Added method to get crt.sh ids and crawl them.

discovery/crtsh.py

@@ -7,8 +7,7 @@ def __init__(self, word):
         self.word = word.replace(' ', '%20')
         self.results = ""
         self.totalresults = ""
-        self.server = "www.google.com"
-        self.hostname = "www.google.com"
+        self.server = "https://crt.sh/?q="
         self.userAgent = "(Mozilla/5.0 (Windows; U; Windows NT 6.0;en-US; rv:1.9.2) Gecko/20100116 Firefox/3.7"
         self.quantity = "100"
         self.counter = 0
@@ -16,7 +15,7 @@ def __init__(self, word):
 
     def do_search(self):
         try:
-            urly = "https://crt.sh/?q=%25" + self.word
+            urly = self.server + self.word
         except Exception as e:
             print(e)
         headers = {'User-Agent':'Mozilla/5.0 (Macintosh; Intel Mac OS X 10.10; rv:34.0) Gecko/20100101 Firefox/34.0'}
@@ -24,8 +23,32 @@ def do_search(self):
             r=requests.get(urly,headers=headers)
         except Exception as e:
             print(e)
-        self.results = r.text
-        self.totalresults += self.results
+        links = self.get_info(r.text)
+        for link in links:
+            r = requests.get(link, headers=headers)
+            self.results = r.text
+            self.totalresults += self.results
+
+    def get_info(self,text):
+        lines = []
+        for line in str(text).splitlines():
+            line = line.strip()
+            if 'id=' in line:
+                lines.append(line)
+        links = []
+        for i in range(len(lines)):
+            if i % 2 == 0:
+                current = lines[i]
+                current = current[43:]
+                link = ''
+                for ch in current:
+                    if ch == '"':
+                        break
+                    else:
+                        link += ch
+                links.append(('https://crt.sh?id=' + str(link)))
+        return links
+
 
     def get_hostnames(self):
         rawres = myparser.parser(self.results, self.word)

myparser.py

@@ -1,4 +1,3 @@
-import string
 import re
 
 

theHarvester.py

@@ -337,13 +337,13 @@ def start(argv):
                             db.store_all(word,all_hosts,'host','threatcrowd')
                         except Exception: pass
 
-                        """print("[-] Searching in CRTSH server..")
+                        print("[-] Searching in CRTSH server..")
                         search = crtsh.search_crtsh(word)
                         search.process()
                         hosts = search.get_hostnames()
                         all_hosts.extend(hosts)
                         db=stash.stash_manager()
-                        db.store_all(word,all_hosts,'host','CRTsh')"""
+                        db.store_all(word,all_hosts,'host','CRTsh')
 
                         print("[-] Searching in Virustotal server..")
                         search = virustotal.search_virustotal(word)