speakers --input @meeting.wav
```

```bash
arkcli +understand ast --input @speech.m4a "译成英文"
```

```bash
arkcli +understand meeting-minutes --input @meeting.m4a
```

```bash
arkcli +understand "<prompt>" --input @photo.jpg
```

```bash
arkcli +understand asr --input @speech.mp3 --stream
```

```bash
arkcli +understand image-caption --input @x.jpg --system-prompt-append "用英文回答"
```

--------------------------------

### Pricing Plans Output Format

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-pricing/references/arkcli-pricing-plans.md

Example of the JSON structure returned by the pricing plans command.

```json
{
  "items": [
    {
      "plan_id": "agent-plan-personal-small",
      "product": "AgentPlan",
      "edition": "personal",
      "tier": "small",
      "price": 40,
      "original_price": 40,
      "period": "monthly",
      "currency": "CNY"
    },
    {
      "plan_id": "agent-plan-enterprise-small",
      "product": "AgentPlan",
      "edition": "enterprise",
      "tier": "small",
      "price": 0,
      "original_price": 0,
      "period": "monthly",
      "currency": "CNY",
      "error": "AccessDenied: caller lacks permission"
    }
  ]
}
```

--------------------------------

### JSON output for keys list

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-profile/references/arkcli-profile-keys.md

Example output format when listing keys in a profile.

```json
{
  "profile": "platform_cn-beijing_default",
  "default_api_key": "392****dab0",
  "available_api_keys": ["392****dab0", "abc****1234", "..."]
}
```

--------------------------------

### Verify SDK Version and Commands

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-train-finetune/references/ark-finetune-sdk.md

Checks the installed SDK version and verifies available command-line tools.

```bash
ark upgrade --dry-run
python -m pip show ark-sdk
ark --help
ark init workspace --help
ark create mcj --help
ark test pipeline_plugin --help
```

--------------------------------

### Activate foundation model services

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-models/references/arkcli-models-activate.md

Examples of activating models with various configurations, including sub-services, dry-run mode, and non-interactive flags.

```bash
# 默认仅开通基础服务（推理 + 精调）— 等价于 SubServices=["base"]
arkcli models activate doubao-seed-1-6-flash

# 同时开通基础服务与低延迟推理子服务（URL2 完整请求示例）
arkcli models activate doubao-seed-1-6-flash --sub-services base,fast-infer

# dry-run 校验（仅校验请求合法性，不实际开通；自动跳过 [Y/N] 提示）
arkcli models activate doubao-seed-1-6-flash --sub-services base,fast-infer --dry-run

# 非交互（CI / 脚本场景，跳过 [Y/N] 确认提示）
arkcli models activate doubao-seed-1-6-flash --yes
```

--------------------------------

### Renew Command Usage

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-plans/references/arkcli-plans-renew.md

Examples of how to use the renew command for guiding, estimating, and executing renewals.

```bash
# 第 1 步: 引导 — 拿到协议清单
arkcli plans renew --plan agent-plan
arkcli plans renew --plan agent-plan-team --duration 3 --seat-ids seat-001,seat-002

# 第 2 步: 询价 (可选)
arkcli plans renew --plan agent-plan --duration 6 --dry-run

# 第 3 步: 用户同意协议后, 真续费
arkcli plans renew --plan agent-plan --yes
arkcli plans renew --plan coding-plan --duration 6 --yes
arkcli plans renew --plan agent-plan-team --duration 3 --seat-ids seat-001,seat-002 --yes
arkcli plans renew --plan coding-plan-team --seat-ids seat-aaa --yes
```

--------------------------------

### Response Result JSON Structure

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-chat/references/caching-thinking.md

Example of the extended ResponsesResult object returned by +chat and chat get.

