This commit is contained in:
overcuriousity
2025-09-10 15:17:17 +02:00
parent 696cec0723
commit ce0e11cf0b
16 changed files with 2577 additions and 484 deletions

View File

@@ -5,11 +5,17 @@ Contains implementations for various reconnaissance data sources.
from .base_provider import BaseProvider, RateLimiter
from .crtsh_provider import CrtShProvider
from .dns_provider import DNSProvider
from .shodan_provider import ShodanProvider
from .virustotal_provider import VirusTotalProvider
__all__ = [
'BaseProvider',
'RateLimiter',
'CrtShProvider'
'CrtShProvider',
'DNSProvider',
'ShodanProvider',
'VirusTotalProvider'
]
__version__ = "1.0.0-phase1"
__version__ = "1.0.0-phase2"

View File

@@ -113,86 +113,134 @@ class BaseProvider(ABC):
pass
def make_request(self, url: str, method: str = "GET",
params: Optional[Dict[str, Any]] = None,
headers: Optional[Dict[str, str]] = None,
target_indicator: str = "") -> Optional[requests.Response]:
"""
Make a rate-limited HTTP request with forensic logging.
params: Optional[Dict[str, Any]] = None,
headers: Optional[Dict[str, str]] = None,
target_indicator: str = "",
max_retries: int = 3) -> Optional[requests.Response]:
"""
Make a rate-limited HTTP request with forensic logging and retry logic.
Args:
url: Request URL
method: HTTP method
params: Query parameters
headers: Additional headers
target_indicator: The indicator being investigated
max_retries: Maximum number of retry attempts
Returns:
Response object or None if request failed
"""
for attempt in range(max_retries + 1):
# Apply rate limiting
self.rate_limiter.wait_if_needed()
start_time = time.time()
response = None
error = None
try:
self.total_requests += 1
# Prepare request
request_headers = self.session.headers.copy()
if headers:
request_headers.update(headers)
print(f"Making {method} request to: {url} (attempt {attempt + 1})")
# Make request
if method.upper() == "GET":
response = self.session.get(
url,
params=params,
headers=request_headers,
timeout=self.timeout
)
elif method.upper() == "POST":
response = self.session.post(
url,
json=params,
headers=request_headers,
timeout=self.timeout
)
else:
raise ValueError(f"Unsupported HTTP method: {method}")
print(f"Response status: {response.status_code}")
response.raise_for_status()
self.successful_requests += 1
# Success - log and return
duration_ms = (time.time() - start_time) * 1000
self.logger.log_api_request(
provider=self.name,
url=url,
method=method.upper(),
status_code=response.status_code,
response_size=len(response.content),
duration_ms=duration_ms,
error=None,
target_indicator=target_indicator
)
return response
except requests.exceptions.RequestException as e:
error = str(e)
self.failed_requests += 1
print(f"Request failed (attempt {attempt + 1}): {error}")
# Check if we should retry
if attempt < max_retries and self._should_retry(e):
backoff_time = (2 ** attempt) * 1 # Exponential backoff: 1s, 2s, 4s
print(f"Retrying in {backoff_time} seconds...")
time.sleep(backoff_time)
continue
else:
break
except Exception as e:
error = f"Unexpected error: {str(e)}"
self.failed_requests += 1
print(f"Unexpected error: {error}")
break
# All attempts failed - log and return None
duration_ms = (time.time() - start_time) * 1000
self.logger.log_api_request(
provider=self.name,
url=url,
method=method.upper(),
status_code=response.status_code if response else None,
response_size=len(response.content) if response else None,
duration_ms=duration_ms,
error=error,
target_indicator=target_indicator
)
return None
def _should_retry(self, exception: requests.exceptions.RequestException) -> bool:
"""
Determine if a request should be retried based on the exception.
