Portscan CDN Detection

bbot/modules/portscan.py

@@ -6,6 +6,9 @@
 from bbot.modules.base import BaseModule
 
 
+# TODO: this module is getting big. It should probably be two modules: one for ping and one for SYN.
+
+
 class portscan(BaseModule):
     flags = ["active", "portscan", "safe"]
     watched_events = ["IP_ADDRESS", "IP_RANGE", "DNS_NAME"]
@@ -27,6 +30,8 @@ class portscan(BaseModule):
         "adapter_ip": "",
         "adapter_mac": "",
         "router_mac": "",
+        "cdn_tags": "cdn-",
+        "allowed_cdn_ports": None,
     }
     options_desc = {
         "top_ports": "Top ports to scan (default 100) (to override, specify 'ports')",
@@ -39,6 +44,8 @@ class portscan(BaseModule):
         "adapter_ip": "Send packets using this IP address. Not needed unless masscan's autodetection fails",
         "adapter_mac": "Send packets using this as the source MAC address. Not needed unless masscan's autodetection fails",
         "router_mac": "Send packets to this MAC address as the destination. Not needed unless masscan's autodetection fails",
+        "cdn_tags": "Comma-separated list of tags to skip, e.g. 'cdn,cloud'",
+        "allowed_cdn_ports": "Comma-separated list of ports that are allowed to be scanned for CDNs",
     }
     deps_common = ["masscan"]
     batch_size = 1000000
@@ -60,7 +67,15 @@ async def setup(self):
             try:
                 self.helpers.parse_port_string(self.ports)
             except ValueError as e:
-                return False, f"Error parsing ports: {e}"
+                return False, f"Error parsing ports '{self.ports}': {e}"
+        self.cdn_tags = [t.strip() for t in self.config.get("cdn_tags", "").split(",")]
+        self.allowed_cdn_ports = self.config.get("allowed_cdn_ports", None)
+        if self.allowed_cdn_ports is not None:
+            try:
+                self.allowed_cdn_ports = [int(p.strip()) for p in self.allowed_cdn_ports.split(",")]
+            except Exception as e:
+                return False, f"Error parsing allowed CDN ports '{self.allowed_cdn_ports}': {e}"
+
         # whether we've finished scanning our original scan targets
         self.scanned_initial_targets = False
         # keeps track of individual scanned IPs and their open ports
@@ -227,9 +242,20 @@ async def emit_open_port(self, ip, port, parent_event):
             parent=parent_event,
             context=f"{{module}} executed a {scan_type} scan against {parent_event.data} and found: {{event.type}}: {{event.data}}",
         )
-        await self.emit_event(event)
+
+        await self.emit_event(event, abort_if=self.abort_if)
         return event
 
+    def abort_if(self, event):
+        if self.allowed_cdn_ports is not None:
+            # if the host is a CDN
+            for cdn_tag in self.cdn_tags:
+                if any(t.startswith(str(cdn_tag)) for t in event.tags):
+                    # and if its port isn't in the list of allowed CDN ports
+                    if event.port not in self.allowed_cdn_ports:
+                        return True, "event is a CDN and port is not in the allowed list"
+        return False
+
     def parse_json_line(self, line):
         try:
             j = json.loads(line)

bbot/test/test_step_2/module_tests/test_module_portscan.py

@@ -109,10 +109,12 @@ def check(self, module_test, events):
                 if e.type == "DNS_NAME" and e.data == "dummy.asdf.evilcorp.net" and str(e.module) == "dummy_module"
             ]
         )
-        assert 2 <= len([e for e in events if e.type == "IP_ADDRESS" and e.data == "8.8.8.8"]) <= 3
-        assert 2 <= len([e for e in events if e.type == "IP_ADDRESS" and e.data == "8.8.4.4"]) <= 3
-        assert 2 <= len([e for e in events if e.type == "IP_ADDRESS" and e.data == "8.8.4.5"]) <= 3
-        assert 2 <= len([e for e in events if e.type == "IP_ADDRESS" and e.data == "8.8.4.6"]) <= 3
+        # the reason these numbers aren't exactly predictable is because we can't predict which one arrives first
+        # to the portscan module. Sometimes, one that would normally be deduped is force-emitted because it led to a new open port.
+        assert 2 <= len([e for e in events if e.type == "IP_ADDRESS" and e.data == "8.8.8.8"]) <= 4
+        assert 2 <= len([e for e in events if e.type == "IP_ADDRESS" and e.data == "8.8.4.4"]) <= 4
+        assert 2 <= len([e for e in events if e.type == "IP_ADDRESS" and e.data == "8.8.4.5"]) <= 4
+        assert 2 <= len([e for e in events if e.type == "IP_ADDRESS" and e.data == "8.8.4.6"]) <= 4
         assert 1 == len([e for e in events if e.type == "OPEN_TCP_PORT" and e.data == "8.8.8.8:443"])
         assert 1 == len([e for e in events if e.type == "OPEN_TCP_PORT" and e.data == "8.8.4.5:80"])
         assert 1 == len([e for e in events if e.type == "OPEN_TCP_PORT" and e.data == "8.8.4.6:631"])

docs/scanning/tips_and_tricks.md

@@ -77,7 +77,38 @@ You can also pair the web spider with subdomain enumeration:
 bbot -t evilcorp.com -f subdomain-enum -c spider.yml
 ```
 
-### Ingesting BBOT Data Into SIEM (Elastic, Splunk)
+### Exclude CDNs from Port Scan
+
+If you want to exclude CDNs (e.g. Cloudflare) from port scanning, you can set the `allowed_cdn_ports` config option in the `portscan` module. For example, to allow only port 80 (HTTP) and 443 (HTTPS), you can do the following:
+
+```bash
+bbot -t evilcorp.com -m portscan -c modules.portscan.allowed_cdn_ports=80,443
+```
+
+By default, if you set `allowed_cdn_ports`, it will skip only providers marked as CDNs. If you want to skip cloud providers as well, you can set `cdn_tags`, which is a comma-separated list of tags to skip (matched against the beginning of each tag).
+
+```bash
+bbot -t evilcorp.com -m portscan -c modules.portscan.allowed_cdn_ports=80,443 modules.portscan.cdn_tags=cdn-,cloud-
+```
+
+...or via a preset:
+
+```yaml title="skip_cdns.yml"
+modules:
+  - portscan
+
+config:
+  modules:
+    portscan:
+      allowed_cdn_ports: 80,443
+      cdn_tags: cdn-,cloud-
+```
+
+```bash
+bbot -t evilcorp.com -p skip_cdns.yml
+```
+
+### Ingest BBOT Data Into SIEM (Elastic, Splunk)
 
 If your goal is to feed BBOT data into a SIEM such as Elastic, be sure to enable this option when scanning: