A CLI tool that validates AI-generated code for common issues, anti-patterns, and security vulnerabilities. It analyzes code produced by AI assistants (like GitHub Copilot, Claude Code, etc.) and generates quality reports with improvement suggestions and confidence scores.

Features

Security Vulnerability Detection: Detect hardcoded secrets, SQL injection risks, command injection, and other security issues
Code Smell Detection: Identify code smells like unused imports, mutable defaults, complex functions, and poor error handling
Multi-language Support: Analyze Python, JavaScript, and TypeScript code
Confidence Scoring: Get a 0-100 score reflecting code quality and security posture
Multiple Output Formats: Terminal output with Rich formatting, JSON, and Markdown export options
Severity-based Filtering: Filter issues by severity level (critical, high, medium, low, info)

Installation

From Source

git clone https://7000pct.gitea.bloupla.net/7000pctAUTO/ai-code-audit-cli.git
cd ai-code-audit-cli
pip install -e .

Using pip

pip install ai-code-audit-cli

Quick Start

# Audit a single file
audit scan path/to/file.py

# Audit a directory recursively
audit scan path/to/project/

# Audit with JSON output
audit scan path/to/code/ --format json

# Audit with verbose output
audit scan path/to/code/ --verbose

# Filter by language
audit scan path/to/code/ --language python

# Filter by severity
audit scan path/to/code/ --severity high

Usage

Basic Scanning

# Scan a Python file
audit scan my_script.py

# Scan a JavaScript file
audit scan index.js

# Scan a TypeScript file
audit scan app.ts

# Scan a directory
audit scan src/

Output Options

# Rich terminal output (default)
audit scan path/to/code/

# JSON output for CI/CD
audit scan path/to/code/ --format json

# Markdown report
audit scan path/to/code/ --format markdown

# Quiet mode (minimal output)
audit scan path/to/code/ --quiet

Filtering Options

# Only scan Python files
audit scan path/to/code/ --language python

# Only show critical and high severity issues
audit scan path/to/code/ --severity critical high

# Skip color output
audit scan path/to/code/ --no-color

# Verbose output with more details
audit scan path/to/code/ --verbose

Configuration

Audit Configuration File

Create an audit.toml file in your project root:

[severity]
critical = true
high = true
medium = true
low = true
info = false

[languages]
python = true
javascript = true
typescript = true

[scanners]
bandit = true
ruff = true
tree_sitter = true

[confidence]
min_score = 70

PyProject.toml Integration

The tool automatically reads Ruff configuration from your pyproject.toml:

[tool.ruff]
line-length = 88
select = ["E", "F", "W", "C90", "I", "N", "UP", "B", "SIM"]
ignore = ["E501"]

[tool.ruff.per-file-ignores]
"__init__.py" = ["F401"]

Output Formats

Terminal Output

The default terminal output uses Rich formatting to display:

  AI Code Audit Report

  Summary:
  ┌─────────────────────────────────────────────┐
  │ Files Scanned      │ 15                    │
  │ Issues Found       │ 23                    │
  │ Confidence Score   │ 78                    │
  └─────────────────────────────────────────────┘

  Issues by Severity:
  ├─ Critical:  2
  ├─ High:      5
  ├─ Medium:    8
  ├─ Low:       6
  └─ Info:      2

JSON Output

{
  "summary": {
    "files_scanned": 15,
    "total_issues": 23,
    "confidence_score": 78
  },
  "issues": [
    {
      "id": "B001",
      "severity": "high",
      "category": "security",
      "message": "Possible hardcoded password detected",
      "file": "src/auth.py",
      "line": 42
    }
  ]
}

Markdown Output

# AI Code Audit Report

## Summary

| Metric | Value |
|--------|-------|
| Files Scanned | 15 |
| Issues Found | 23 |
| Confidence Score | 78 |

## Issues

### Critical

| File | Line | Issue |
|------|------|-------|
| src/auth.py | 42 | Possible hardcoded password |

Confidence Score

The confidence score (0-100) reflects the overall quality and security posture of the analyzed code:

90-100: Excellent - Code is well-structured and secure
70-89: Good - Minor issues found, generally safe to use
50-69: Moderate - Several issues, review recommended
30-49: Poor - Significant issues, refactoring advised
0-29: Critical - Major problems, do not deploy

Score Calculation

The score is calculated based on:

Security Issues (40% weight): Critical issues have highest impact
Code Smells (30% weight): Anti-patterns and poor practices
Complexity (20% weight): Function complexity and file size
Error Handling (10% weight): Missing exception handling

Supported Languages

Language	Security Scanning	Linting	Pattern Detection
Python	✅ Bandit	✅ Ruff	✅ Tree-sitter
JavaScript	⚠️ Basic	✅ Ruff	✅ Tree-sitter
TypeScript	⚠️ Basic	✅ Ruff	✅ Tree-sitter

Scanner Details

Bandit Scanner

Detects common security issues in Python code:

Hardcoded passwords and secrets
SQL injection vulnerabilities
Command injection risks
Insecure cryptographic usage
Path traversal issues

Ruff Scanner

Fast linting for code quality:

PEP 8 compliance
Import sorting
Code complexity
Anti-patterns
Unused code

Tree-sitter Scanner

Multi-language pattern detection:

API key and credential patterns
SQL injection patterns
Dangerous function usage
Deprecated API calls

API Usage

Python API

from src.core.scanner import Scanner
from src.reporting.confidence import ConfidenceScorer

# Initialize scanner
scanner = Scanner()

# Scan a directory
results = scanner.scan_directory("path/to/code")

# Calculate confidence score
scorer = ConfidenceScorer()
score = scorer.calculate(results)

# Print results
scanner.print_report(results)

Custom Configuration

from src.core.config import AuditConfig

config = AuditConfig(
    severity_filter={"critical", "high", "medium"},
    language_filter={"python"},
    scanners={"bandit", "ruff"}
)

scanner = Scanner(config=config)
results = scanner.scan("path/to/code")

Contributing

Fork the repository
Create a feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request

Development Setup

git clone https://7000pct.gitea.bloupla.net/7000pctAUTO/ai-code-audit-cli.git
cd ai-code-audit-cli
pip install -e ".[dev]"

# Run tests
pytest tests/ -v --cov=src

# Run linting
ruff check .

# Type checking
mypy src/

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

Bandit for Python security scanning
Ruff for fast linting
Tree-sitter for code parsing
Typer for CLI interface
Rich for beautiful terminal output