Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
bytedance avatar

Byted Data Label

  • 31 installs
  • 411 repo stars
  • Updated August 4, 2026
  • bytedance/agentkit-samples

Byted data label is a Claude skill for the Seederive platform that uses an LLM to batch-label and analyze text, audio and image data.

About

This skill operates the Seederive platform to batch-label unstructured text, audio and image data with an LLM. It supports analysis types such as emotion detection, shill detection, opinion extraction, content scoring, translation and tag classification. A developer uses it to classify or annotate a batch of comments or records, starting with a lightweight quick-preview before creating full batch tasks.

  • Seederive platform skill that uses an LLM to batch-analyze and label text, audio and image data
  • Supports emotion detection, shill/water-army detection, opinion extraction, content scoring, translation and tag classif
  • Offers a quick-preview path for fast tests plus batch tasks, tag bases and prompt optimization

Byted Data Label by the numbers

  • 31 all-time installs (skills.sh)
  • Ranked #1,102 of 2,064 Data Science & ML skills by installs in the Skillselion catalog
  • Data as of Aug 5, 2026 (Skillselion catalog sync)
At a glance

byted-data-label capabilities & compatibility

Requires Volcengine AK/SK for the Seederive platform.

Capabilities
data analysis · translation
Use cases
data analysis · translation
Runs
Runs locally
Pricing
Bring your own API key
From the docs

What byted-data-label says it does

Seederive 用 LLM 对文本/语音/图片数据做情感分析、标签分类、观点提取等批量处理。
SKILL.md
python3 ${SKILL_DIR}/scripts/seederive.py task quick-preview \
SKILL.md
| `--max-rows` | 否 | 最大处理行数(默认 10,上限 50) |
SKILL.md
npx skills add https://github.com/bytedance/agentkit-samples --skill byted-data-label

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs31
repo stars411
Last updatedAugust 4, 2026
Repositorybytedance/agentkit-samples

What it does

Batch-label and analyze text, audio and image data (emotion, tags, opinions, translation) with an LLM via Seederive.

Who is it for?

Batch classification, sentiment analysis, tagging, opinion extraction or translation over a set of records.

Skip if: One-off single-item generation; it is built for batch analysis of text/audio/image datasets.

When should I use this skill?

The user wants to label, annotate, classify, translate or score a batch of text/comments, or detect shills in data.

What you get

Returns per-record labels such as emotion, tags, opinions, scores or translations, optionally exported to CSV.

  • Labeled records (JSON or CSV)
  • Emotion/opinion/score/tag outputs

By the numbers

  • 8 analysis node types (emotion, shill, opinion, scoring, translation, tag, subject, custom)
  • quick-preview default 10 rows, max 50 rows

Files

SKILL.mdMarkdownGitHub ↗

Seederive 非结构化打标平台

你是 Seederive 平台的操作助手。所有 Seederive 操作从这里开始。

什么是 Seederive

Seederive 用 LLM 对文本/语音/图片数据做情感分析、标签分类、观点提取等批量处理。

认证配置

使用前需要设置 AK/SK 环境变量:

环境变量说明必填
VOLCENGINE_ACCESS_KEYAccess Key
VOLCENGINE_SECRET_KEYSecret Key

验证连通性

设置好环境变量后,执行以下命令验证:

python3 ${SKILL_DIR}/scripts/seederive.py task list --page-size 1

如果返回 "code": 0 表示连通成功。如果返回认证错误,请检查 AK/SK 是否正确。

执行命令的方式

python3 ${SKILL_DIR}/scripts/seederive.py <子命令和参数>

第一步:判断用户意图

阅读用户的需求,对照下表确定属于哪个场景:

场景用户说了什么(示例)下一步
A. 快速试效果"帮我分析这几条评论" / "试一下情感分析" / "看看这些文本的标签"→ 直接用 quick-preview,见下方「场景 A」
B. 创建批量任务"帮我对这个数据表做情感分析" / "建一个打标任务"→ 读取 ${SKILL_DIR}/references/task.md 获取详细指引
C. 需要标签体系"按我们的标签分类" / "建一个标签库" / "主体识别"→ 读取 ${SKILL_DIR}/references/tag-base.md 获取详细指引
D. 优化效果"效果不好" / "帮我优化" / "上传错题" / "换个模型"→ 读取 ${SKILL_DIR}/references/optimize.md 获取详细指引
E. 不确定"我有一批数据想处理" / "能做什么"→ 先问用户数据是什么、想得到什么结果,再回到本表判断
重要:场景 B/C/D 的具体操作步骤、参数说明、JSON 格式都在对应的参考文件中。你必须用 Read 工具读取对应文件后再执行,本文件不包含这些细节。

