Buckets:

download
raw
5.85 kB
"""
Test script for AI integration with all providers.
NOTE: This file contains intentional use of dangerous patterns (os.system, subprocess with shell=True, etc.)
for the purpose of testing, demonstration, and validation of security scanners. These are NOT used in production code.
Tests: Ollama (Gemma 3:12b), OpenAI, Phi-3, Pattern-based
"""
import asyncio
import logging
from dotenv import load_dotenv
# Load environment variables
load_dotenv()
# Setup logging
logging.basicConfig(
level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
from decoyable.llm.model_router import get_router
# Test vulnerable code sample
VULNERABLE_CODE = '''
import os
import sqlite3
def get_user_data(user_id):
# SQL Injection vulnerability
query = f"SELECT * FROM users WHERE id = {user_id}"
conn = sqlite3.connect('database.db')
cursor = conn.cursor()
cursor.execute(query)
return cursor.fetchall()
def run_command(cmd):
# Command injection vulnerability
os.system(f"ping -c 1 {cmd}")
def get_api_key():
# Hardcoded secret
api_key = "sk-1234567890abcdef"
return api_key
'''
async def test_ai_providers():
"""Test all AI providers."""
print("\n" + "="*60)
print("๐Ÿค– TESTING DECOYABLE AI INTEGRATION")
print("="*60 + "\n")
# Get router instance
router = get_router()
# Print status
print("๐Ÿ“Š AVAILABLE PROVIDERS:")
router.print_status()
print("\n" + "="*60)
print("๐Ÿงช TESTING AI ANALYSIS ON VULNERABLE CODE")
print("="*60 + "\n")
print("๐Ÿ“ Test Code:")
print("-" * 40)
print(VULNERABLE_CODE[:200] + "...")
print("-" * 40 + "\n")
# Test analysis
print("๐Ÿ” Running AI analysis...")
try:
result = await router.analyze(
code=VULNERABLE_CODE,
context={
"file": "test.py",
"framework": "flask"
},
task="security_analysis"
)
if result:
print("\nโœ… AI ANALYSIS COMPLETE!")
print("="*60)
# Show provider used
provider = result.get("_provider", "unknown")
cost = result.get("_cost", 0.0)
local = result.get("_local", False)
print(f"\n๐Ÿค– Provider Used: {provider.upper()}")
print(f"๐Ÿ’ฐ Cost: ${cost:.4f}")
print(f"๐Ÿ”’ Local: {'Yes' if local else 'No'}")
# Show vulnerabilities found
vulns = result.get("vulnerabilities", [])
print(f"\n๐ŸŽฏ Vulnerabilities Found: {len(vulns)}")
for i, vuln in enumerate(vulns, 1):
print(f"\n {i}. {vuln.get('type', 'Unknown')}")
print(f" Severity: {vuln.get('severity', 'Unknown')}")
print(f" Confidence: {vuln.get('confidence', 0)*100:.0f}%")
desc = vuln.get('description', 'No description')
print(f" Description: {desc[:100]}...")
# Show risk score
risk = result.get("risk_score", 0)
print(f"\n๐Ÿ“Š Overall Risk Score: {risk}/100")
# Show summary
summary = result.get("summary", "No summary available")
print(f"\n๐Ÿ“ Summary:")
print(f" {summary[:200]}...")
# Show recommendations
recs = result.get("recommendations", [])
if recs:
print(f"\n๐Ÿ’ก Recommendations:")
for i, rec in enumerate(recs[:3], 1):
print(f" {i}. {rec[:100]}...")
print("\n" + "="*60)
print("โœ… TEST SUCCESSFUL!")
print("="*60)
else:
print("\nโŒ Analysis returned no results")
except Exception as e:
print(f"\nโŒ Error during analysis: {e}")
import traceback
traceback.print_exc()
async def test_ollama_direct():
"""Test Ollama directly."""
print("\n" + "="*60)
print("๐Ÿฆ™ TESTING OLLAMA + GEMMA 3:12B DIRECTLY")
print("="*60 + "\n")
from decoyable.llm.ollama_client import OllamaClient
client = OllamaClient()
print(f"Model: {client.model}")
print(f"Available: {client.available}")
print(f"Base URL: {client.base_url}\n")
if client.available:
print("โœ… Ollama is running!")
# Get available models
models = client.get_available_models()
print(f"\n๐Ÿ“ฆ Available Models ({len(models) if isinstance(models, list) else 0}):")
if isinstance(models, list):
for model in models:
if isinstance(model, dict):
name = model.get('name', 'Unknown')
size = model.get('size', 0) / (1024**3) # Convert to GB
params = model.get('details', {}).get('parameter_size', 'Unknown')
print(f" - {name} ({size:.1f}GB, {params})")
else:
print(f" - {model}")
print("\n๐Ÿงช Testing quick analysis...")
result = await client.analyze_security(
code="password = '123456' # Hardcoded password",
task="security_analysis"
)
if result:
print("โœ… Analysis successful!")
print(f"Found {len(result.get('vulnerabilities', []))} issues")
else:
print("โŒ Analysis failed")
else:
print("โŒ Ollama is not available")
if __name__ == "__main__":
print("\n๐Ÿš€ DECOYABLE v1.2.0 - AI Integration Test\n")
# Run tests
asyncio.run(test_ollama_direct())
asyncio.run(test_ai_providers())
print("\nโœ… All tests complete!\n")

Xet Storage Details

Size:
5.85 kB
ยท
Xet hash:
1aa418a00604bd75909b5e460683663ddd93dffeb3afafaa66c5ddf1906ea44e

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