```json
{
  "id": "resp_...",
  "object": "response",
  "status": "completed",
  "created_at": 1746673200,
  "model": "...",
  "content": "...",
  "reasoning_content": "...",
  "usage": { ... },
  "store": true,
  "expire_at": 1746676800,
  "previous_response_id": "...",
  "caching": {"type": "enabled", "prefix": null},
  "thinking": {"type": "disabled"},
  "reasoning": {"effort": "medium"}
}
```

--------------------------------

### CLI 实测配置管理命令

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-config/references/evals.md

使用临时 HOME 目录执行配置的创建、查看、切换、删除及重置操作，避免污染真实环境。

```bash
tmp_home="$(mktemp -d)"

# 1) 创建 profile (0.1.16+: 走 profile 子树, 旧 `config init` 已 deprecated)
HOME="$tmp_home" arkcli profile create \
  --type platform \
  --region cn-beijing \
  --project default \
  --format json

# 2) list/show 只读排障（建议优先 show/list）
HOME="$tmp_home" arkcli profile list --format json
HOME="$tmp_home" arkcli profile show --format json --transform 'base_url'
HOME="$tmp_home" arkcli profile show --profile default --format json --transform 'config.base_url'

# 3) use/delete/reset 写操作（谨慎）
HOME="$tmp_home" arkcli profile use default --format json
HOME="$tmp_home" arkcli profile delete default --format json
HOME="$tmp_home" arkcli config reset --format json  # 整库清理仍走 config reset
```

--------------------------------

### Initialize Volcengine Profile via CLI

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-auth/references/arkcli-auth-init-volc.md

Execute the initialization command to create a platform profile from environment variables or perform a dry run to preview the configuration.

```bash
arkcli init-volc            # 消费 VOLC_INIT_*, 零交互, 落 platform profile + 设 default
arkcli init-volc --dry-run  # 只预览将创建的 profile, 不写盘
```

--------------------------------

### Create profiles using profile create

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-config/references/arkcli-config-init.md

Recommended replacement for config init, requiring an explicit profile type.

```bash
arkcli profile create --type platform --region cn-beijing --project default --set-default
arkcli profile create --type agent-plan --plan-tier medium --set-default
arkcli profile create --type coding-plan --plan-tier lite --set-default
```

--------------------------------

### Install AI Agent Skills

Source: https://github.com/volcengine/ark-cli/blob/main/README.md

Install built-in Agent Skills to enable natural language interaction with Ark services.

```shell
arkcli +connect
```

--------------------------------

### Initialize SDK Workspace

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-train-finetune/references/ark-finetune-sdk.md

Initializes a new project workspace using a specific template.

```bash
ark init workspace <workspace-dir> --template <template-name>
```

--------------------------------

### List Models and Resources

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-profile/SKILL.md

List models for plan-based profiles or show platform resources.

```bash
arkcli profile models list
arkcli profile show
```

--------------------------------

### Integrate with --store and chat get

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-chat/references/stream-events.md

Extracting a response ID from a streaming session to perform a subsequent lookup using chat get.

```bash
# 流式跑完, 从 completed 事件拿 id
RESP_ID=$(arkcli +chat "记住我叫小明" --model ep-xxx --store --stream --include-events \
  | jq -r 'select(.type == "response.completed") | .response.id' | head -1)

# 后续用 chat get 查完整响应
arkcli chat get "$RESP_ID" --format json | jq '{id, content, usage}'
```

--------------------------------

### View ArkCLI Fine-tune Help

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-train-finetune/references/create.md

Use these commands to inspect available flags and options for the fine-tuning creation process.

```bash
arkcli train finetune --help
arkcli train finetune create --help
```

--------------------------------

### Update infer endpoint examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-infer-endpoint/references/arkcli-infer-endpoint-update.md

Various examples for updating endpoint attributes, including renaming, clearing descriptions, adjusting rate limits, and performing dry runs.

```bash
arkcli infer endpoint update ep-20260428194457-p8b47 --name seed-2-0-pro-demo-123
```

```bash
arkcli infer endpoint update ep-20260428194457-p8b47 --description ""
```

```bash
arkcli infer endpoint update ep-20260428194457-p8b47 --rpm 60 --tpm 100000
```

```bash
arkcli infer endpoint update ep-20260507230659-skvgn \
  --cg-concurrent-requests 10 \
  --cg-create-task-rpm 60
```

```bash
arkcli infer endpoint update ep-20260428194457-p8b47 --name foo --dry-run
```

--------------------------------

### Define Agent Goal

Source: https://github.com/volcengine/ark-cli/blob/main/README.md

Example of providing a natural language goal to the Agent.