场景 A:快速试效果(唯一可以直接执行的场景)

这是最轻量的路径,无需创建任务,传几条文本就能看结果。

支持的分析类型

分析类型nodeType 值输出额外参数
情感分析EMOTION_DETECTION正面/负面/中性 + 原因
营销水军识别SHILL_DETECTION是/否 + 原因
观点提取OPINION_SUMMARY核心观点 + 理由
内容评分CONTENT_SCORING质量/原创/有用/合规评分
翻译TRANSLATION翻译结果--target-language
标签分类TAG_DETECTION多级标签--tag-base-id(需要先建标签库,见场景 C)
主体识别SUBJECT_DETECTION多级主体--tag-base-id(需要先建标签库,见场景 C)
自定义分析CUSTOM_APPLICATION自定义--prompt + --output-fields

执行方式

方式一:直接传文本(推荐,最快)

python3 ${SKILL_DIR}/scripts/seederive.py task quick-preview \
  --raw-data '["文本1", "文本2", "文本3"]' \
  --node-type EMOTION_DETECTION \
  --input-column "评论内容"

方式二:上传文件

python3 ${SKILL_DIR}/scripts/seederive.py task quick-preview \
  --file data.csv \
  --node-type EMOTION_DETECTION \
  --input-column "评论内容"

方式三:导出结果为 CSV 文件

python3 ${SKILL_DIR}/scripts/seederive.py task quick-preview \
  --raw-data '["文本1", "文本2"]' \
  --node-type EMOTION_DETECTION \
  --input-column "评论内容" \
  --response-format csv --output result.csv

自定义分析示例

python3 ${SKILL_DIR}/scripts/seederive.py task quick-preview \
  --raw-data '["今天天气真好", "堵车堵了两小时"]' \
  --node-type CUSTOM_APPLICATION \
  --input-column "内容" \
  --prompt "提取关键词和情绪强度" \
  --output-fields '[{"fieldName":"keywords","fieldType":"String"},{"fieldName":"intensity","fieldType":"String"}]'

quick-preview 全部参数

参数必填说明
--raw-data与 file 二选一JSON 字符串数组或对象数组
--raw-data-file与上二选一JSON 文件路径
--file与 raw-data 二选一CSV / Excel 文件
--node-type分析类型,见上表
--input-column待处理文本的列名
--max-rows最大处理行数(默认 10,上限 50)
--tag-base-idTAG/SUBJECT 需要标签库 ID
--promptCUSTOM 需要自定义提示词
--output-fieldsCUSTOM 需要输出字段 JSON 数组
--target-languageTRANSLATION 用目标语言(默认"中文")
--response-formatjson(默认)或 csv
--outputCSV 输出文件路径

场景之间的流转

场景 A(试效果)
  │
  ├─ 效果满意 + 数据量大 → 场景 B(建正式任务批量跑)
  │                          → 读取 ${SKILL_DIR}/references/task.md
  │
  ├─ 需要标签分类 → 场景 C(先建标签库)→ 回到 A 或 B
  │                  → 读取 ${SKILL_DIR}/references/tag-base.md
  │
  └─ 效果不满意 → 场景 D(优化提示词/换模型)→ 回到 A 验证
                   → 读取 ${SKILL_DIR}/references/optimize.md

关键原则

1. 先试后建:建议用户先用 quick-preview 试效果,满意后再创建正式任务 2. 渐进披露:不要一次给用户灌输所有概念,按需引导到对应参考文件 3. 按需加载:只有需要执行场景 B/C/D 时才去读取对应参考文件

Related skills

FAQ

What data types can it label?

Text, audio and image data, batch-processed with an LLM.

What analysis types are supported?

Emotion detection, shill detection, opinion summary, content scoring, translation, tag detection, subject detection and custom analysis.

How do you test quickly?

Use quick-preview to pass a few texts and see results without creating a full task.

Data Science & MLanalyticspipelines

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.