Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

amazon-research-reviews-skill

v1.0.0

AI驱动的电商评论深度分析工具,支持22维度智能标签、用户画像识别、VOC洞察和可视化看板生成。 当用户需要以下功能时触发: - 分析电商产品评论(Amazon/eBay/AliExpress等平台) - 从评论中提取用户画像、痛点和VOC(客户之声) - 生成产品洞察报告和机会点分析 - 创建专业的可视化分析看...

0· 118·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for liangdabiao/amazon-research-reviews-skill.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "amazon-research-reviews-skill" (liangdabiao/amazon-research-reviews-skill) from ClawHub.
Skill page: https://clawhub.ai/liangdabiao/amazon-research-reviews-skill
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install amazon-research-reviews-skill

ClawHub CLI

Package manager switcher

npx clawhub@latest install amazon-research-reviews-skill
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description (review analysis, VOC, personas, dashboards) align with the included scripts: CSV reading, batch creation, prompt generation, tag/stats merging and HTML report generation. No unrelated credentials, binaries, or network endpoints are requested by the skill. The scripts produce prompts for an LLM and expect LLM responses to be saved back as JSON—this matches the declared AI-driven workflow.
Instruction Scope
SKILL.md instructs the agent to read CSV files, create batched AI prompts, and save prompt/result files; the included scripts create and write prompt files containing raw review text and sample rows. This is consistent with the purpose, but it explicitly involves sending potentially sensitive customer review contents to an LLM (the skill relies on the agent/LLM to process prompts). Users should be aware that review text (which may include PII) will be included in prompts sent to the model.
Install Mechanism
There is no declared install spec (instruction-only), which minimizes upfront install risk. However the Python code will attempt to pip-install openpyxl at runtime if an Excel file is detected (os.system('pip install openpyxl -q')). That modifies the runtime environment and can pull packages from PyPI. Shell scripts expect common tools (python3, jq, iconv) to be present. No remote download URLs or archive extraction were observed.
Credentials
The skill declares no environment variables, credentials, or config paths. The scripts read and write files under the skill/project output directories and do not request unrelated secrets. This is proportionate to a CSV-based analysis tool.
Persistence & Privilege
always:false and typical agent invocation are used. The skill writes output files into output/{product}_{date} and creates prompt/result files—this is expected. It does not request permanent system-level privileges or modify other skills' configurations.
Assessment
This skill appears to do what it claims: parse CSV reviews, create batched prompts, and produce labeled CSV/Markdown/HTML reports. Before installing or running it: 1) Understand that review text (including any PII in reviews) will be embedded in prompts and sent to an LLM — do not upload sensitive customer data unless you accept that exposure. 2) The code may pip-install openpyxl at runtime and expects python3, jq, iconv and common Unix tools; run it in an environment where installing packages is acceptable (or pre-install dependencies). 3) The skill writes files into the skill/project output folders—review where those files are stored and their permissions. 4) If you need stricter data governance, sanitize PII from reviews before analysis or run the skill in an isolated environment. If you want, I can list all files and the exact lines that create prompts or perform runtime installs so you can inspect them further.

Like a lobster shell, security has layers — review code before you run it.

latestvk9745henqqgygc4j48kkackb8983hw8n
118downloads
0stars
1versions
Updated 1mo ago
v1.0.0
MIT-0

Simple Review Analyzer

AI驱动的电商评论深度分析工具,提供22维度智能标签、用户画像识别和专业的可视化看板。

核心特性

22维度智能标签系统

全面覆盖评论信息的8大维度:

  • 人群维度 (4): 性别、年龄段、职业、购买角色
  • 场景维度 (1): 使用场景
  • 功能维度 (2): 满意度、具体功能
  • 质量维度 (3): 材质、做工、耐用性
  • 服务维度 (5): 发货速度、包装质量、客服响应、退换货、保修
  • 体验维度 (4): 舒适度、易用性、外观设计、价格感知
  • 市场维度 (2): 竞品对比、复购意愿
  • 情感维度 (1): 总体评价

三位一体输出

  1. CSV标签数据: 原始评论 + 22维度 AI 标签
  2. Markdown洞察报告: 战略机会点、痛点、优化建议
  3. HTML可视化看板: 6个交互式图表、黑金设计

四位一体VOC系统

  • 用户画像识别(3-4类典型用户)
  • 黄金样本(每类6条精选评论)
  • 3D头像系统(6种风格)
  • 情感分析(分布与关键词)

快速开始

首次使用

  1. 准备 CSV 文件(支持自动模糊匹配列名)
  2. 在 Claude Code 中调用:
    请分析这个产品的评论:reviews.csv
    

CSV 文件格式要求

必需列(自动模糊匹配):