```text
帮我选择一个适合图文理解的模型，配置到本地 Agent，
然后用它总结一段视频，并告诉我本月用了多少 token。
```

--------------------------------

### 执行在线插件测试

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-train-finetune/references/testing-debug-monitoring.md

使用 ark CLI 工具查看插件测试命令帮助。

```bash
ark test pipeline_plugin --help
```

--------------------------------

### Stderr output for keys refresh

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-profile/references/arkcli-profile-keys.md

Example stderr message displayed after a successful refresh.

```text
[arkcli] profile "platform_cn-beijing_default" keys refreshed: +2 -0 (total 5)
```

--------------------------------

### Initialization Output Format

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-auth/references/arkcli-auth-init-volc.md

The command returns a JSON object confirming the status and details of the created profile without exposing sensitive credentials.

```json
{"ready":true,"profile":"platform_cn-beijing_accountwide","type":"platform","region":"cn-beijing","project":"账号全部资源","identity_key":"volc-2100000001","is_root":false,"has_api_key":true}
```

--------------------------------

### Execute arkcli API commands

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-api-explorer/SKILL.md

Examples for listing available actions, executing a specific action with parameters, and transforming the output using gjson paths.

```bash
arkcli api --list
arkcli api model.list_foundation_models --params '{"PageSize":10,"PageNumber":1}'

# 只提取 items 里的 name（示例 path，按实际输出结构调整）
arkcli api model.list_foundation_models --params '{"PageSize":10,"PageNumber":1}' --transform 'Result.Items.#.Name'
```

--------------------------------

### Example JSON response for seat-list

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-plans/references/arkcli-plans-team-seat-list.md

The structure of the JSON output returned by the seat-list command.

```json
{
  "plan": "agent-plan-team",
  "scene": "agent_plan_enterprise",
  "total": 42,
  "seats": [
    {
      "seat_id": "seat-20260608110804-gjrwr",
      "tier": "medium",
      "seat_status": "Active",
      "billing_status": "Running",
      "user_id": "83144215",
      "user_name": "yinfan.ivan",
      "project_name": "default",
      "instance_id": "ins-...",
      "order_time": 1700000000000,
      "expired_time": 1800000000000,
      "create_time": 1690000000000,
      "update_time": 1700000000000
    }
  ]
}
```

--------------------------------

### Prepare Model for Deployment

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-custommodel/SKILL.md

Sequence for listing ready models and selecting one as a target for deployment.

```bash
custommodel list --mine --statuses ready  → 选目标 cm-xxxxx
        → +deploy --model cm-xxxxx ...   （若已有 Running Endpoint 会复用；详见 ../arkcli-deploy/SKILL.md）
```

--------------------------------

### ImageStreamDelta JSON Structure

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-gen/references/image-stream.md

Example of the NDJSON event structure output by the CLI during streaming.

```json
{
  "event_type": "image_generation.partial_succeeded",
  "model": "doubao-seedream-5-0-260128",
  "created": 1700000000,
  "image_index": 0,
  "url": "https://...",
  "size": "1024x1024"
}
```

--------------------------------

### 监控任务状态

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-train-finetune/references/testing-debug-monitoring.md

使用 ark CLI 命令查看任务列表、详情及拉取插件产物。

```bash
ark list mcj
ark get mcj <mcj-id>
ark pull mcj <mcj-id> --include-plugin
```

--------------------------------

### Invoke Embedding Action

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-api-explorer/references/arkcli-api.md

Example of calling the arkruntime.create_embeddings action to generate text embeddings.

```bash
arkcli api arkruntime.create_embeddings --params '{"model":"doubao-embedding-large-text-250515","input":"火山方舟"}'
```

--------------------------------

### Querying Account Balances with arkcli

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-usage/references/arkcli-usage-balance.md

Examples of using the balance command with different types to retrieve free-quota, media-asset, and subscription plan information.

```bash
# 各模型的免费推理额度 / 资源包余额(走 ListModelChargeItems)
arkcli usage balance --type free-quota
arkcli usage balance --type free-quota --model doubao-seed-1-6
arkcli usage balance --type free-quota --modality LLM --page-size 20

# 媒资库容量(走 /open/GetAssetQuota)
arkcli usage balance --type media-asset
arkcli usage balance --type media-asset --project my-project

# 订阅套餐余额(走 QueryPlanUsage,精简版输出)
arkcli usage balance --type plan
arkcli usage balance --type plan --product agent-plan
arkcli usage balance --type plan --all
```

--------------------------------

### JSON output for keys refresh

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-profile/references/arkcli-profile-keys.md

Example output format after refreshing keys from the control plane.

