Code Engine

Security analysis of code using AST parsing and pattern detection.

Overview

The Code Engine uses static analysis to detect security vulnerabilities:

• Dangerous function calls
• Code injection risks
• Unsafe imports
• Security anti-patterns

Supported Languages

Language	Detection
Python	Full AST + patterns
JavaScript	Pattern matching
Shell	Pattern matching

Usage

code = """
import os
os.system('rm -rf /')
"""

result = client.verify_code(code, language="python")
print(result.verified)  # False
print(result.status)    # "BLOCKED"

for vuln in result.vulnerabilities:
    print(f"{vuln.severity}: {vuln.message}")
    # CRITICAL: Shell command execution (os.system)

Detection Categories

Critical

Pattern	Risk
eval()	Code execution
exec()	Code execution
os.system()	Shell execution
subprocess.*	Process spawning

High

Pattern	Risk
pickle.load()	Deserialization
__import__()	Dynamic import
open(..., 'w')	File overwrite

Medium

Pattern	Risk
requests.get()	External request
sqlite3.connect()	Database access

Response Format

{
    "status": "BLOCKED",
    "verified": False,
    "vulnerabilities": [
        {
            "type": "os.system",
            "severity": "critical",
            "line": 2,
            "message": "Shell command execution"
        }
    ]
}

Safe Code Example

good_code = """
def calculate_sum(a, b):
    return a + b

result = calculate_sum(5, 10)
print(f"Sum: {result}")
"""

result = client.verify_code(good_code)
print(result.verified)  # True
print(result.status)    # "VERIFIED"