# *=-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-==-=-=-==-=-
# XOOPS CUBE BRASIL - ROBOTS.TXT / by Mikhail Miguel
# ----------------------------------------------------
# Robots like Google and Yahoo understand the 2.0 specs of robots.txt,
# so we can allow them a bit more freedom about what to do.
#
# Other bots are known to be trouble, particularly those
# designed to copy entire sites.
#
# Good to remember: robots.txt file is more than 5000 characters.
# Google tool can analyze robots.txt files that are a maximum of 5000 characters.
#
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-==-=-=-==-=-*


User-agent: sitecheck.internetseer.com
Disallow: /

User-agent: Zealbot
Disallow: /

User-agent: SiteSnagger
Disallow: /

User-agent: WebStripper
Disallow: /

User-agent: WebCopier
Disallow: /

User-agent: Fetch
Disallow: /

User-agent: Offline Explorer
Disallow: /

User-agent: Teleport
Disallow: /

User-agent: TeleportPro
Disallow: /

User-agent: WebZIP
Disallow: /

User-agent: linko
Disallow: /

User-agent: HTTrack
Disallow: /

User-agent: Xenu
Disallow: /

User-agent: larbin
Disallow: /

User-agent: libwww
Disallow: /

User-agent: ZyBORG
Disallow: /

User-agent: Download Ninja
Disallow: /

User-agent: wget
Disallow: /

User-agent: grub-client
Disallow: /

User-agent: k2spider
Disallow: /

# *=-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-
# Hits many times per second, not acceptable
# http://www.nameprotect.com/botinfo.html
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

User-agent: NPBot
Disallow: /

# *=-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-
# A capture bot, downloads gazillions of pages
# with no public benefit: http://www.webreaper.net/
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

User-agent: WebReaper
Disallow: /

# *=-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-
# Allow the wayback-machine to index all/any pages
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

User-agent: ia_archiver
Allow: /*


# *=-=-=-==-=-=-==-=-=-==-=-|
# GOOGLE ------------------|
# =-=-=-==-=-=-==-=-=-==-=-*|

User-agent: Googlebot

Disallow: /*/admin/
Disallow: /*?page=print
Disallow: /*PHPSESSID
Disallow: /index.php?PHPSESSID
Disallow: /?PHPSESSID
Disallow: /*print.
Disallow: /*ratefile
Disallow: /*ratenews.
Disallow: /*reply
Disallow: /*sort=


# *=-=-=-==-=-=-==-=-=-==-=-
# ROOT FILES
# Because root files have high priority for google,
# and I don't want to see files without content like 'register.php'
# in the first page of results...
# =-=-=-==-=-=-==-=-=-==-=-*

Disallow: /admin.php
Disallow: /banners.php
Disallow: /notifications.php
Disallow: /pmlite.php
Disallow: /readpmsg.php
Disallow: /register.php
Disallow: /user.php
Disallow: /viewpmsg.php
Disallow: /edituser.php
Disallow: /footer.php
Disallow: /header.php
Disallow: /image.php
Disallow: /imagemanager.php
Disallow: /lostpass.php
Disallow: /mainfile.dist.php
Disallow: /mainfile.php
Disallow: /misc.php

# *=-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-
# MODULES
# ---------------------------------------------------
# Because each directory 'url' have high priority for google,
# and I don't want to see this paths without content in results
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

Disallow: /modules/pm/
Disallow: /modules/stdCache/
Disallow: /modules/system/
Disallow: /modules/tellafriend/
Disallow: /modules/xclBlocks/
Disallow: /modules/xclMail/
Disallow: /modules/xclObjects/
Disallow: /modules/d3forum/index.php?page=newtopic
Disallow: /modules/d3forum/index.php?page=reply
Disallow: /modules/legacyRender/
Disallow: /modules/mylinks/ratelink.php
Disallow: /modules/newbb/action.transfer.php
Disallow: /modules/mylinks/myheader.php


# *=-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-
# GOOGLE DON'T HAVE FRIENDS TO TELL...
# ---------------------------------------------------
# Because I don't want duplicate content
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

Disallow: /modules/shop/tell_a_friend.php
Disallow: /modules/articles/email.php


# *=-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-
# Because GOOGLE DON'T HAVE A PRINT
# ---------------------------------------------------
# To avoid duplicate content and
# direct access to the print pages (from google results)
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