  • 评论内容:内容/评价/body/review/text/comment
  • 评分:打分/rating/score/star

可选列

  • 时间:时间/date/日期/time
  • 标题:标题/title/summary
  • 用户名:用户/user/username

核心工作流程

第一步:收集参数

使用 AskUserQuestion 收集:

  • 分析数量(100条/300条/全部)

第二步:执行分析

  1. 读取并解析 CSV 文件
  2. 批量标签提取(每批最多30条)
  3. 统计分析与用户画像识别
  4. 生成洞察报告
  5. 生成 HTML 可视化看板

第三步:展示结果

output/ 目录下按产品名_日期创建文件夹,生成三种报告:

output/
├── 产品A_20260320/
│   ├── reviews_labeled.csv
│   ├── 分析洞察报告.md
│   └── 可视化洞察报告.html
├── 产品B_20260321/
│   ├── reviews_labeled.csv
│   ├── 分析洞察报告.md
│   └── 可视化洞察报告.html
└── ...

命名规则

  • 文件夹:{产品名}_{日期YYYYMMDD}
  • CSV:reviews_labeled.csv
  • Markdown:分析洞察报告.md
  • HTML:可视化洞察报告.html

技术架构

文件结构

simple-review-analyzer/
├── skill.md              # Skill 定义文件
├── prompts/              # 提示词模板
│   ├── tagging.txt      # 单条评论打标
│   ├── tagging_batch.txt # 批量打标
│   └── insights.txt     # 洞察报告生成
├── templates/            # 输出模板
│   └── report.html      # HTML 可视化模板
└── utils/               # 工具脚本
    ├── transform_logic.py  # JSON↔CSV双向转换工具
    └── csv_reader.sh    # CSV 读取辅助

输出文件:生成到 output/{产品名}_{日期}/ 目录

提示词模板

  • 单条打标: 基于 prompts/tagging.txt,分析单条评论返回22维度标签
  • 批量打标: 基于 prompts/tagging_batch.txt,一次处理最多30条评论
  • 洞察报告: 基于 prompts/insights.txt,生成深度分析报告

使用场景

场景1:产品优化

分析自己产品的评论,发现用户痛点,优化产品功能和设计。

场景2:竞品分析

分析竞品评论,了解竞争对手的优势和劣势,寻找差异化机会。

场景3:市场调研

批量分析多个产品的评论,了解市场需求、用户偏好和行业趋势。

场景4:用户洞察

深度了解目标用户群体,构建精准用户画像,优化营销策略。

输出文件说明

CSV 标签数据

  • 原始评论数据(保留所有原始列)
  • 22维度 AI 标签列(新增)
  • 用途:数据分析、导入 BI 工具、筛选过滤

Markdown 洞察报告

  • 执行摘要:分析概述、关键发现
  • 战略机会点:市场机会、产品差异化
  • 用户痛点:问题分类、改进建议
  • 用户画像:典型用户类型、需求特点
  • VOC 分析:黄金样本、情感分布

HTML 可视化看板

  • 黑金奢华设计风格
  • 6个交互式 Chart.js 图表
  • 响应式设计,支持移动端
  • 纯 HTML 文件,可直接浏览器打开

工具使用

transform_logic.py - 数据格式转换工具

位置:utils/transform_logic.py

功能:JSON(嵌套结构)与 CSV(扁平格式)之间的双向转换

使用方法

# JSON → CSV(扁平化输出,22维度标签展开为独立列)
python3 .claude/skills/simple-review-analyzer/utils/transform_logic.py json \
    tagged_reviews.json \
    reviews_labeled.csv

# CSV → JSON(恢复嵌套结构,22维度标签聚合为tags对象)
python3 .claude/skills/simple-review-analyzer/utils/transform_logic.py csv \
    reviews_labeled.csv \
    tagged_reviews.json

数据结构对比

JSON (嵌套)CSV (扁平)
tags.人群_性别人群_性别
tags.场景_使用场景场景_使用场景
tags.功能_满意度功能_满意度
......

应用场景

  • 将AI分析结果导出为Excel/BI工具兼容格式
  • 从已有标签数据恢复嵌套JSON结构
  • 数据格式验证和转换

注意事项

  1. 分析数量: 推荐选择 100 条评论以平衡速度与质量
  2. 并发限制: 每批最多处理 30 条评论
  3. 输出目录: output/{产品名}_{日期}/ (自动创建)
  4. 文件命名: 固定文件名,便于对比同一产品的多次分析
  5. 编码支持: UTF-8/GBK/GB2312 自动检测
  6. 工具使用: 可单独使用 transform_logic.py 进行数据格式转换

Comments

Loading comments...