Args:
url: Request URL
method: HTTP method
params: Query parameters
headers: Additional headers
target_indicator: The indicator being investigated
exception: The request exception that occurred
Returns:
Response object or None if request failed
True if the request should be retried
"""
# Apply rate limiting
self.rate_limiter.wait_if_needed()
start_time = time.time()
response = None
error = None
try:
self.total_requests += 1
# Prepare request
request_headers = self.session.headers.copy()
if headers:
request_headers.update(headers)
print(f"Making {method} request to: {url}")
# Make request
if method.upper() == "GET":
response = self.session.get(
url,
params=params,
headers=request_headers,
timeout=self.timeout
)
elif method.upper() == "POST":
response = self.session.post(
url,
json=params,
headers=request_headers,
timeout=self.timeout
)
else:
raise ValueError(f"Unsupported HTTP method: {method}")
print(f"Response status: {response.status_code}")
response.raise_for_status()
self.successful_requests += 1
except requests.exceptions.RequestException as e:
error = str(e)
self.failed_requests += 1
print(f"Request failed: {error}")
except Exception as e:
error = f"Unexpected error: {str(e)}"
self.failed_requests += 1
print(f"Unexpected error: {error}")
# Calculate duration and log request
duration_ms = (time.time() - start_time) * 1000
self.logger.log_api_request(
provider=self.name,
url=url,
method=method.upper(),
status_code=response.status_code if response else None,
response_size=len(response.content) if response else None,
duration_ms=duration_ms,
error=error,
target_indicator=target_indicator
)
return response if error is None else None
# Retry on connection errors, timeouts, and 5xx server errors
if isinstance(exception, (requests.exceptions.ConnectionError,
requests.exceptions.Timeout)):
return True
if isinstance(exception, requests.exceptions.HTTPError):
if hasattr(exception, 'response') and exception.response:
# Retry on server errors (5xx) but not client errors (4xx)
return exception.response.status_code >= 500
return False
def log_relationship_discovery(self, source_node: str, target_node: str,
relationship_type: RelationshipType,

View File

@@ -0,0 +1,338 @@
"""
DNS resolution provider for DNSRecon.
Discovers domain relationships through DNS record analysis.
"""
import socket
import dns.resolver
import dns.reversename
from typing import List, Dict, Any, Tuple, Optional
from .base_provider import BaseProvider
from core.graph_manager import RelationshipType, NodeType
class DNSProvider(BaseProvider):
"""
Provider for standard DNS resolution and reverse DNS lookups.
Discovers domain-to-IP and IP-to-domain relationships through DNS records.
"""
def __init__(self):
"""Initialize DNS provider with appropriate rate limiting."""
super().__init__(
name="dns",
rate_limit=100, # DNS queries can be faster
timeout=10
)
# Configure DNS resolver
self.resolver = dns.resolver.Resolver()
self.resolver.timeout = 5
self.resolver.lifetime = 10
def get_name(self) -> str:
"""Return the provider name."""
return "dns"
def is_available(self) -> bool:
"""DNS is always available - no API key required."""
return True
def query_domain(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""
Query DNS records for the domain to discover relationships.
Args:
domain: Domain to investigate
Returns:
List of relationships discovered from DNS analysis
"""
if not self._is_valid_domain(domain):
return []
relationships = []
# Query A records
relationships.extend(self._query_a_records(domain))
# Query AAAA records (IPv6)
relationships.extend(self._query_aaaa_records(domain))
# Query CNAME records
relationships.extend(self._query_cname_records(domain))
# Query MX records
relationships.extend(self._query_mx_records(domain))
# Query NS records
relationships.extend(self._query_ns_records(domain))
return relationships
def query_ip(self, ip: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""
Query reverse DNS for the IP address.
Args:
ip: IP address to investigate
Returns:
List of relationships discovered from reverse DNS
"""
if not self._is_valid_ip(ip):
return []
relationships = []
try:
# Perform reverse DNS lookup
reverse_name = dns.reversename.from_address(ip)
response = self.resolver.resolve(reverse_name, 'PTR')
for ptr_record in response:
hostname = str(ptr_record).rstrip('.')
