Skip to content

Commit

Permalink
Merge branch 'dev' into extract_internal_module
Browse files Browse the repository at this point in the history
  • Loading branch information
domwhewell-sage authored Nov 12, 2024
2 parents 95b4cbb + 5de9452 commit 363cd6d
Show file tree
Hide file tree
Showing 33 changed files with 1,108 additions and 706 deletions.
2 changes: 1 addition & 1 deletion .github/workflows/distro_tests.yml
Original file line number Diff line number Diff line change
Expand Up @@ -61,4 +61,4 @@ jobs:
export BBOT_DISTRO_TESTS=true
poetry env use python3.11
poetry install
poetry run pytest --reruns 2 -o timeout_func_only=true --timeout 1200 --disable-warnings --log-cli-level=INFO .
poetry run pytest --reruns 2 -o timeout_func_only=true --timeout 1200 --disable-warnings --log-cli-level=INFO -k gowitness .
2 changes: 1 addition & 1 deletion .github/workflows/tests.yml
Original file line number Diff line number Diff line change
Expand Up @@ -48,7 +48,7 @@ jobs:
poetry install
- name: Run tests
run: |
poetry run pytest --exitfirst --reruns 2 -o timeout_func_only=true --timeout 1200 --disable-warnings --log-cli-level=INFO --cov-config=bbot/test/coverage.cfg --cov-report xml:cov.xml --cov=bbot .
poetry run pytest --exitfirst --reruns 2 -o timeout_func_only=true --timeout 1200 --disable-warnings --log-cli-level=INFO --cov-config=bbot/test/coverage.cfg --cov-report xml:cov.xml --cov=bbot -k gowitness .
- name: Upload Debug Logs
uses: actions/upload-artifact@v3
with:
Expand Down
1 change: 1 addition & 0 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -357,6 +357,7 @@ For details, see [Configuration](https://www.blacklanternsecurity.com/bbot/Stabl
- **Modules**
- [List of Modules](https://www.blacklanternsecurity.com/bbot/Stable/modules/list_of_modules)
- [Nuclei](https://www.blacklanternsecurity.com/bbot/Stable/modules/nuclei)
- [Custom YARA Rules](https://www.blacklanternsecurity.com/bbot/Stable/modules/custom_yara_rules)
- **Misc**
- [Contribution](https://www.blacklanternsecurity.com/bbot/Stable/contribution)
- [Release History](https://www.blacklanternsecurity.com/bbot/Stable/release_history)
Expand Down
42 changes: 30 additions & 12 deletions bbot/core/event/base.py
Original file line number Diff line number Diff line change
Expand Up @@ -503,12 +503,13 @@ def scope_distance(self, scope_distance):
for t in list(self.tags):
if t.startswith("distance-"):
self.remove_tag(t)
if scope_distance == 0:
self.add_tag("in-scope")
self.remove_tag("affiliate")
else:
self.remove_tag("in-scope")
self.add_tag(f"distance-{new_scope_distance}")
if self.host:
if scope_distance == 0:
self.add_tag("in-scope")
self.remove_tag("affiliate")
else:
self.remove_tag("in-scope")
self.add_tag(f"distance-{new_scope_distance}")
self._scope_distance = new_scope_distance
# apply recursively to parent events
parent_scope_distance = getattr(self.parent, "scope_distance", None)
Expand Down Expand Up @@ -1018,20 +1019,21 @@ def __init__(self, *args, **kwargs):
class DictPathEvent(DictEvent):
def sanitize_data(self, data):
new_data = dict(data)
new_data["path"] = str(new_data["path"])
file_blobs = getattr(self.scan, "_file_blobs", False)
folder_blobs = getattr(self.scan, "_folder_blobs", False)
blob = None
try:
data_path = Path(data["path"])
if data_path.is_file():
self._data_path = Path(data["path"])
if self._data_path.is_file():
self.add_tag("file")
if file_blobs:
with open(data_path, "rb") as file:
with open(self._data_path, "rb") as file:
blob = file.read()
elif data_path.is_dir():
elif self._data_path.is_dir():
self.add_tag("folder")
if folder_blobs:
blob = self._tar_directory(data_path)
blob = self._tar_directory(self._data_path)
except KeyError:
pass
if blob:
Expand Down Expand Up @@ -1540,7 +1542,23 @@ def _pretty_string(self):


