Cve-search - Common Vulnerabilities and Exposures local search tool

cve-search is a tool to import CVE (Common Vulnerabilities and Exposures) and CPE (Common Platform Enumeration) into a MongoDB to facilitate search and processing of CVEs.

This project is maintained by adulau


cve-search is a tool to import CVE (Common Vulnerabilities and Exposures) and CPE (Common Platform Enumeration) into a MongoDB to facilitate search and processing of CVEs.

The main objective of the software is to avoid doing direct and public lookup into the public CVE databases. This is usually faster to do local lookups and limits your sensitive queries via the Internet.

cve-search visualization


The requirements can be installed with pip:

sudo pip3 install -r requirements.txt

Installation of MongoDB

First, you'll need to have a Python 3 installation (3.2 or 3.3 preferred). Then you need to install MongoDB (2.2) from source (this should also work with any standard packages from your favorite distribution). Don't forget to install the headers for development while installing MongoDB. You can go to for to get the packages for your distribution, or for the source code.

Populating the database

For the initial run, you need to populate the CVE database by running:

./ -p
./ -c

It will fetch all the existing XML files from the Common Vulnerabilities and Exposures database and the Common Platform Enumeration.

A more detailed documentation can be found in the Documentations folder of the project.

Database and collections

The MongoDB database is called cvedb and there are 8 collections:

Updating the database

An updater script helps to start the db_mgmt_*

./ -v

You can run it in a crontab, logging is done in syslog by default.


You can search the database using

./ -p cisco:ios:12.4
./ -p cisco:ios:12.4 -o json
./ -f nagios -n
./ -p microsoft:windows_7 -o html

If you want to search all the WebEx vulnerabilities and only printing the official references from the supplier.

./ -p webex: -o csv  -v "cisco"

You can also dump the JSON for a specific CVE ID.

./ -c CVE-2010-3333

Or you can use the XMPP bot

./ -j -p strongpassword

Or dump the last 2 CVE entries in RSS or Atom format

./ -f atom -l 2

Or you can use the webinterface.


Usage of the ranking database

There is a ranking database allowing to rank software vulnerabilities based on their common platform enumeration name. The ranking can be done per organization or department within your organization or any meaningful name for you.

As an example, you can add a partial CPE name like "sap:netweaver" which is very critical for your accounting department.

./python3.3  -c "sap:netweaver" -g "accounting" -r 3

and then you can lookup the ranking (-r option) for a specific CVE-ID:

./python3.3 -c CVE-2012-4341  -r  -n

Advanced usage

As cve-search is based on a set of tools, it can be used and combined with standard Unix tools. If you ever wonder what are the top vendors using the term "unknown" for their vulnerabilities:

python3 -q unknown -f | jq -r '. | .vulnerable_configuration[0]' | cut -f3 -d: | sort  | uniq -c  | sort -nr | head -10

1500 oracle
381 sun
372 hp
232 google
208 ibm
126 mozilla
103 microsoft
100 adobe
 78 apple
 68 linux

You can compare CVSS (Common Vulnerability Scoring System ) values of some products based on their CPE name. Like comparing oracle:java versus sun:jre and using R to make some statistics about their CVSS values:

python3 -p oracle:java -o json  | jq -r '.cvss' | Rscript -e 'summary(as.numeric(read.table(file("stdin"))[,1]))'
Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
1.800   5.350   9.300   7.832  10.000  10.000

python3 -p sun:jre -o json  | jq -r '.cvss' | Rscript -e 'summary(as.numeric(read.table(file("stdin"))[,1]))'
Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
0.000   5.000   7.500   7.333  10.000  10.000

Fulltext indexing

If you want to index all the CVEs from your current MongoDB collection:


and you query the fulltext index (to get a list of matching CVE-ID):

./python3.3 -q NFS -q Linux

or to query the fulltext index and output the JSON object for each CVE-ID:

./python3.3 -q NFS -q Linux -j

Fulltext visualization

The fulltext indexer visualization is using the fulltext indexes to build a list of the most common keywords used in CVE. NLTK is required to generate the keywords with the most common English stopwords and lemmatize the output. NTLK for Python 3 exists but you need to use the alpha version of NLTK.

./python3.3  -g -s >cve.json

You can see a visualization on the demo site.

Web interface

The web interface is a minimal interface to see the last CVE entries and query a specific CVE. You'll need flask in order to run the website and Flask-PyMongo. To start the web interface:

cd ./web

Then you can connect on to browser the last CVE.

Software using cve-search


cve-search is free software released under the "Modified BSD license"

Copyright (c) 2012 Wim Remes -
Copyright (c) 2012-2014 Alexandre Dulaunoy -
Copyright (c) 2014 Pieter-Jan Moreels -