if self._is_valid_domain(hostname):
raw_data = {
'query_type': 'PTR',
'ip_address': ip,
'hostname': hostname,
'ttl': response.ttl
}
relationships.append((
ip,
hostname,
RelationshipType.A_RECORD, # Reverse relationship
RelationshipType.A_RECORD.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=ip,
target_node=hostname,
relationship_type=RelationshipType.A_RECORD,
confidence_score=RelationshipType.A_RECORD.default_confidence,
raw_data=raw_data,
discovery_method="reverse_dns_lookup"
)
except Exception as e:
self.logger.logger.debug(f"Reverse DNS lookup failed for {ip}: {e}")
return relationships
def _query_a_records(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""Query A records for the domain."""
relationships = []
#if not DNS_AVAILABLE:
# return relationships
try:
response = self.resolver.resolve(domain, 'A')
for a_record in response:
ip_address = str(a_record)
raw_data = {
'query_type': 'A',
'domain': domain,
'ip_address': ip_address,
'ttl': response.ttl
}
relationships.append((
domain,
ip_address,
RelationshipType.A_RECORD,
RelationshipType.A_RECORD.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=domain,
target_node=ip_address,
relationship_type=RelationshipType.A_RECORD,
confidence_score=RelationshipType.A_RECORD.default_confidence,
raw_data=raw_data,
discovery_method="dns_a_record"
)
except Exception as e:
self.logger.logger.debug(f"A record query failed for {domain}: {e}")
return relationships
def _query_aaaa_records(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""Query AAAA records (IPv6) for the domain."""
relationships = []
#if not DNS_AVAILABLE:
# return relationships
try:
response = self.resolver.resolve(domain, 'AAAA')
for aaaa_record in response:
ip_address = str(aaaa_record)
raw_data = {
'query_type': 'AAAA',
'domain': domain,
'ip_address': ip_address,
'ttl': response.ttl
}
relationships.append((
domain,
ip_address,
RelationshipType.A_RECORD, # Using same type for IPv6
RelationshipType.A_RECORD.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=domain,
target_node=ip_address,
relationship_type=RelationshipType.A_RECORD,
confidence_score=RelationshipType.A_RECORD.default_confidence,
raw_data=raw_data,
discovery_method="dns_aaaa_record"
)
except Exception as e:
self.logger.logger.debug(f"AAAA record query failed for {domain}: {e}")
return relationships
def _query_cname_records(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""Query CNAME records for the domain."""
relationships = []
#if not DNS_AVAILABLE:
# return relationships
try:
response = self.resolver.resolve(domain, 'CNAME')
for cname_record in response:
target_domain = str(cname_record).rstrip('.')
if self._is_valid_domain(target_domain):
raw_data = {
'query_type': 'CNAME',
'source_domain': domain,
'target_domain': target_domain,
'ttl': response.ttl
}
relationships.append((
domain,
target_domain,
RelationshipType.CNAME_RECORD,
RelationshipType.CNAME_RECORD.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=domain,
target_node=target_domain,
relationship_type=RelationshipType.CNAME_RECORD,
confidence_score=RelationshipType.CNAME_RECORD.default_confidence,
raw_data=raw_data,
discovery_method="dns_cname_record"
)
except Exception as e:
self.logger.logger.debug(f"CNAME record query failed for {domain}: {e}")
return relationships
def _query_mx_records(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""Query MX records for the domain."""
relationships = []
#if not DNS_AVAILABLE:
# return relationships
try:
response = self.resolver.resolve(domain, 'MX')
for mx_record in response:
mx_host = str(mx_record.exchange).rstrip('.')
if self._is_valid_domain(mx_host):
raw_data = {
'query_type': 'MX',
'domain': domain,
'mx_host': mx_host,
'priority': mx_record.preference,
'ttl': response.ttl
}
relationships.append((
domain,
mx_host,
RelationshipType.MX_RECORD,
RelationshipType.MX_RECORD.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=domain,
target_node=mx_host,
relationship_type=RelationshipType.MX_RECORD,
confidence_score=RelationshipType.MX_RECORD.default_confidence,
raw_data=raw_data,
discovery_method="dns_mx_record"
)
except Exception as e:
self.logger.logger.debug(f"MX record query failed for {domain}: {e}")
return relationships
def _query_ns_records(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""Query NS records for the domain."""
relationships = []
#if not DNS_AVAILABLE:
# return relationships
try:
response = self.resolver.resolve(domain, 'NS')
for ns_record in response:
ns_host = str(ns_record).rstrip('.')