class FILESYSTEM(DictPathEvent):
pass
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
if self._data_path.is_file():
# detect type of file content using magic
from bbot.core.helpers.libmagic import get_magic_info, get_compression

extension, mime_type, description, confidence = get_magic_info(self.data["path"])
self.data["magic_extension"] = extension
self.data["magic_mime_type"] = mime_type
self.data["magic_description"] = description
self.data["magic_confidence"] = confidence
# detection compression
compression = get_compression(mime_type)
if compression:
self.add_tag("compressed")
self.add_tag(f"{compression}-archive")
self.data["compression"] = compression


class RAW_DNS_RECORD(DictHostEvent, DnsEvent):
Expand Down
68 changes: 68 additions & 0 deletions bbot/core/helpers/libmagic.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
import puremagic


def get_magic_info(file):

magic_detections = puremagic.magic_file(file)
if magic_detections:
magic_detections.sort(key=lambda x: x.confidence, reverse=True)
detection = magic_detections[0]
return detection.extension, detection.mime_type, detection.name, detection.confidence
return "", "", "", 0


def get_compression(mime_type):
mime_type = mime_type.lower()
# from https://github.com/cdgriffith/puremagic/blob/master/puremagic/magic_data.json
compression_map = {
"application/gzip": "gzip", # Gzip compressed file
"application/zip": "zip", # Zip archive
"application/x-bzip2": "bzip2", # Bzip2 compressed file
"application/x-xz": "xz", # XZ compressed file
"application/x-7z-compressed": "7z", # 7-Zip archive
"application/vnd.rar": "rar", # RAR archive
"application/x-lzma": "lzma", # LZMA compressed file
"application/x-compress": "compress", # Unix compress file
"application/zstd": "zstd", # Zstandard compressed file
"application/x-lz4": "lz4", # LZ4 compressed file
"application/x-tar": "tar", # Tar archive
"application/x-zip-compressed-fb2": "zip", # Zip archive (FB2)
"application/epub+zip": "zip", # EPUB book (Zip archive)
"application/pak": "pak", # PAK archive
"application/x-lha": "lha", # LHA archive
"application/arj": "arj", # ARJ archive
"application/vnd.ms-cab-compressed": "cab", # Microsoft Cabinet archive
"application/x-sit": "sit", # StuffIt archive
"application/binhex": "binhex", # BinHex encoded file
"application/x-lrzip": "lrzip", # Long Range ZIP
"application/x-alz": "alz", # ALZip archive
"application/x-tgz": "tgz", # Gzip compressed Tar archive
"application/x-gzip": "gzip", # Gzip compressed file
"application/x-lzip": "lzip", # Lzip compressed file
"application/x-zstd-compressed-tar": "zstd", # Zstandard compressed Tar archive
"application/x-lz4-compressed-tar": "lz4", # LZ4 compressed Tar archive
"application/vnd.comicbook+zip": "zip", # Comic book archive (Zip)
"application/vnd.palm": "palm", # Palm OS data
"application/fictionbook2+zip": "zip", # FictionBook 2.0 (Zip)
"application/fictionbook3+zip": "zip", # FictionBook 3.0 (Zip)
"application/x-cpio": "cpio", # CPIO archive
"application/x-java-pack200": "pack200", # Java Pack200 archive
"application/x-par2": "par2", # PAR2 recovery file
"application/x-rar-compressed": "rar", # RAR archive
"application/java-archive": "zip", # Java Archive (JAR)
"application/x-webarchive": "zip", # Web archive (Zip)
"application/vnd.android.package-archive": "zip", # Android package (APK)
"application/x-itunes-ipa": "zip", # iOS application archive (IPA)
"application/x-stuffit": "sit", # StuffIt archive
"application/x-archive": "ar", # Unix archive
"application/x-qpress": "qpress", # Qpress archive
"application/x-xar": "xar", # XAR archive
"application/x-ace": "ace", # ACE archive
"application/x-zoo": "zoo", # Zoo archive
"application/x-arc": "arc", # ARC archive
"application/x-zstd-compressed-tar": "zstd", # Zstandard compressed Tar archive
"application/x-lz4-compressed-tar": "lz4", # LZ4 compressed Tar archive
"application/vnd.comicbook-rar": "rar", # Comic book archive (RAR)
}

