report-generator

name	report-generator
description	Generate structured Markdown and PDF reports from analysis results. Use when (1) rendering per-document analysis results into a readable report, (2) generating a full system review report from 7-dimension analysis, (3) producing a next-directions report from evolution tracking and gap analysis, (4) generating a PM development suggestion report with scoring cards, (5) producing a PRD draft report based on historical analysis, (6) creating Mermaid diagrams for evolution chains and dependency graphs, (7) converting Markdown reports to PDF. Also use when a user mentions 'generate report', 'export report', 'PDF report', 'review report', or 'analysis report'.
license	MIT
compatibility	opencode

Report Generator

Generate structured Markdown and PDF reports from all upstream Skill analysis results. The final deliverable step.

Report Types

Report	Source Skills	Description
Per-Analysis Report	Skill 2 + 3	Per-document 6-dimension analysis with quality scores
Full Review Report	Skill 4 (full_report)	Complete 7-dimension system review
Next Directions Report	Skill 4 + 5	Requirement direction recommendations
PM Development Report	Skill 4 (quality_assessment)	PM writing/thinking scoring card + growth path
PRD Draft Report	Skill 4 (prd_draft)	New PRD draft based on historical analysis
Insights Report	Skill 5	Evolution tracking + gap analysis with coverage matrix

Output Formats

Format	Description
Markdown	Structured .md file with tables, headings, Mermaid code blocks
PDF	Converted from Markdown with Chinese font support

Quick Start

All commands below assume the working directory is the skill root (skills/report-generator/). Install dependencies first: pip install -r requirements.txt

Generate reports

python scripts/generate.py <classify_json> <analysis_dir> <review_json> <output_dir> [options]

# Generate all reports from all upstream results
python scripts/generate.py classify.json ./analysis/ review.json ./reports/

# Specific report type only
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --report-type per_analysis
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --report-type full_review
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --report-type pm_development
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --report-type next_directions
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --report-type insights

# Include insights data (Skill 5 output)
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --insights-json insights.json

# Generate PDF output
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --format pdf

# Both MD and PDF
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --format all

# Polish report with LLM (optional)
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --polish

# Select specific sections
python scripts/generate.py classify.json ./analysis/ review.json ./reports/ --sections overview,per_analysis,evolution

Input

Required

1. Classify result JSON — from prd-overview-classify
2. Analysis directory — containing prd-per-analysis output JSONs
3. Review result JSON — from system-review
4. Output directory — where reports will be written

Optional

• Insights JSON — via --insights-json from requirement-insights (adds evolution/gap sections)
• Report type — via --report-type (default: all)
• Output format — via --format (md, pdf, all; default: md)
• Section selection — via --sections for cherry-picking sections
• Polish — via --polish flag to run LLM-based report polishing

Output Format

{
  "project_name": "智能联动",
  "files": [
    {
      "type": "markdown",
      "path": "/output/逐篇分析报告.md",
      "size": 94822
    },
    {
      "type": "markdown",
      "path": "/output/体系Review报告.md",
      "size": 48988
    },
    {
      "type": "pdf",
      "path": "/output/体系Review报告.pdf",
      "size": 89000
    }
  ],
  "mermaid_charts": [
    {
      "type": "evolution",
      "chart_id": "dynamic-delay",
      "code": "flowchart TD\n  ..."
    }
  ],
  "summary": {
    "total_reports": 3,
    "total_md_size": 143810,
    "chart_count": 5
  }
}

Report Structures

Per-Analysis Report

# {{project_name}} 需求文档逐篇分析报告

## 一、文档概览
（文档数量、分类分布、版本链概要）

## 二、逐篇分析
### 2.1 {{doc_title}}
- 核心问题：...
- 边界外问题：⚠️ 未解决 / ✅ 已解决
- 质量评分：★★★★☆ 4/5

## 三、边界外问题追踪汇总
| 问题 | 来源版本 | 解决状态 | 解决版本 |

## 四、需求演进脉络
（Mermaid 演进图）

## 五、文档质量评价

## 六、产品经理特征总结

Full Review Report

Same structure as Skill 4's generate_full_report_md() output — 7 sections matching the 7 dimensions.

Insights Report

# {{project_name}} 需求洞察报告

## 一、演进链追踪
（Mermaid 演进图 × N + 问题收敛统计）

## 二、功能覆盖矩阵
| 功能维度 | 覆盖文档 | 状态 |

## 三、需求缺口
（缺口列表 + 严重程度 + 建议）

## 四、功能重叠
（重叠列表 + 评估）

Key Features

• Template-driven rendering: Data tables auto-generated from JSON, no LLM needed for basic reports
• Mermaid diagram builder: Evolution chains, dependency graphs from structured data
• Optional LLM polish: --polish flag for human-like report refinement
• PDF support: Markdown → PDF with Chinese fonts
• Selective generation: --report-type and --sections for cherry-picking
• Composable: Each report type is independent, can be generated separately

Dependencies

pip install -r requirements.txt
# Core: pydantic
# PDF: reportlab (optional, for PDF output)

Scripts Reference

Script	Purpose
generate.py	Main entry: render reports + optional PDF
mermaid_builder.py	Mermaid diagram code generation

Prompts Reference

Prompt	Purpose
report-polish.md	Optional LLM-based report polishing

For detailed API and customization, see references/usage-guide.md.