if self._is_valid_domain(ns_host):
raw_data = {
'query_type': 'NS',
'domain': domain,
'ns_host': ns_host,
'ttl': response.ttl
}
relationships.append((
domain,
ns_host,
RelationshipType.NS_RECORD,
RelationshipType.NS_RECORD.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=domain,
target_node=ns_host,
relationship_type=RelationshipType.NS_RECORD,
confidence_score=RelationshipType.NS_RECORD.default_confidence,
raw_data=raw_data,
discovery_method="dns_ns_record"
)
except Exception as e:
self.logger.logger.debug(f"NS record query failed for {domain}: {e}")
return relationships

View File

@@ -0,0 +1,299 @@
"""
Shodan provider for DNSRecon.
Discovers IP relationships and infrastructure context through Shodan API.
"""
import json
from typing import List, Dict, Any, Tuple, Optional
from urllib.parse import quote
from .base_provider import BaseProvider
from core.graph_manager import RelationshipType
from config import config
class ShodanProvider(BaseProvider):
"""
Provider for querying Shodan API for IP address and hostname information.
Requires valid API key and respects Shodan's rate limits.
"""
def __init__(self):
"""Initialize Shodan provider with appropriate rate limiting."""
super().__init__(
name="shodan",
rate_limit=60, # Shodan API has various rate limits depending on plan
timeout=30
)
self.base_url = "https://api.shodan.io"
self.api_key = config.get_api_key('shodan')
def get_name(self) -> str:
"""Return the provider name."""
return "shodan"
def is_available(self) -> bool:
"""
Check if Shodan provider is available (has valid API key).
"""
return self.api_key is not None and len(self.api_key.strip()) > 0
def query_domain(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""
Query Shodan for information about a domain.
Uses Shodan's hostname search to find associated IPs.
Args:
domain: Domain to investigate
Returns:
List of relationships discovered from Shodan data
"""
if not self._is_valid_domain(domain) or not self.is_available():
return []
relationships = []
try:
# Search for hostname in Shodan
search_query = f"hostname:{domain}"
url = f"{self.base_url}/shodan/host/search"
params = {
'key': self.api_key,
'query': search_query,
'minify': True # Get minimal data to reduce bandwidth
}
response = self.make_request(url, method="GET", params=params, target_indicator=domain)
if not response or response.status_code != 200:
return []
data = response.json()
if 'matches' not in data:
return []
# Process search results
for match in data['matches']:
ip_address = match.get('ip_str')
hostnames = match.get('hostnames', [])
if ip_address and domain in hostnames:
raw_data = {
'ip_address': ip_address,
'hostnames': hostnames,
'country': match.get('location', {}).get('country_name', ''),
'city': match.get('location', {}).get('city', ''),
'isp': match.get('isp', ''),
'org': match.get('org', ''),
'ports': match.get('ports', []),
'last_update': match.get('last_update', '')
}
relationships.append((
domain,
ip_address,
RelationshipType.A_RECORD, # Domain resolves to IP
RelationshipType.A_RECORD.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=domain,
target_node=ip_address,
relationship_type=RelationshipType.A_RECORD,
confidence_score=RelationshipType.A_RECORD.default_confidence,
raw_data=raw_data,
discovery_method="shodan_hostname_search"
)
# Also create relationships to other hostnames on the same IP
for hostname in hostnames:
if hostname != domain and self._is_valid_domain(hostname):
hostname_raw_data = {
'shared_ip': ip_address,
'all_hostnames': hostnames,
'discovery_context': 'shared_hosting'
}
relationships.append((
domain,
hostname,
RelationshipType.PASSIVE_DNS, # Shared hosting relationship
0.6, # Lower confidence for shared hosting
hostname_raw_data
))
self.log_relationship_discovery(
source_node=domain,
target_node=hostname,
relationship_type=RelationshipType.PASSIVE_DNS,
confidence_score=0.6,
raw_data=hostname_raw_data,
discovery_method="shodan_shared_hosting"
)
except json.JSONDecodeError as e:
self.logger.logger.error(f"Failed to parse JSON response from Shodan: {e}")
except Exception as e:
self.logger.logger.error(f"Error querying Shodan for domain {domain}: {e}")
return relationships
def query_ip(self, ip: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""
Query Shodan for information about an IP address.
