Auto Weekly Report System

全自动周报系统。定时收集v3.5、InStreet、价格监控等数据，生成周报并发布到企业微信文档。当需要自动化报告生成、团队数据同步、定时数据汇总时使用。

Security Scan

The code and instructions are consistent with an automated weekly-report generator that reads local /tmp logs and produces a Markdown report for manual or scripted publishing to WeCom; behavior appears proportional to the stated purpose, with only minor documentation/invocation inconsistencies to fix before deployment.

Purpose & Capability

The skill declares an auto weekly report system and the included Python scripts implement data collection (from /tmp logs), report generation, and publishing helper commands for WeCom — these requirements align with the stated purpose. The files accessed (/tmp/* logs and json) match the described data dependencies.

Instruction Scope

SKILL.md and README instruct running generation and publish steps and scheduling a cron job; this stays within the expected scope. However, SKILL.md references scripts/generate_weekly.py (and README mentions generate_weekly.py) but the repo provides generator.py (which implements the generator) — a filename mismatch that will confuse users. The publisher does not perform live API calls but emits wecom_mcp command templates and saved shell scripts for manual execution; this is explicit in the code.

Install Mechanism

No install spec or external downloads are present (instruction-only install). All code is provided in the skill bundle; nothing is pulled from external URLs during install, so install risk is low.

Credentials

The skill requests no environment variables or credentials, which is coherent because publisher.py only prints wecom_mcp commands and requires the user to supply credentials manually. Be aware that actually publishing to WeCom will require the wecom_mcp CLI or API credentials (not requested by the skill). The scripts read local files under /tmp which could contain sensitive information; reading those files is proportional to the stated data-collection purpose but warrants review of those files before use.

Persistence & Privilege

always is false and the skill does not attempt to modify other skills or global agent settings. The scripts write temporary outputs and shell scripts under /tmp (weekly_data.json, weekly_report_auto.md, publish scripts) which is reasonable for this utility but should be considered when choosing the runtime user and scheduling cron jobs.

Assessment

This skill is internally coherent for producing weekly reports from local logs and preparing commands to publish to WeCom, but do the following before installing or scheduling it: 1) Fix the invocation mismatch: SKILL.md/README reference scripts/generate_weekly.py but the repository provides generator.py (run python3 scripts/generator.py or use full_pipeline.py). 2) Verify /tmp data files (agent_v35_production.log, instreet_reply.log, price_monitor_db.json) — they may contain sensitive info; ensure the process has permission only to needed files. 3) Understand publishing is manual: publisher.py prints wecom_mcp commands and writes shell scripts; you must have the wecom_mcp CLI and appropriate WeCom credentials in your environment to perform actual publishing. 4) Cron: run the scheduled job as a non-privileged user and consider using a dedicated directory (not world-writable /tmp) if you want to reduce risk of other users tampering with generated scripts. 5) Review the saved shell scripts (/tmp/publish_weekly.sh, /tmp/publish_commands.sh) before executing. If you need tighter security, run the workflow inside an isolated container or VM and audit the logs the skill will read.

Auto Weekly System

快速使用

生成本周报告：

python3 scripts/generate_weekly.py

发布到企业微信：

python3 scripts/publish.py --title "项目周报 2026-03-21"

完整流程（生成+发布）：

python3 scripts/full_pipeline.py

功能模块

1. 数据收集 (collector.py)

• v3.5 生产数据
• InStreet 回复统计
• 价格监控变动
• 系统健康检查

2. 报告生成 (generator.py)

• Markdown 格式化
• 数据可视化（表格、趋势）
• 异常高亮

3. 发布系统 (publisher.py)

• 创建企业微信文档
• 写入 Markdown 内容
• 返回分享链接

定时任务

建议设置：

# 每周五下午5点自动生成并发布
cron add --name "auto-weekly" --schedule "0 17 * * 5" \
  --command "python3 /path/to/full_pipeline.py"

数据依赖

• /tmp/v35_migration_config.json - v3.5配置
• /tmp/agent_v35_production.log - v3.5运行日志
• /tmp/instreet_reply.log - InStreet日志
• /tmp/price_monitor_db.json - 价格数据库