Current Release: v0.3.9 (2019.11.26)
dirsearch is a simple command line tool designed to brute force directories and files in websites.
git clone https://github.com/maurosoria/dirsearch.git
cd dirsearch
python3 dirsearch.py -u <URL> -e <EXTENSION>
you can also use this alias to send directly to proxy
python3 /path/to/dirsearch/dirsearch.py --http-proxy=localhost:8080
Options:
-h, --help show this help message and exit
Mandatory:
-u URL, --url=URL URL target
-L URLLIST, --url-list=URLLIST
URL list target
-e EXTENSIONS, --extensions=EXTENSIONS
Extension list separated by comma (Example: php,asp)
-E, --extensions-list
Use predefined list of common extensions
Dictionary Settings:
-w WORDLIST, --wordlist=WORDLIST
-l, --lowercase
-f, --force-extensions
Force extensions for every wordlist entry (like in
DirBuster)
General Settings:
-s DELAY, --delay=DELAY
Delay between requests (float number)
-r, --recursive Bruteforce recursively
-R RECURSIVE_LEVEL_MAX, --recursive-level-max=RECURSIVE_LEVEL_MAX
Max recursion level (subdirs) (Default: 1 [only
rootdir + 1 dir])
--suppress-empty, --suppress-empty
--scan-subdir=SCANSUBDIRS, --scan-subdirs=SCANSUBDIRS
Scan subdirectories of the given -u|--url (separated
by comma)
--exclude-subdir=EXCLUDESUBDIRS, --exclude-subdirs=EXCLUDESUBDIRS
Exclude the following subdirectories during recursive
scan (separated by comma)
-t THREADSCOUNT, --threads=THREADSCOUNT
Number of Threads
-x EXCLUDESTATUSCODES, --exclude-status=EXCLUDESTATUSCODES
Exclude status code, separated by comma (example: 301,
500)
--exclude-texts=EXCLUDETEXTS
Exclude responses by texts, separated by comma
(example: "Not found", "Error")
--exclude-regexps=EXCLUDEREGEXPS
Exclude responses by regexps, separated by comma
(example: "Not foun[a-z]{1}", "^Error$")
-c COOKIE, --cookie=COOKIE
--ua=USERAGENT, --user-agent=USERAGENT
-F, --follow-redirects
-H HEADERS, --header=HEADERS
Headers to add (example: --header "Referer:
example.com" --header "User-Agent: IE"
--random-agents, --random-user-agents
Connection Settings:
--timeout=TIMEOUT Connection timeout
--ip=IP Resolve name to IP address
--proxy=HTTPPROXY, --http-proxy=HTTPPROXY
Http Proxy (example: localhost:8080
--http-method=HTTPMETHOD
Method to use, default: GET, possible also: HEAD;POST
--max-retries=MAXRETRIES
-b, --request-by-hostname
By default dirsearch will request by IP for speed.
This forces requests by hostname
Reports:
--simple-report=SIMPLEOUTPUTFILE
Only found paths
--plain-text-report=PLAINTEXTOUTPUTFILE
Found paths with status codes
--json-report=JSONOUTPUTFILE
- Windows XP/7/8/10
- GNU/Linux
- MacOSX
- Multithreaded
- Keep alive connections
- Support for multiple extensions (-e|--extensions asp,php)
- Reporting (plain text, JSON)
- Heuristically detects invalid web pages
- Recursive brute forcing
- HTTP proxy support
- User agent randomization
- Batch processing
- Request delaying
- Option to remove dot from extension when forcing (--nd, example%EXT% instead of example.%EXT%)
- Options to display only items with response length from range (--min & --max)
- Option to whitelist response codes (-i 200,500)
- Option to remove output from console (-q, keeps output to files)
- Option to add custom suffixes to filenames without dots (--suff .BAK,.old, example.%EXT%%SUFFIX%)
Dictionaries must be text files. Each line will be processed as such, except that the special word %EXT% is used, which will generate one entry for each extension (-e | --extension) passed as an argument.
Example:
- example/
- example.%EXT%
Passing the extensions "asp" and "aspx" will generate the following dictionary:
- example/
- example.asp
- example.aspx
You can also use -f | --force-extensions switch to append extensions to every word in the wordlists (like DirBuster).
Some examples how to use dirsearch - those are the most common arguments. If you need all, just use the "-h" argument.
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt --recursive -R 2
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt --recursive -R 4 --scan-subdirs=/,/wp-content/,/wp-admin/
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt --exclude-texts=This,AndThat
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt -H "User-Agent: IE"
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt -t 20
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt --random-agents
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt --json-report=reports/target.json
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt --simple-report=reports/target-paths.txt
python3 dirsearch.py -e php,txt,zip -u https://target -w db/dicc.txt --plain-text-report=reports/target-paths-and-status.json
Install Docker
curl -fsSL https://get.docker.com | bash
To use docker you need superuser power
To create image
docker build -t "dirsearch:v0.3.8" .
dirsearch this is name the image and v0.3.8 is version
For using
docker run -it --rm "dirsearch:v0.3.8" -u target -e php,html,png,js,jpg
target is the site or IP
Copyright (C) Mauro Soria (maurosoria at gmail dot com)
License: GNU General Public License, version 2
Special thanks for these people.
- mzfr
- Damian89
- Bo0oM
- liamosaur
- redshark1802
- SUHAR1K
- FireFart
- k2l8m11n2
- vlohacks
- r0p0s3c
- V-Rico
- russtone