Args:
ip: IP address to investigate
Returns:
List of relationships discovered from Shodan IP data
"""
if not self._is_valid_ip(ip) or not self.is_available():
return []
relationships = []
try:
# Query Shodan host information
url = f"{self.base_url}/shodan/host/{ip}"
params = {'key': self.api_key}
response = self.make_request(url, method="GET", params=params, target_indicator=ip)
if not response or response.status_code != 200:
return []
data = response.json()
# Extract hostname relationships
hostnames = data.get('hostnames', [])
for hostname in hostnames:
if self._is_valid_domain(hostname):
raw_data = {
'ip_address': ip,
'hostname': hostname,
'country': data.get('country_name', ''),
'city': data.get('city', ''),
'isp': data.get('isp', ''),
'org': data.get('org', ''),
'asn': data.get('asn', ''),
'ports': data.get('ports', []),
'last_update': data.get('last_update', ''),
'os': data.get('os', '')
}
relationships.append((
ip,
hostname,
RelationshipType.A_RECORD, # IP resolves to hostname
RelationshipType.A_RECORD.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=ip,
target_node=hostname,
relationship_type=RelationshipType.A_RECORD,
confidence_score=RelationshipType.A_RECORD.default_confidence,
raw_data=raw_data,
discovery_method="shodan_host_lookup"
)
# Extract ASN relationship if available
asn = data.get('asn')
if asn:
asn_name = f"AS{asn}"
asn_raw_data = {
'ip_address': ip,
'asn': asn,
'isp': data.get('isp', ''),
'org': data.get('org', '')
}
relationships.append((
ip,
asn_name,
RelationshipType.ASN_MEMBERSHIP,
RelationshipType.ASN_MEMBERSHIP.default_confidence,
asn_raw_data
))
self.log_relationship_discovery(
source_node=ip,
target_node=asn_name,
relationship_type=RelationshipType.ASN_MEMBERSHIP,
confidence_score=RelationshipType.ASN_MEMBERSHIP.default_confidence,
raw_data=asn_raw_data,
discovery_method="shodan_asn_lookup"
)
except json.JSONDecodeError as e:
self.logger.logger.error(f"Failed to parse JSON response from Shodan: {e}")
except Exception as e:
self.logger.logger.error(f"Error querying Shodan for IP {ip}: {e}")
return relationships
def search_by_organization(self, org_name: str) -> List[Dict[str, Any]]:
"""
Search Shodan for hosts belonging to a specific organization.
Args:
org_name: Organization name to search for
Returns:
List of host information dictionaries
"""
if not self.is_available():
return []
try:
search_query = f"org:\"{org_name}\""
url = f"{self.base_url}/shodan/host/search"
params = {
'key': self.api_key,
'query': search_query,
'minify': True
}
response = self.make_request(url, method="GET", params=params, target_indicator=org_name)
if response and response.status_code == 200:
data = response.json()
return data.get('matches', [])
except Exception as e:
self.logger.logger.error(f"Error searching Shodan by organization {org_name}: {e}")
return []
def get_host_services(self, ip: str) -> List[Dict[str, Any]]:
"""
Get service information for a specific IP address.
Args:
ip: IP address to query
Returns:
List of service information dictionaries
"""
if not self._is_valid_ip(ip) or not self.is_available():
return []
try:
url = f"{self.base_url}/shodan/host/{ip}"
params = {'key': self.api_key}
response = self.make_request(url, method="GET", params=params, target_indicator=ip)
if response and response.status_code == 200:
data = response.json()
return data.get('data', []) # Service banners
except Exception as e:
self.logger.logger.error(f"Error getting Shodan services for IP {ip}: {e}")
return []

View File

@@ -0,0 +1,334 @@
"""
VirusTotal provider for DNSRecon.
Discovers domain relationships through passive DNS and URL analysis.
