SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsweekly-report-generator

weekly-report-generator

Office & productivity

View repository

Updated 10-16 09:34

About this skill

Weekly Report Generator

Automatically generate professional work weekly reports from Git commit records, supporting single-week/multi-week batch generation.

Core Principles

  • Main-sub separation: The main agent is responsible for scheduling, and sub-agents are responsible for actual generation
  • Concise output: Only output the commit message array, without redundant information
  • Read-only Git logs: Reading project source code files is prohibited
  • Content cleaning is mandatory: Convert technical terminology into business language (executed by AI, cannot be skipped)
  • Task management: Use the TodoWrite tool to create a task list and track the generation status of each weekly report

Quick Start

Generating a weekly report only requires 2 steps：

Step 1: 运行编排脚本准备数据 → Step 2: 并行处理每个周报

Step 1: Run the orchestration script (prepare data)

python scripts/orchestrate_reports.py \
  --paths "E:\项目1,E:\项目2" \
  --time "本月" \
  --output "E:\周报输出" \
  --template "E:\模板.docx" \
  --format docx \
  --naming "第一周周报,第二周周报,第三周周报,第四周周报"

The script automatically completes：

  • Parse time expressions (this week/this month/2025-1-1-2025-1-31)
  • Validate project paths and output paths
  • Analyze template files (identify sections that need to be supplemented)
  • Generate an independent task configuration file for each week
  • Generate secondary confirmation information and Claude Code invocation instructions

Output files(located in {输出路径}/tmp/）：

  • time_result.json - Time parsing result
  • template_structure.json - Template structure (if a template is provided)
  • week_XX-task.json - Task configuration for each week
  • claude_instruction.md - Claude Code invocation instructions

Step 2: Process each weekly report in parallel (executed by AI)

Create a task list(required):

# 使用 TodoWrite 工具创建任务清单
todos = [
  {"content": "读取调用说明和参考文档", "status": "pending", "activeForm": "读取调用说明和参考文档"},
  {"content": "为第1周启动子智能体", "status": "pending", "activeForm": "为第1周启动子智能体"},
  {"content": "为第2周启动子智能体", "status": "pending", "activeForm": "为第2周启动子智能体"},
  # ... 为每个周创建一个任务
  {"content": "汇总所有结果并清理临时文件", "status": "pending", "activeForm": "汇总所有结果并清理临时文件"}
]

Read reference documents：

  • Script parameters required: Read references/script-api-reference.md to view detailed parameters

Workflow：

  1. Read invocation instructions：{输出路径}/tmp/claude_instruction.md(mark the first task as in_progress)
  2. Launch sub-agents in parallel: Use the Task tool to launch an independent general-purpose sub-agent for each week
    • 📘 Invocation method: Refer to references/workflow.md Section 2.2 for detailed examples
    • Call the Task tool multiple times in the same response to achieve parallel processing
    • Mark the status for each subtask (→ in_progress when started, → completed when finished)
    • ⚠️ Important: Sub-agents must only return a brief success/failure status, and must not output detailed reports
  3. Aggregate results: Collect the success/failure status of all subtasks (mark as in_progress)
  4. Clean up temporary files: Delete successful temporary files, and keep failed ones for debugging (mark as completed)

⚠️ Sub-agent output specification(important):

  • ✅ Only return a brief status: "✅ Week X weekly report generated successfully" or "❌ Week X weekly report generation failed: [reason]"
  • ❌ Do not output detailed work content, data statistics, terminology conversion examples, etc.
  • ❌ Do not output detailed execution step descriptions
  • ✅ Strictly output according to the template：
    • If a template is provided (--template parameter): output only the weekly report file in the template format

Complete workflow and Task tool invocation examples：📘 workflow.md

Reference documentation

📘 workflow.md

Complete workflow guide, including:

  • Detailed Step 2 execution steps
  • Sub-agent task prompt template (ready to copy and use)
  • Temporary file structure and cleanup rules

📘 examples.md

Common usage scenario examples:

  • Example 1: Generate this month's weekly report (Markdown format)
  • Example 2: Use a Word template
  • Example 3: Batch generate weekly reports for multiple weeks
  • Example 4: Multi-project summary

📘 script-api-reference.md

Detailed Python script invocation parameters and usage examples:

  • parse_time.py - Time parsing
  • analyze_template.py - Template analysis
  • get_git_logs.py - Git log retrieval (compact mode)
  • fill_template.py - Template filling and export
  • orchestrate_reports.py - Orchestrate all steps

Parameter description

orchestrate_reports.py parameters

  • --paths(required): list of project paths, comma-separated
  • --time(required): time expression
    • Relative time:本周、上周、本月、上月、本年、去年
    • Absolute time:YYYY-MM-DD or YYYY-MM-DD-YYYY-MM-DD
  • --output(required): output directory path
  • --template(optional): template file path (supports .md and .docx)
  • --format(optional, default md): output format (md or docx）
  • --naming(optional): list of weekly report naming rules, comma-separated
    • Recommended: specify a custom filename for each weekly report
    • Example:--naming "华电第一周周报,华电第二周周报"
    • The number of naming rules must match the number of generated weekly reports
    • When not provided, default naming is used:第X周周报.docx

Time expression examples

Expression Description Generated weekly reports
本周 Monday to Friday of the current week 1 report
本月 From the 1st to the last day of this month (divided by week) 4-5 reports
2025-1-1-2025-1-31 Specified time range Divided by week
2025-1-15 The week containing the specified date 1 report

Frequently asked questions

Q: Where are temporary files saved? A: {输出路径}/tmp/ directory, automatically cleaned up after success