PhonePe · Prateek-Thakare · Feb 7, 2025 · Feb 1, 2025 · Feb 1, 2025 · Feb 3, 2025
diff --git a/Dockerfile b/Dockerfile
@@ -1,3 +1,10 @@
+FROM golang:1.22 as go-wayback-builder
+RUN git clone https://github.com/Abhinandan-Khurana/go-wayback.git
+WORKDIR go-wayback
+RUN GOOS=linux GOARCH=amd64 go build -o go-wayback v2/main.go
+RUN chmod +x go-wayback
+RUN cp go-wayback /usr/bin/
+
 FROM --platform=linux/amd64 python:3.9-slim
 
 # Install wget
@@ -20,7 +27,11 @@ RUN rm -rf *
 RUN echo "Installing Go_Virustotal"
 RUN wget https://github.com/Abhinandan-Khurana/go_virustotal/releases/download/v1.0.1/go_virustotal-linux-v1.0.1
 RUN mv go_virustotal-linux-v1.0.1 go_virustotal
-RUN mv go_virustotal /usr/bin
+RUN chmod +x go_virustotal
+RUN mv go_virustotal /usr/bin/
+
+# Install Go_Wayback
+COPY --from=go-wayback-builder /usr/bin/go-wayback /usr/bin
 
 # Install HTTPX
 RUN echo "Installing HTTPX"

diff --git a/configs/local.yml b/configs/local.yml
@@ -14,7 +14,7 @@ workflow:
     cmd: []
     workflowConfig:
       - moduleName : discovery
-        tools: ['Subfinder','Go_Virustotal']
+        tools: ['Subfinder','Go_Virustotal','Go_Wayback']
         order: 1
       - moduleName: prerecon
         tools: ['FindCDN', 'Naabu'] 

diff --git a/mantis/modules/discovery/Go_Virustotal.py b/mantis/modules/discovery/Go_Virustotal.py
@@ -9,6 +9,10 @@
 import logging
 
 '''
+Tool Author: Abhinandan Khurana
+Github: github.com/abhinandan-khurana
+Tool Link: github.com/abhinandan-khurana/go_virustotal
+
 Go_Virustotal module enumerates subdomain of the TLDs which are fetched from database. 
 Output file: .json 
 ------------------
@@ -68,7 +72,6 @@ def parse_report(self, outfile):
                     'tool_source': 'Go_Virustotal'
                 }
                 output_dict_list.append(domain_dict)
-        print(output_dict_list)
         return output_dict_list
 
     async def db_operations(self, tool_output_dict, asset=None):

diff --git a/mantis/modules/discovery/Go_Wayback.py b/mantis/modules/discovery/Go_Wayback.py
@@ -0,0 +1,72 @@
+from mantis.constants import ASSET_TYPE_SUBDOMAIN
+from mantis.utils.crud_utils import CrudUtils
+from mantis.tool_base_classes.toolScanner import ToolScanner
+from mantis.models.args_model import ArgsModel
+from mantis.utils.tool_utils import get_assets_grouped_by_type
+from mantis.constants import ASSET_TYPE_TLD
+import json
+import os
+import logging
+
+'''
+Tool Author: Abhinandan Khurana
+Github: github.com/abhinandan-khurana
+Tool Link: github.com/abhinandan-khurana/go-wayback
+
+Go_Wayback module enumerates subdomain of the TLDs which are fetched from database. 
+Output file: .txt 
+Each subdomain discovered is inserted into the database as a new asset. 
+'''
+
+class Go_Wayback(ToolScanner):
+
+    def __init__(self) -> None:
+        super().__init__()
+
+    async def get_commands(self, args: ArgsModel):
+        self.org = args.org
+        self.base_command = 'go-wayback  -o {output_file_path} -subdomain {input_domain}'
+        self.outfile_extension = ".txt"
+        self.assets = await get_assets_grouped_by_type(self, args, ASSET_TYPE_TLD)
+        return super().base_get_commands(self.assets)
+
+    def clean_url(self, url):
+        """Clean URL to extract subdomain only."""
+        # Remove protocol if present
+        if '://' in url:
+            url = url.split('://')[-1]
+
+        # Remove URL parameters and paths
+        url = url.split('?')[0]
+        url = url.split('/')[0]
+
+        # Remove port numbers if present
+        url = url.split(':')[0]
+
+        return url.strip()
+
+    def parse_report(self, outfile):
+        output_dict_list = []
+        wayback_output = open(outfile).readlines()
+        seen_domains = set()  # To avoid duplicates
+
+        for domain in wayback_output:
+            clean_domain = self.clean_url(domain.rstrip('\n'))
+
+            # Skip if domain already processed or empty
+            if not clean_domain or clean_domain in seen_domains:
+                continue
+
+            seen_domains.add(clean_domain)
+            domain_dict = {
+                '_id': clean_domain,
+                'asset': clean_domain,
+                'asset_type': ASSET_TYPE_SUBDOMAIN,
+                'org': self.org,
+                'tools_source': 'go-wayback'
+            }
+            output_dict_list.append(domain_dict)
+        return output_dict_list
+
+    async def db_operations(self, tool_output_dict, asset=None):
+        await CrudUtils.insert_assets(tool_output_dict)