"""
import json
from typing import List, Dict, Any, Tuple, Optional
from .base_provider import BaseProvider
from core.graph_manager import RelationshipType
from config import config
class VirusTotalProvider(BaseProvider):
"""
Provider for querying VirusTotal API for passive DNS and domain reputation data.
Requires valid API key and strictly respects free tier rate limits.
"""
def __init__(self):
"""Initialize VirusTotal provider with strict rate limiting for free tier."""
super().__init__(
name="virustotal",
rate_limit=4, # Free tier: 4 requests per minute
timeout=30
)
self.base_url = "https://www.virustotal.com/vtapi/v2"
self.api_key = config.get_api_key('virustotal')
def get_name(self) -> str:
"""Return the provider name."""
return "virustotal"
def is_available(self) -> bool:
"""
Check if VirusTotal provider is available (has valid API key).
"""
return self.api_key is not None and len(self.api_key.strip()) > 0
def query_domain(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""
Query VirusTotal for domain information including passive DNS.
Args:
domain: Domain to investigate
Returns:
List of relationships discovered from VirusTotal data
"""
if not self._is_valid_domain(domain) or not self.is_available():
return []
relationships = []
# Query domain report
domain_relationships = self._query_domain_report(domain)
relationships.extend(domain_relationships)
# Query passive DNS for the domain
passive_dns_relationships = self._query_passive_dns_domain(domain)
relationships.extend(passive_dns_relationships)
return relationships
def query_ip(self, ip: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""
Query VirusTotal for IP address information including passive DNS.
Args:
ip: IP address to investigate
Returns:
List of relationships discovered from VirusTotal IP data
"""
if not self._is_valid_ip(ip) or not self.is_available():
return []
relationships = []
# Query IP report
ip_relationships = self._query_ip_report(ip)
relationships.extend(ip_relationships)
# Query passive DNS for the IP
passive_dns_relationships = self._query_passive_dns_ip(ip)
relationships.extend(passive_dns_relationships)
return relationships
def _query_domain_report(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""Query VirusTotal domain report."""
relationships = []
try:
url = f"{self.base_url}/domain/report"
params = {
'apikey': self.api_key,
'domain': domain,
'allinfo': 1 # Get comprehensive information
}
response = self.make_request(url, method="GET", params=params, target_indicator=domain)
if not response or response.status_code != 200:
return []
data = response.json()
if data.get('response_code') != 1:
return []
# Extract resolved IPs
resolutions = data.get('resolutions', [])
for resolution in resolutions:
ip_address = resolution.get('ip_address')
last_resolved = resolution.get('last_resolved')
if ip_address and self._is_valid_ip(ip_address):
raw_data = {
'domain': domain,
'ip_address': ip_address,
'last_resolved': last_resolved,
'source': 'virustotal_domain_report'
}
relationships.append((
domain,
ip_address,
RelationshipType.PASSIVE_DNS,
RelationshipType.PASSIVE_DNS.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=domain,
target_node=ip_address,
relationship_type=RelationshipType.PASSIVE_DNS,
confidence_score=RelationshipType.PASSIVE_DNS.default_confidence,
raw_data=raw_data,
discovery_method="virustotal_domain_resolution"
)
# Extract subdomains
subdomains = data.get('subdomains', [])
for subdomain in subdomains:
if subdomain != domain and self._is_valid_domain(subdomain):
raw_data = {
'parent_domain': domain,
'subdomain': subdomain,
'source': 'virustotal_subdomain_discovery'
}
relationships.append((
domain,
subdomain,
RelationshipType.PASSIVE_DNS,
0.7, # Medium-high confidence for subdomains
raw_data
))
self.log_relationship_discovery(
source_node=domain,
target_node=subdomain,
relationship_type=RelationshipType.PASSIVE_DNS,
confidence_score=0.7,
raw_data=raw_data,
discovery_method="virustotal_subdomain_discovery"
)
except json.JSONDecodeError as e:
self.logger.logger.error(f"Failed to parse JSON response from VirusTotal: {e}")
except Exception as e:
self.logger.logger.error(f"Error querying VirusTotal domain report for {domain}: {e}")
return relationships
def _query_ip_report(self, ip: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""Query VirusTotal IP report."""