return compression_map.get(mime_type, "")
16 changes: 13 additions & 3 deletions bbot/core/shared_deps.py
Original file line number Diff line number Diff line change
Expand Up @@ -79,13 +79,23 @@
"ignore_errors": True,
},
{
"name": "Install Chromium dependencies (Debian)",
"name": "Install Chromium dependencies (Ubuntu 24.04)",
"package": {
"name": "libasound2,libatk-bridge2.0-0,libatk1.0-0,libcairo2,libcups2,libdrm2,libgbm1,libnss3,libpango-1.0-0,libxcomposite1,libxdamage1,libxfixes3,libxkbcommon0,libxrandr2",
"name": "libasound2t64,libatk-bridge2.0-0,libatk1.0-0,libcairo2,libcups2,libdrm2,libgbm1,libnss3,libpango-1.0-0,libglib2.0-0,libxcomposite1,libxdamage1,libxfixes3,libxkbcommon0,libxrandr2",
"state": "present",
},
"become": True,
"when": "ansible_facts['os_family'] == 'Debian'",
"when": "ansible_facts['distribution'] == 'Ubuntu' and ansible_facts['distribution_version'] == '24.04'",
"ignore_errors": True,
},
{
"name": "Install Chromium dependencies (Other Debian-based)",
"package": {
"name": "libasound2,libatk-bridge2.0-0,libatk1.0-0,libcairo2,libcups2,libdrm2,libgbm1,libnss3,libpango-1.0-0,libglib2.0-0,libxcomposite1,libxdamage1,libxfixes3,libxkbcommon0,libxrandr2",
"state": "present",
},
"become": True,
"when": "ansible_facts['os_family'] == 'Debian' and not (ansible_facts['distribution'] == 'Ubuntu' and ansible_facts['distribution_version'] == '24.04')",
"ignore_errors": True,
},
{
Expand Down
147 changes: 147 additions & 0 deletions bbot/db/sql/models.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,147 @@
# This file contains SQLModel (Pydantic + SQLAlchemy) models for BBOT events, scans, and targets.
# Used by the SQL output modules, but portable for outside use.

import json
import logging
from datetime import datetime
from pydantic import ConfigDict
from typing import List, Optional
from typing_extensions import Annotated
from pydantic.functional_validators import AfterValidator
from sqlmodel import inspect, Column, Field, SQLModel, JSON, String, DateTime as SQLADateTime


log = logging.getLogger("bbot_server.models")


def naive_datetime_validator(d: datetime):
"""
Converts all dates into UTC, then drops timezone information.
This is needed to prevent inconsistencies in sqlite, because it is timezone-naive.
"""
# drop timezone info
return d.replace(tzinfo=None)


NaiveUTC = Annotated[datetime, AfterValidator(naive_datetime_validator)]


class CustomJSONEncoder(json.JSONEncoder):
def default(self, obj):
# handle datetime
if isinstance(obj, datetime):
return obj.isoformat()
return super().default(obj)


class BBOTBaseModel(SQLModel):
model_config = ConfigDict(extra="ignore")

def __init__(self, *args, **kwargs):
self._validated = None
super().__init__(*args, **kwargs)

@property
def validated(self):
try:
if self._validated is None:
self._validated = self.__class__.model_validate(self)
return self._validated
except AttributeError:
return self

def to_json(self, **kwargs):
return json.dumps(self.validated.model_dump(), sort_keys=True, cls=CustomJSONEncoder, **kwargs)

@classmethod
def _pk_column_names(cls):
return [column.name for column in inspect(cls).primary_key]

def __hash__(self):
return hash(self.to_json())

def __eq__(self, other):
return hash(self) == hash(other)


### EVENT ###


class Event(BBOTBaseModel, table=True):

def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
data = self._get_data(self.data, self.type)
self.data = {self.type: data}
if self.host:
self.reverse_host = self.host[::-1]

def get_data(self):
return self._get_data(self.data, self.type)

