Text Classifier Pro

Upload text or CSV files for AI-powered multi-template or custom text classification with structured labels, confidence scores, batch processing, and multi-f...

Security Scan

The package implements a local text-classification tool and its runtime requirements and network calls align with the stated purpose; nothing in the files indicates covert exfiltration or unrelated privileges, though there are minor documentation/config mismatches to be aware of.

Purpose & Capability

The name/description (text classification) match the included code (classify_text, classify_batch, parsing, web UI). Requested credentials (OpenAI API key and an optional license token) are consistent with the functionality. Minor inconsistencies: registry metadata lists no required env vars while SKILL.md/README recommend OPENAI_API_KEY and SECRET_KEY; requirements.txt includes the 'openai' package but code calls OpenAI via direct HTTP (requests) rather than the openai client.

Instruction Scope

SKILL.md instructs running the local Flask web app or CLI and to provide an OpenAI API key and optional license token. The runtime instructions and code only reference relevant files and endpoints (api.openai.com for classification and geo-api.yk-global.com for license validation). The license validation step posts the provided license token to an external domain — this is expected for license checking but is the sole external token transmission to note.

Install Mechanism

No install spec is provided (instruction-only install), and dependencies are declared in requirements.txt. There are no high-risk remote downloads or archive extracts. Users are expected to pip install the listed Python packages before running; this is normal for a local Python tool.

Credentials

The tool requires an OpenAI API key (used to call api.openai.com) and optionally a license token (sent to geo-api.yk-global.com for validation). These credentials are proportionate to the stated features. Caveats: registry metadata did not declare these env vars even though SKILL.md/README recommend OPENAI_API_KEY and SECRET_KEY; the license token is transmitted to a vendor domain for validation, so only supply it if you trust that domain.

Persistence & Privilege

The skill does not request elevated platform privileges (always: false). It runs a local Flask server and uses in-process memory for simple rate/count tracking. There is no evidence it modifies other skills or system-wide configuration.

Assessment

This skill appears to be a normal local text-classifier: it will ask you for your OpenAI API key (used to call OpenAI) and may ask for a license token that it posts to https://geo-api.yk-global.com for validation. Review and trust the geo-api.yk-global.com domain before supplying license tokens. Note minor documentation mismatches (env vars recommended in README but not declared in registry; requirements list 'openai' though code uses requests). If you plan to run it, inspect the repository (especially any remaining code under the truncated history/save functions) and run it locally rather than on a public server; only provide keys you trust being used in your local environment.

SKILL.md — Text Classifier Pro

触发词

文本分类, 文本打标, 文本分类器, text classifier, 批量分类, 文本归类, 文本标签, 分类打标

功能说明

上传文本/CSV → AI 自动分类打标 → 输出结构化标签结果 + 置信度

核心能力

• 多输入方式：纯文本粘贴、TXT/CSV 文件上传
• 6 大预设分类模板：意图/情感/行业/风险/优先级/内容分类
• 自定义分类：支持自定义标签体系和 Prompt
• 多格式导出：屏幕显示（Markdown 表格）/ CSV / Excel / JSON
• Token 验证：支持 4 档套餐（FREE/STD/PRO/MAX），本地 5 分钟缓存

套餐差异

功能	FREE	STD ¥9.9	PRO ¥29.9	MAX ¥99
分类次数	20次	200次/天	2000次/天	不限
批量处理	❌	✅ 50条	✅ 500条	✅ 5000条
预设分类器	3种	10种	50种	不限
自定义分类	❌	✅	✅	✅
置信度显示	❌	✅	✅	✅
历史记录	❌	7天	90天	365天
API 访问	❌	❌	✅	✅

Token 验证规则

• 验证接口：POST https://geo-api.yk-global.com/validate
• 请求头：Authorization: Bearer {api_key}，请求体：{}
• 响应字段：**valid**（非 success）
• 前缀：TEXT-CLS-*（TEXT-CLS-FREE / TEXT-CLS-STD / TEXT-CLS-PRO / TEXT-CLS-MAX）
• 网络错误 → 自动降级 FREE，不阻断使用
• 本地缓存 5 分钟

使用方式

1. Web 界面（推荐）

cd /home/gem/workspace/agent/workspace/text-classifier
pip install -r requirements.txt
python scripts/web_app.py
# 访问 http://localhost:5000

2. 命令行

# 单条文本分类
python -m scripts.classifier --text "这个产品很好用" --classifier "情感分类" --api-key "sk-..."

# 批量 CSV 文件分类
python -m scripts.classifier --file data.csv --classifier "意图分类" --api-key "sk-..." --output csv --output-path results.csv

# 使用自定义标签
python -m scripts.classifier --text "紧急问题需要处理" --custom-labels "高,中,低" --custom-prompt "请判断以下问题的优先级" --api-key "sk-..."

3. Python API

from scripts.classifier import classify_text, validate_token

# Token 验证
tier = validate_token("TEXT-CLS-PRO-xxxx")
print(f"Tier: {tier}")

# 单条分类
result = classify_text(
    text="这个产品非常好用，客服态度也很棒",
    classifier_name="情感分类",
    api_key="sk-...",
    show_confidence=True
)
print(result)
# {'label': '正面', 'confidence': 0.85, 'raw': '正面', ...}

文件结构

text-classifier/
├── scripts/
│   ├── __init__.py
│   ├── classifier.py      # 核心分类逻辑
│   └── web_app.py          # Flask Web 界面
├── SKILL.md
├── README.md
├── requirements.txt
└── test_text_classifier.py

注意事项

• CSV 文件自动识别文本列，支持 GBK/GB2312/UTF-8 编码
• 批量处理自动限速（5 req/sec）避免触发 API 限流
• 历史记录仅 STD 及以上套餐保留
• API Key 建议通过环境变量 OPENAI_API_KEY 注入

来源

本技能由 YK-Global 开发，如需购买收费版请访问 YK-Global.com