Code Auditor Plugin
Comprehensive code auditing and assessment plugin for Claude Code with parallel analysis agents, reusable assessment skills, and professional report generation.
Installation
# From the Sigistry marketplace
/plugin marketplace add code-auditor
# Or install directly
/plugin install https://github.com/Sigistry/marketplace/tree/main/plugins/code-auditor
Quick Start
Run a full codebase audit:
/full-audit
This dispatches 5 analysis agents in parallel, security, code quality, architecture, performance, and test coverage, then compiles a unified report with a weighted composite score.
Commands
| Command | Description |
|---|---|
/full-audit |
Run comprehensive audit with all agents in parallel |
/generate-report |
Generate professional HTML report from assessment results |
Agents
Agents run autonomously and can also be triggered automatically from natural conversation.
| Agent | Color | Scope |
|---|---|---|
security-scanner |
Red | Vulnerability scanning, injection detection, hardcoded secrets, dependency CVEs, compliance |
code-quality-analyzer |
Cyan | Code smells, complexity, duplication, standards violations, codebase metrics |
architecture-analyzer |
Blue | Architecture discovery, design patterns, dependency analysis, Mermaid diagrams |
performance-analyzer |
Yellow | Bottleneck detection, memory leaks, bundle optimization, caching analysis |
test-coverage-analyzer |
Green | Test coverage mapping, quality assessment, gap identification |
Skills
Skills provide reusable assessment methodology that agents and commands consume.
| Skill | Purpose |
|---|---|
assessment-scoring |
Weighted scoring methodology (Code Quality 25%, Performance 25%, Security 20%, Maintainability 20%, Testing 10%) |
security-methodology |
OWASP Top 10, CWE/CVE classification, compliance frameworks (PCI DSS, GDPR, HIPAA, SOC 2) |
architecture-diagramming |
C4 Model methodology, Mermaid diagram patterns, reusable diagram templates |
Typical Workflow
- Run
/full-auditfor comprehensive analysis - Review findings from each agent
- Check the weighted composite score (0-10)
- Follow the prioritized remediation roadmap
- Run
/generate-reportfor a shareable HTML report
Plugin Structure
code-auditor/
├── .claude-plugin/
│ └── plugin.json
├── commands/
│ ├── full-audit.md
│ └── generate-report.md
├── agents/
│ ├── security-scanner.md
│ ├── code-quality-analyzer.md
│ ├── architecture-analyzer.md
│ ├── performance-analyzer.md
│ └── test-coverage-analyzer.md
├── skills/
│ ├── assessment-scoring/
│ │ ├── SKILL.md
│ │ └── references/
│ ├── security-methodology/
│ │ ├── SKILL.md
│ │ └── references/
│ └── architecture-diagramming/
│ ├── SKILL.md
│ ├── references/
│ └── assets/diagram-templates/
└── README.md
Tech Agnostic
All agents and skills are language and framework agnostic. They automatically detect the technology stack and apply relevant analysis patterns. Supported ecosystems include JavaScript/TypeScript, Python, Java, Go, Ruby, PHP, C#/.NET, Rust, and more.
Requirements
- Claude Code CLI
License
MIT