@staticmethod
def _get_data(data, type):
# handle SIEM-friendly format
if isinstance(data, dict) and list(data) == [type]:
return data[type]
return data

uuid: str = Field(
primary_key=True,
index=True,
nullable=False,
)
id: str = Field(index=True)
type: str = Field(index=True)
scope_description: str
data: dict = Field(sa_type=JSON)
host: Optional[str]
port: Optional[int]
netloc: Optional[str]
# store the host in reversed form for efficient lookups by domain
reverse_host: Optional[str] = Field(default="", exclude=True, index=True)
resolved_hosts: List = Field(default=[], sa_type=JSON)
dns_children: dict = Field(default={}, sa_type=JSON)
web_spider_distance: int = 10
scope_distance: int = Field(default=10, index=True)
scan: str = Field(index=True)
timestamp: NaiveUTC = Field(index=True)
parent: str = Field(index=True)
tags: List = Field(default=[], sa_type=JSON)
module: str = Field(index=True)
module_sequence: str
discovery_context: str = ""
discovery_path: List[str] = Field(default=[], sa_type=JSON)
parent_chain: List[str] = Field(default=[], sa_type=JSON)


### SCAN ###


class Scan(BBOTBaseModel, table=True):
id: str = Field(primary_key=True)
name: str
status: str
started_at: NaiveUTC = Field(index=True)
finished_at: Optional[NaiveUTC] = Field(default=None, sa_column=Column(SQLADateTime, nullable=True, index=True))
duration_seconds: Optional[float] = Field(default=None)
duration: Optional[str] = Field(default=None)
target: dict = Field(sa_type=JSON)
preset: dict = Field(sa_type=JSON)


### TARGET ###


class Target(BBOTBaseModel, table=True):
name: str = "Default Target"
strict_scope: bool = False
seeds: List = Field(default=[], sa_type=JSON)
whitelist: List = Field(default=None, sa_type=JSON)
blacklist: List = Field(default=[], sa_type=JSON)
hash: str = Field(sa_column=Column("hash", String, unique=True, primary_key=True, index=True))
scope_hash: str = Field(sa_column=Column("scope_hash", String, index=True))
seed_hash: str = Field(sa_column=Column("seed_hashhash", String, index=True))
whitelist_hash: str = Field(sa_column=Column("whitelist_hash", String, index=True))
blacklist_hash: str = Field(sa_column=Column("blacklist_hash", String, index=True))
2 changes: 1 addition & 1 deletion bbot/modules/baddns.py
Original file line number Diff line number Diff line change
Expand Up @@ -22,7 +22,7 @@ class baddns(BaseModule):
"enabled_submodules": "A list of submodules to enable. Empty list (default) enables CNAME, TXT and MX Only",
}
module_threads = 8
deps_pip = ["baddns~=1.1.864"]
deps_pip = ["baddns~=1.4.13"]

def select_modules(self):
selected_submodules = []
Expand Down
2 changes: 1 addition & 1 deletion bbot/modules/baddns_direct.py
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@ class baddns_direct(BaseModule):
"custom_nameservers": "Force BadDNS to use a list of custom nameservers",
}
module_threads = 8
deps_pip = ["baddns~=1.1.864"]
deps_pip = ["baddns~=1.4.13"]

scope_distance_modifier = 1

Expand Down
2 changes: 1 addition & 1 deletion bbot/modules/baddns_zone.py
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ class baddns_zone(baddns_module):
"only_high_confidence": "Do not emit low-confidence or generic detections",
}
module_threads = 8
deps_pip = ["baddns~=1.1.864"]
deps_pip = ["baddns~=1.4.13"]

def set_modules(self):
self.enabled_submodules = ["NSEC", "zonetransfer"]
Expand Down
2 changes: 1 addition & 1 deletion bbot/modules/badsecrets.py
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@ class badsecrets(BaseModule):
options_desc = {
"custom_secrets": "Include custom secrets loaded from a local file",
}
deps_pip = ["badsecrets~=0.4.490"]
deps_pip = ["badsecrets~=0.6.21"]

async def setup(self):
self.custom_secrets = None
Expand Down
Loading

0 comments on commit 363cd6d

Please sign in to comment.