How It Works

Learn about the scanning process, vulnerability assessment, and AI-powered fixes.

Scanning Process

Coach uses a multi-step process to analyze repositories for security vulnerabilities:

Repository Cloning

Coach clones the target repository to analyze its contents safely.

Code Parsing

The code is parsed to identify language-specific patterns and structures.

Vulnerability Detection

Multiple detection algorithms are applied to identify potential vulnerabilities.

Vulnerability Assessment

Coach categorizes vulnerabilities by severity level to help you prioritize remediation efforts:

Critical

Severe vulnerabilities that require immediate attention

High

Important issues that should be fixed soon

Medium

Moderate risk issues that should be addressed

Low

Minor issues that pose limited risk

Assessment Criteria

Exploit Potential: How easily can the vulnerability be exploited?
Attack Surface: How exposed is the vulnerability to potential attackers?
Impact: What damage could result from exploiting this vulnerability?
Prevalence: How common is this vulnerability in the codebase?

AI-Powered Fixes

Coach uses the Gemini API to generate intelligent fixes for detected vulnerabilities:

Vulnerable Code

Python

def process_user_input(user_input):
    # Vulnerable code with SQL injection risk
    query = f"SELECT * FROM users WHERE username = '{user_input}'"
    return db.execute(query)

AI-Generated Fix

Python

def process_user_input(user_input):
    # Using parameterized query to prevent SQL injection
    query = "SELECT * FROM users WHERE username = %s"
    return db.execute(query, (user_input,))

Explanation

The AI detects the SQL injection vulnerability and generates a fix that uses parameterized queries to safely handle user input without risking injection attacks.