arkcli +understand:多模态理解工作流,通过 12 个任务型 sub-skill(4 模态:image/video/audio/file)在数据面 Responses API 引擎上做专项理解。覆盖图片描述/OCR、视觉定位(bbox grounding)、GUI 操作识别、PDF/文档字段抽取、视频总结/视频问答、音视频联合理解、语音转写(ASR)、语音翻译(AST)、SRT 字幕打轴、多说话人转写、会议纪要。每个 sub-skill 自带专家 system prompt 与默认模型。当用户要『转写这段录音 / 把图里目标框出来 / 抽取这份 PDF 的字段 / 总结这段视频 / 生成字幕 / 识别会议纪要 / 描述这张图』这类有明确产出形态的多模态理解任务时使用;开放式带图对话用 +chat,生成图/视频用 +gen。
$npx -y skills add volcengine/ark-cli --skill arkcli-understandInstalls into the current project.
Run `npx skills use "https://github.com/volcengine/ark-cli" --skill "volcengine/ark-cli/arkcli-understand"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/volcengine/ark-cli" that are relevant to the current task. Run `npx skills add "https://github.com/volcengine/ark-cli"` and select the relevant skills, then follow their instructions.