```json
{
  "profile": "platform_cn-beijing_default",
  "default_api_key": "392****dab0",
  "available_api_keys": ["...", "..."],
  "added_count": 2,
  "removed_count": 0,
  "refresh_status": "ok"
}
```

--------------------------------

### Custom model upload examples

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-custommodel/references/arkcli-custommodel-upload.md

Various usage scenarios including minimal upload, uploading with quantization and description, and performing a dry-run to preview the request.

```bash
# 最简：必填三项
arkcli models custommodel upload \
  --name my-finetune-v1 \
  --base-model doubao-seed-1-6-251015 \
  --tos tos://my-bucket/finetune/v1/
```

```bash
# 带描述与上传时量化
arkcli models custommodel upload \
  --name my-finetune-v1-int8 \
  --base-model doubao-seed-1-6-251015 \
  --tos tos://my-bucket/finetune/v1/ \
  --quantization int8 \
  --description "lora-sft on customer support corpus, int8 quantized"
```

```bash
# dry-run 预览请求体（不实际提交）
arkcli models custommodel upload --name X --base-model Y --tos tos://b/p --dry-run
```

--------------------------------

### 执行 video-summary 视频总结

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-understand/references/sub-skills.md

按指定粒度对视频内容进行结构化总结。

```bash
arkcli +understand video-summary --input @lecture.mp4 "按 chapter 总结，每章给起始时间点"
```

--------------------------------

### arkcli models custommodel get

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-custommodel/references/arkcli-custommodel-get.md

Retrieves the details or status of a custom model identified by its ID.

```APIDOC
## arkcli models custommodel get

### Description
Retrieves the details or status of a custom model. Users can optionally filter the returned fields using the --transform flag to reduce output size.

### Usage
arkcli models custommodel get <id> [flags]

### Parameters
#### Path Parameters
- **id** (string) - Required - The custom model ID (e.g., cm-xxxxx).

#### Flags
- **--transform** (string list) - Optional - A whitelist of fields to return (e.g., id,name,status). Supported fields include id, name, status, foundation_model, active_endpoints, and artifact_types.

### Examples
# Retrieve full details
arkcli models custommodel get cm-xxxxx

# Retrieve specific fields
arkcli models custommodel get cm-xxxxx --transform id,name,active_endpoints,artifact_types

# Retrieve only status
arkcli models custommodel get cm-xxxxx --transform status

### Output
Returns a JSON object containing model details such as id, name, description, status, base_model, customization_type, source, active_endpoints, artifact_types, create_time, and update_time.
```

--------------------------------

### Get custom model details

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-custommodel/references/arkcli-custommodel-get.md

Basic command to retrieve full details of a custom model.

```bash
arkcli models custommodel get <id> [flags]
```

--------------------------------

### List and compare model prices by modality

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-pricing/references/arkcli-pricing-models.md

Retrieve pricing for all models within a specific modality for comparison.

```bash
arkcli pricing models --modality LLM --page-size 100
# Agent 在结果里按 vendor / FoundationModelName 排序汇总
```

--------------------------------

### List enterprise seats with arkcli

Source: https://github.com/volcengine/ark-cli/blob/main/skills/arkcli-plans/references/arkcli-plans-team-seat-list.md

Examples of using the seat-list command with various filters and pagination options.

```bash
# 列 Agent Plan 团队版全部席位（默认页大小）
arkcli plans team seat-list --plan agent-plan-team

# Coding Plan 团队版
arkcli plans team seat-list --plan coding-plan-team

# 按席位档位过滤
arkcli plans team seat-list --plan agent-plan-team --type medium

# 精准查指定席位（最多 1000 个 SeatID）
arkcli plans team seat-list --plan agent-plan-team --seat-ids seat-001,seat-002

# 按绑定子用户名过滤
arkcli plans team seat-list --plan agent-plan-team --user-name yinfan.ivan

# 仅看空闲席位（Idle=1）
arkcli plans team seat-list --plan agent-plan-team --seat-status 1

# 仅看 Running 计费状态
arkcli plans team seat-list --plan agent-plan-team --billing-status 2

# 翻页
arkcli plans team seat-list --plan agent-plan-team --page-number 2 --page-size 50
```