relationships = []
try:
url = f"{self.base_url}/ip-address/report"
params = {
'apikey': self.api_key,
'ip': ip
}
response = self.make_request(url, method="GET", params=params, target_indicator=ip)
if not response or response.status_code != 200:
return []
data = response.json()
if data.get('response_code') != 1:
return []
# Extract resolved domains
resolutions = data.get('resolutions', [])
for resolution in resolutions:
hostname = resolution.get('hostname')
last_resolved = resolution.get('last_resolved')
if hostname and self._is_valid_domain(hostname):
raw_data = {
'ip_address': ip,
'hostname': hostname,
'last_resolved': last_resolved,
'source': 'virustotal_ip_report'
}
relationships.append((
ip,
hostname,
RelationshipType.PASSIVE_DNS,
RelationshipType.PASSIVE_DNS.default_confidence,
raw_data
))
self.log_relationship_discovery(
source_node=ip,
target_node=hostname,
relationship_type=RelationshipType.PASSIVE_DNS,
confidence_score=RelationshipType.PASSIVE_DNS.default_confidence,
raw_data=raw_data,
discovery_method="virustotal_ip_resolution"
)
except json.JSONDecodeError as e:
self.logger.logger.error(f"Failed to parse JSON response from VirusTotal: {e}")
except Exception as e:
self.logger.logger.error(f"Error querying VirusTotal IP report for {ip}: {e}")
return relationships
def _query_passive_dns_domain(self, domain: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""Query VirusTotal passive DNS for domain."""
# Note: VirusTotal's passive DNS API might require a premium subscription
# This is a placeholder for the endpoint structure
return []
def _query_passive_dns_ip(self, ip: str) -> List[Tuple[str, str, RelationshipType, float, Dict[str, Any]]]:
"""Query VirusTotal passive DNS for IP."""
# Note: VirusTotal's passive DNS API might require a premium subscription
# This is a placeholder for the endpoint structure
return []
def get_domain_reputation(self, domain: str) -> Dict[str, Any]:
"""
Get domain reputation information from VirusTotal.
Args:
domain: Domain to check reputation for
Returns:
Dictionary containing reputation data
"""
if not self._is_valid_domain(domain) or not self.is_available():
return {}
try:
url = f"{self.base_url}/domain/report"
params = {
'apikey': self.api_key,
'domain': domain
}
response = self.make_request(url, method="GET", params=params, target_indicator=domain)
if response and response.status_code == 200:
data = response.json()
if data.get('response_code') == 1:
return {
'positives': data.get('positives', 0),
'total': data.get('total', 0),
'scan_date': data.get('scan_date', ''),
'permalink': data.get('permalink', ''),
'reputation_score': self._calculate_reputation_score(data)
}
except Exception as e:
self.logger.logger.error(f"Error getting VirusTotal reputation for domain {domain}: {e}")
return {}
def get_ip_reputation(self, ip: str) -> Dict[str, Any]:
"""
Get IP reputation information from VirusTotal.
Args:
ip: IP address to check reputation for
Returns:
Dictionary containing reputation data
"""
if not self._is_valid_ip(ip) or not self.is_available():
return {}
try:
url = f"{self.base_url}/ip-address/report"
params = {
'apikey': self.api_key,
'ip': ip
}
response = self.make_request(url, method="GET", params=params, target_indicator=ip)
if response and response.status_code == 200:
data = response.json()
if data.get('response_code') == 1:
return {
'positives': data.get('positives', 0),
'total': data.get('total', 0),
'scan_date': data.get('scan_date', ''),
'permalink': data.get('permalink', ''),
'reputation_score': self._calculate_reputation_score(data)
}
except Exception as e:
self.logger.logger.error(f"Error getting VirusTotal reputation for IP {ip}: {e}")
return {}
def _calculate_reputation_score(self, data: Dict[str, Any]) -> float:
"""Calculate a normalized reputation score (0.0 to 1.0)."""
positives = data.get('positives', 0)
total = data.get('total', 1) # Avoid division by zero
if total == 0:
return 1.0 # No data means neutral
# Score is inverse of detection ratio (lower detection = higher reputation)
return max(0.0, 1.0 - (positives / total))