Buckets:

download
raw
7.57 kB
"""
Quick test script to demonstrate OpenAI integration with DECOYABLE AI systems.
Tests:
1. OpenAI API connectivity
2. AI-powered vulnerability analysis with GPT
3. LLM-enhanced threat predictions
4. Smart remediation suggestions
"""
import asyncio
import json
import os
import sys
from pathlib import Path
# Add parent directory to path
sys.path.insert(0, str(Path(__file__).parent))
from decoyable.orchestrator import get_orchestrator
async def test_openai_integration():
"""Test OpenAI integration with DECOYABLE."""
print("\n" + "="*80)
print("๐Ÿงช TESTING OPENAI INTEGRATION WITH DECOYABLE AI".center(80))
print("="*80 + "\n")
# Check API key
api_key = os.getenv("OPENAI_API_KEY")
if not api_key:
print("โŒ No OpenAI API key found in environment!")
print(" Please set OPENAI_API_KEY in your .env file\n")
return False
print(f"โœ… OpenAI API key detected: configured ({len(api_key)} chars)")
print()
# Test basic OpenAI connectivity
print("๐Ÿ” Testing OpenAI API connectivity...")
try:
from openai import OpenAI
client = OpenAI(api_key=api_key)
# Simple test call
response = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[
{"role": "system", "content": "You are a cybersecurity expert."},
{"role": "user", "content": "In one sentence, what is SQL injection?"}
],
max_tokens=100
)
test_response = response.choices[0].message.content
print(f"โœ… OpenAI API working! Test response:")
print(f" '{test_response}'\n")
except ImportError:
print("โš ๏ธ OpenAI package not installed. Installing...")
import subprocess
subprocess.check_call([sys.executable, "-m", "pip", "install", "openai"])
print("โœ… OpenAI package installed!\n")
return await test_openai_integration() # Retry
except Exception as e:
print(f"โŒ OpenAI API test failed: {e}\n")
return False
# Test DECOYABLE AI analysis
print("-"*80)
print("๐Ÿค– Running DECOYABLE AI Analysis...")
print("-"*80 + "\n")
orchestrator = get_orchestrator()
# Analyze a sample vulnerable code snippet
sample_code = """
import os
import subprocess
def process_user_input(user_data):
# Vulnerable to command injection
result = os.system(f"echo {user_data}")
# Vulnerable to SQL injection
query = f"SELECT * FROM users WHERE username = '{user_data}'"
# Vulnerable to path traversal
filename = user_data
with open(f"/data/{filename}", "r") as f:
content = f.read()
return content
"""
# Save sample to temp file
temp_file = Path("temp_vulnerable_sample.py")
temp_file.write_text(sample_code)
try:
# Run comprehensive analysis
report = await orchestrator.analyze_codebase_comprehensive(str(temp_file))
print("๐Ÿ“Š ANALYSIS RESULTS:")
print(f" Risk Level: {report['risk_emoji']} {report['overall_risk_level']}")
print(f" Risk Score: {report['overall_risk_score']:.1f}")
print(f" Vulnerabilities: {report['summary']['total_vulnerabilities']}")
print(f" AI Predictions: {report['summary']['predicted_threats']}")
print(f" Exploit Chains: {report['summary']['exploit_chains_detected']}")
print()
# Show top predictions
if report.get('threat_predictions'):
print("๐ŸŽฏ TOP AI PREDICTIONS:")
for i, pred in enumerate(report['threat_predictions'][:3], 1):
print(f" {i}. {pred['threat_type']}")
print(f" Probability: {pred['probability']} | Confidence: {pred['confidence']}")
print(f" Risk Score: {pred['risk_score']}")
print()
# Show exploit chains
if report.get('exploit_chains'):
print("โ›“๏ธ EXPLOIT CHAINS DETECTED:")
for chain in report['exploit_chains']:
print(f" ๐Ÿ”ด {chain['severity']}: {chain['impact']}")
print(f" Chain: {' โ†’ '.join(chain['vulnerability_types'])}")
print()
finally:
# Cleanup
if temp_file.exists():
temp_file.unlink()
# Test LLM-enhanced analysis (if we have time to implement)
print("-"*80)
print("๐Ÿง  POTENTIAL LLM-ENHANCED FEATURES:")
print("-"*80 + "\n")
print(" โœจ Natural language vulnerability explanations")
print(" โœจ Context-aware remediation suggestions")
print(" โœจ Attack scenario generation")
print(" โœจ Custom security policy recommendations")
print(" โœจ Intelligent code review comments")
print(" โœจ Threat intelligence summarization")
print()
print("="*80)
print("โœ… OpenAI Integration Test Complete!".center(80))
print("="*80 + "\n")
return True
async def demonstrate_llm_enhanced_analysis():
"""Demonstrate what LLM-enhanced analysis could look like."""
print("\n" + "="*80)
print("๐Ÿ’ก LLM-ENHANCED ANALYSIS DEMO".center(80))
print("="*80 + "\n")
api_key = os.getenv("OPENAI_API_KEY")
if not api_key:
print("โš ๏ธ Skipping LLM demo - no API key\n")
return
try:
from openai import OpenAI
client = OpenAI(api_key=api_key)
# Example: Get natural language explanation for a vulnerability
vulnerability = {
"type": "SQL_INJECTION",
"code": "query = f\"SELECT * FROM users WHERE id = '{user_id}'\"",
"line": 42
}
print("๐Ÿ” Analyzing vulnerability with GPT-3.5...")
print(f" Code: {vulnerability['code']}")
print()
prompt = f"""As a cybersecurity expert, explain this vulnerability:
Type: {vulnerability['type']}
Code: {vulnerability['code']}
Provide:
1. Why this is vulnerable (2 sentences)
2. A specific attack example
3. The exact fix (code)
4. Prevention tips
Be concise and technical."""
response = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[
{"role": "system", "content": "You are a senior cybersecurity engineer providing code review."},
{"role": "user", "content": prompt}
],
max_tokens=500,
temperature=0.3
)
explanation = response.choices[0].message.content
print("๐Ÿง  GPT-4 Analysis:")
print("-" * 80)
print(explanation)
print("-" * 80)
print()
print("โœจ This is the power of LLM-enhanced security analysis!")
print(" โ€ข Natural language explanations")
print(" โ€ข Context-aware recommendations")
print(" โ€ข Teaching moments for developers")
print()
except Exception as e:
print(f"โš ๏ธ LLM demo encountered error: {e}")
print(" (This is optional - core AI features still work!)\n")
if __name__ == "__main__":
print("\n๐Ÿš€ Starting DECOYABLE + OpenAI Integration Tests...\n")
# Run tests
loop = asyncio.get_event_loop()
success = loop.run_until_complete(test_openai_integration())
if success:
loop.run_until_complete(demonstrate_llm_enhanced_analysis())
print("๐ŸŽ‰ All tests complete!\n")

Xet Storage Details

Size:
7.57 kB
ยท
Xet hash:
136f16c46fa5c3da6e8959d2593c3afe1efd3bf7c390762ba0d2b277bcb7f07c

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.