nirholas

repo-intelligence-guide

Guide to Lyra Intel — an intelligence platform for analyzing repositories of any size, from small projects to enterprise monorepos with millions of lines of code. 70+ analysis components including security scanning, AI integration, dependency mapping, and code quality metrics.

nirholas 247 69 Actualitzat fa 5 mesos
GitHub

Instal·la

npx skillscat add nirholas/three-ws/repo-intelligence-guide

Instal·la des del registre de SkillsCat.

SKILL.md

Lyra Intel — Repository Intelligence Guide

Lyra Intel is a deep repository analysis platform with 70+ analysis components. It handles everything from small projects to enterprise monorepos with millions of lines of code.

Analysis Categories

Category Components What It Checks
Security 12 Vulnerabilities, secrets, dependencies
Code Quality 10 Complexity, duplication, patterns
AI Readiness 8 AGENTS.md, llms.txt, MCP support
Architecture 8 Structure, coupling, layering
Dependencies 7 Outdated, vulnerable, license issues
Documentation 6 README, API docs, inline comments
Testing 6 Coverage, quality, patterns
Performance 5 Build time, bundle size, hot paths
CI/CD 4 Pipeline quality, deployment
Community 4 Contributing, COC, issue templates

Quick Start

# Install
pip install lyra-intel

# Analyze a repo
lyra-intel analyze https://github.com/nirholas/sperax

# Analyze local directory
lyra-intel analyze ./my-project

# Analyze with specific focus
lyra-intel analyze ./my-project --focus security,ai-readiness

Security Scanning

What It Detects

Issue Severity Example
Hardcoded secrets Critical API keys, private keys in code
SQL injection High Unsanitized database queries
XSS vulnerabilities High Unsafe HTML rendering
Vulnerable deps Variable Known CVEs in node_modules
Insecure crypto Medium Weak hashing algorithms
Path traversal High Unsanitized file paths
SSRF High Unvalidated URLs in requests

Example Output

Security Score: 87/100

Critical:
  ❌ Hardcoded API key in src/config.ts:15
  ❌ Private key pattern in .env.example

High:
  ⚠️ 3 vulnerable dependencies (npm audit)
  ⚠️ SQL injection in server/queries.ts:42

Medium:
  ℹ️ Missing Content-Security-Policy
  ℹ️ CORS allows all origins

AI Readiness Analysis

Lyra Intel checks how well a repo supports AI agent integration:

Check Description
AGENTS.md Has AI agent instructions?
CLAUDE.md Has Claude-specific guidelines?
llms.txt Has machine-readable documentation?
MCP Support Has MCP server configuration?
API Surface OpenAPI/GraphQL specs available?
Type Safety TypeScript/typed language?
Test Coverage Enough tests for AI to validate changes?
CI Pipeline Automated checks for AI-generated PRs?

Architecture Analysis

Project Structure Score: 91/100

├── src/
│   ├── app/         ✅ Clear separation (Next.js convention)
│   ├── store/       ✅ Centralized state management
│   ├── services/    ✅ Service layer pattern
│   ├── server/      ✅ Server-side logic separated
│   └── components/  ✅ UI components isolated
├── packages/        ✅ Monorepo packages
└── tests/           ✅ Test directory

Issues:
  ⚠️ 2 circular dependencies detected
  ⚠️ src/utils/ has 45 files (consider splitting)

MCP Server

{
  "mcpServers": {
    "lyra-intel": {
      "command": "lyra-intel",
      "args": ["serve"]
    }
  }
}

Tools:

  • analyzeRepo — Full repo analysis
  • securityScan — Security-focused scan
  • getScore — Quick quality score
  • getDependencyReport — Dependency health
  • getArchitectureMap — Architecture visualization

Report Formats

  • Terminal — Colorful CLI output with score badges
  • JSON — Machine-readable for CI/CD pipelines
  • Markdown — GitHub-ready report
  • HTML — Visual dashboard with charts

Links