hibariba

gee-project-analyzer

0
0
# Install this skill:
npx skills add hibariba/skills --skill "gee-project-analyzer"

Install specific skill from multi-skill repository

# Description

Generate architecture, design, or methods reports using parallel analysis agents. Creates short index report with detailed sections and appendix. Applies SOLID principles for modular output. Use when analyzing projects, reviewing architecture, exploring design patterns, or documenting system components.

# SKILL.md


name: gee-project-analyzer
description: Generate architecture, design, or methods reports using parallel analysis agents. Creates short index report with detailed sections and appendix. Applies SOLID principles for modular output. Use when analyzing projects, reviewing architecture, exploring design patterns, or documenting system components.


Project Analysis

Generate project analysis reports through parallel agent workflow.

Flow

Discovery → Parallel Analysis → Index Report
  1. Discovery: Collect report type, depth, scope
  2. Analysis: Launch parallel agents, write to section files
  3. Assembly: Generate index report with links to sections and appendix

Output

.workflow/.scratchpad/analyze-{timestamp}/
├── {TYPE}-REPORT.md           # Short index (~50 lines)
├── sections/
│   ├── 01-overview.md
│   ├── 02-layers.md
│   ├── 03-dependencies.md
│   └── ...
└── appendix/
    ├── recommendations.md
    └── methodology.md

Report Types

Type Focus Areas
architecture overview, layers, dependencies, dataflow, entrypoints
design patterns, classes, interfaces, state
methods algorithms, paths, apis, logic
comprehensive all above

Execution

Phase 1: Discovery

Ask user for configuration:

const config = {
  type: "architecture|design|methods|comprehensive",
  depth: "overview|detailed|deep-dive", 
  scope: "**/*|src/**/*|custom"
};

See references/phase-01-discovery.md for interaction flow.

Phase 2: Parallel Analysis

Launch agents based on report type. Each returns markdown content directly.

const results = await Promise.all(
  focusAreas.map(area => Task({
    subagent_type: "explore",
    prompt: buildPrompt(area, config)
  }))
);

See references/phase-02-analysis.md for agent configuration.

Phase 3: Assembly

Merge agent outputs into index report with linked sections:

# {Report Title}

> Scope: {scope} | Depth: {depth}

## Executive Summary
{synthesized overview}

## Contents

| Section | Summary | Link |
|---------|---------|------|
| Overview | Core architecture | [View](sections/01-overview.md) |
| ... | ... | ... |

## Key Findings
1. {finding}
2. {finding}

## Appendix
- [Recommendations](appendix/recommendations.md)
- [Methodology](appendix/methodology.md)

See references/phase-03-assembly.md for details.

Templates

Template Purpose
templates/index-report.md Index report structure
templates/section-file.md Section file format
templates/recommendations.md Recommendations format

Examples

Example Scenario
examples/01-architecture-nodejs.md Express API architecture
examples/02-design-react.md React design patterns
examples/03-methods-algorithms.md Algorithm analysis
examples/04-comprehensive.md Full project analysis

Scripts

Script Purpose
scripts/validate-report.sh Validate report structure and links
scripts/check-quality.sh Run quality gates
scripts/extract-findings.py Extract key findings

Reference

Document Purpose
references/phase-01-discovery.md User interaction
references/phase-02-analysis.md Agent prompts
references/phase-03-assembly.md Report assembly
references/writing-style.md Paragraph-style standards
references/quality-standards.md Quality gates
references/methodology.md Analysis methodology

# Supported AI Coding Agents

This skill is compatible with the SKILL.md standard and works with all major AI coding agents:

Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.