投放数据分析

基于超级直播、淘宝直播及财务数据，自动识别文件并计算ROI、转化率等关键指标，生成综合投放数据分析报告。

Security Scan

The skill's code and instructions are coherent with its stated purpose (local CSV/XLSX data processing and report generation); it does not request credentials or perform network activity, but there are minor documentation/configuration mismatches you should be aware of.

Purpose & Capability

Name/description match the code: modules read local 超级直播/淘宝直播/财务 files, compute ROI and other KPIs, and generate HTML/CSV reports. No network calls or external services are required. Minor mismatch: SKILL.md and examples refer to environment variables (TOUFANG_DATE_RANGE, TOUFANG_DATA_DIR, TOUFANG_OUTPUT_DIR) and a CLI name '投放数据分析', but the package metadata declares no required env vars and the provided entry points are Python scripts (main.py / clawhub_main.py). This is a documentation/configuration inconsistency, not an obvious malicious indicator.

Instruction Scope

Runtime instructions and code operate on local filesystem paths (default /Users/zhouhao/Documents/投放数据 and ~/Desktop outputs) and use environment variables if present. They do not read unrelated system config or network endpoints. Notes: SKILL.md shows single env var examples but the code references several optional env vars (TOUFANG_DATA_DIR, TOUFANG_OUTPUT_DIR, TOUFANG_METRICS, TOUFANG_OUTPUT_FORMAT). The README/usage examples mention a CLI name that isn't implemented as a packaged command — usage is via python main.py.

Install Mechanism

No install spec or remote downloads are used. Dependencies are listed in requirements.txt (pandas, numpy, chardet, openpyxl, etc.), which is appropriate for the stated purpose. No URLs, archive extraction, or third-party install hooks are present.

Credentials

The skill requests no credentials and does not access networked secrets. It does read and write local files (including user document and Desktop paths) which is expected given the purpose. Minor concern: the manifest does not declare any required env vars even though SKILL.md and code expect optional env vars; this is a transparency/configuration gap but not excessive privilege.

Persistence & Privilege

always is false and the skill does not attempt to modify other skills or system-wide settings. It creates output files in the user's configured output directory (default Desktop) and otherwise runs transient analysis — this matches normal behavior for a data-processing tool.

Assessment

This skill appears to be what it says: a local data-processing/reporting tool that reads CSV/XLSX files, computes KPIs, and writes HTML/CSV reports. Before installing or running: 1) Note the default data and output directories (/Users/zhouhao/... and ~/Desktop) and change them if you prefer to keep data elsewhere. 2) The SKILL.md shows environment variables (TOUFANG_*) but the registry listing does not declare them — the code will still read those env vars if you set them; decide which you want to use. 3) Install dependencies in a virtualenv (pip install -r requirements.txt) and inspect the three main scripts (main.py, clawhub_main.py, simple_run.py) if you want to confirm behavior. 4) Run on non-sensitive sample data first to confirm outputs and column name expectations (some cross-report formulas assume specific columns like 退货率 or 保量佣金). 5) If you need assurance about network/exfiltration risk: there are no network calls in the code, but if you modify the skill to add integrations, review those changes. Overall: functionally coherent and proportionate; proceed with normal caution for any tool that reads local files.

投放数据分析技能

技能概述

基于《数据分析基础概念和逻辑v3.md》文档开发的标准化投放数据分析技能，用于处理超级直播、淘宝直播和财务报表数据。

适用场景

• 万相台投放数据统计分析
• 直播数据多维度分析
• 财务数据与业务数据关联分析
• ROI优化和成本分析

ClawHub使用

通过环境变量配置：

# 设置环境变量
TOUFANG_DATE_RANGE="2026-01-01:2026-01-31"

# 运行技能
python3 clawhub_main.py

# 或者一次性设置
TOUFANG_DATE_RANGE="2026-01-01:2026-01-31" python3 clawhub_main.py

输入要求

• 时间范围：YYYY-MM-DD格式的日期范围
• 数据文件：自动识别三类数据文件
  • 超级直播数据（包含"超级直播"关键词）
  • 淘宝直播数据（包含"淘宝直播"关键词）
  • 财务数据（包含"财务"关键词）

输出内容

1. HTML汇总报表（桌面保存）
2. 数据质量检查报告
3. 关键指标计算
4. 优化建议

核心功能

1. 数据自动识别

# 自动识别数据文件
def auto_detect_files(data_dir):
    """自动识别三类数据文件"""
    super_files = find_files(data_dir, "超级直播")
    taobao_files = find_files(data_dir, "淘宝直播") 
    financial_files = find_files(data_dir, "财务")
    return super_files, taobao_files, financial_files

2. 编码自动处理

# 自动检测和处理编码
def auto_detect_encoding(file_path):
    """自动检测文件编码格式"""
    # 支持GBK、UTF-8等常见编码
    # 自动转换和统一处理

3. 字段映射计算

基于文档中的字段定义和计算公式：

超级直播关键计算：

• ROI = 总成交金额 / 花费
• 观看成本 = 花费 / 观看次数
• 订单成本 = 花费 / 总成交笔数
• 加购成本 = 花费 / (总收藏数 + 总购物车数)

淘宝直播关键计算：

• 成交转化率 = 成交人数 / 商品点击人数
• 客单价 = 成交金额 / 成交人数
• 笔单价 = 成交金额 / 成交笔数

财务报表关键计算：

• 业务口径收入 = 品牌费 + 切片 + 保量佣金 + 预估结算机构佣金 + 预估结算线下佣金
• 财务口径收入 = 业务口径收入 / 1.06
• 毛利率 = 毛利 / 财务口径收入

4. 跨报表关联分析

# 跨报表数据关联
def cross_report_analysis(super_df, taobao_df, financial_df):
    """基于文档的跨报表关联分析"""
    # 超级直播去退ROI参考值
    roi_adjusted = (super_df['总成交金额'] * (1 - taobao_df['退货率'])) / super_df['花费']
    
    # 推广投入回报率
    promotion_roi = (financial_df[['保量佣金','预估结算线下佣金','预估结算机构佣金']].sum() * 
                    (super_df['总成交笔数'] / taobao_df['成交笔数'])) / super_df['花费']
    
    return roi_adjusted, promotion_roi

使用示例

基本使用

# 调用投放数据分析技能
投放数据分析 --date-range "2026-01-01:2026-01-31" --data-dir "/Users/zhouhao/Documents/投放数据"

高级使用

# 包含特定指标计算
投放数据分析 --date-range "2026-01-01:2026-01-31" \
              --metrics "ROI,观看成本,订单成本" \
              --output-format "html,csv"

文件结构

投放数据分析技能/
├── SKILL.md          # 技能说明文档
├── requirements.txt  # Python依赖
├── main.py           # 主程序
├── data_processor.py # 数据处理模块
├── calculator.py     # 指标计算模块
├── reporter.py       # 报表生成模块
└── config.py         # 配置文件

依赖要求

• pandas >= 1.5.0
• numpy >= 1.21.0
• chardet >= 5.0.0

输出示例

技能执行后会生成：

1. YYYY-MM-DD_投放数据分析报告.html - 完整HTML报表
2. YYYY-MM-DD_数据质量检查.csv - 数据质量报告
3. YYYY-MM-DD_关键指标汇总.csv - 指标计算结果

错误处理

• 自动处理文件编码问题
• 字段缺失时的智能处理
• 数据质量异常预警

版本历史

• v1.0.0: 初始版本，基于数据分析基础概念和逻辑v3.md

作者

江风 - 交个朋友直播间