Disallow: /modules/AMS/print.php
Disallow: /modules/addresses/print.php
Disallow: /modules/altern8news/print.php
Disallow: /modules/amevents/print.php
Disallow: /modules/amreviews/print.php
Disallow: /modules/articles/print.php
Disallow: /modules/artigos/print.php
Disallow: /modules/bulletin/?page=print
Disallow: /modules/bulletin/index.php?page=print
Disallow: /modules/catads/print.php
Disallow: /modules/content/print.php
Disallow: /modules/dictionary/print.php
Disallow: /modules/eNoticias/print.php
Disallow: /modules/ecotut/print.php
Disallow: /modules/eguide/print.php
Disallow: /modules/extcal/print.php
Disallow: /modules/info/print.php
Disallow: /modules/lexikon/print.php
Disallow: /modules/magazine/print.php
Disallow: /modules/mx-directory/print.php
Disallow: /modules/myMedia/print.php
Disallow: /modules/newbb/print.php
Disallow: /modules/news/print.php
Disallow: /modules/noticias/print.php
Disallow: /modules/piCal/print.php
Disallow: /modules/smartcontent/print.php
Disallow: /modules/smartfaq/print.php
Disallow: /modules/smartmedia/print.php
Disallow: /modules/smartobject/print.php
Disallow: /modules/smartsection/print.php
Disallow: /modules/soapbox/print.php
Disallow: /modules/stories/print.php
Disallow: /modules/tinycontent/print.php
Disallow: /modules/tinyd0/print.php
Disallow: /modules/userpage/print.php
Disallow: /modules/wakka/handlers/page/print.php
Disallow: /modules/weblinks/print.php
Disallow: /modules/weblog/print.php
Disallow: /modules/webmailfx/print.php
Disallow: /modules/wffaq/print.php
Disallow: /modules/wfsection/print.php
Disallow: /modules/wordbook/print.php
Disallow: /modules/xentcdmanager/print.php
Disallow: /modules/xfsection/print.php
Disallow: /modules/xmail/print.php
Disallow: /modules/zentrack/actions/print.php
Disallow: /modules/zmagazine/print.php

# Because GOOGLE DON'T REPORT BROKEN LINKS/FILES
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

Disallow: /modules/PDdownloads/brokenfile.php
Disallow: /modules/PDlinks/brokenfile.php
Disallow: /modules/addresses/brokenlink.php
Disallow: /modules/bizdir/brokenlink.php
Disallow: /modules/booklists/brokenlink.php
Disallow: /modules/downloadse/brokenfile.php
Disallow: /modules/gsdownloads/brokenfile.php
Disallow: /modules/hiperlinks/brokenlink.php
Disallow: /modules/magazine/brokenfile.php
Disallow: /modules/mx-directory/brokenlink.php
Disallow: /modules/myReviews/brokenfile.php
Disallow: /modules/mydownloads/brokenfile.php
Disallow: /modules/myed2k/brokenlink.php
Disallow: /modules/mylinks/brokenlink.php
Disallow: /modules/opencontent/brokenlink.php
Disallow: /modules/rha7downloads/brokenfile.php
Disallow: /modules/rmdp/broken.php
Disallow: /modules/weblinks/brokenlink.php
Disallow: /modules/webshow/brokenlink.php
Disallow: /modules/wfdownloads/brokenfile.php
Disallow: /modules/wflinks/brokenlink.php
Disallow: /modules/wfsection/brokenfile.php
Disallow: /modules/xdirectory/brokenlink.php
Disallow: /modules/xfsection/brokenfile.php
Disallow: /modules/zmagazine/brokenfile.php


# Because GOOGLE DON'T REPORT CONTENT
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

Disallow: /modules/asms/report.php
Disallow: /modules/mrbs/report.php
Disallow: /modules/newbb/report.php
Disallow: /modules/xhelp/report.php
Disallow: /modules/zentrack/reports/report.php


# Because GOOGLE DON'T REVIEW ANYTHING
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

Disallow: /modules/amreviews/review.php
Disallow: /modules/chatmx/review.php
Disallow: /modules/magazine/review.php
Disallow: /modules/mydownloads/review.php
Disallow: /modules/uu_cart/review.php
Disallow: /modules/wfdownloads/review.php
Disallow: /modules/wfsection/review.php
Disallow: /modules/xclDownloads/review.php


# OTHERS MODULES FILES
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

Disallow: /modules/mydownloads/modfile.php
Disallow: /modules/mydownloads/deletefile.php


# HEY, GOOOOOGLE! STAY HERE! DON'T GO AWAY!!!
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

Disallow: /modules/mydownloads/visit.php
Disallow: /modules/AMS/visit.php
Disallow: /modules/PDdownloads/visit.php
Disallow: /modules/PDlinks/visit.php
Disallow: /modules/addresses/visit.php
Disallow: /modules/adresses/visit.php
Disallow: /modules/bizdir/visit.php
Disallow: /modules/booklists/visit.php
Disallow: /modules/downloadse/visit.php
Disallow: /modules/hiperlinks/visit.php
Disallow: /modules/mx-directory/visit.php
Disallow: /modules/mydownloads/visit.php
Disallow: /modules/myed2k/visit.php
Disallow: /modules/mylinks/visit.php
Disallow: /modules/news/visit.php
Disallow: /modules/noticias/visit.php
Disallow: /modules/opencontent/visit.php
Disallow: /modules/rha7downloads/visit.php
Disallow: /modules/smartpartner/visit.php
Disallow: /modules/smartsection/visit.php
Disallow: /modules/weblinks/visit.php
Disallow: /modules/webshow/visit.php
Disallow: /modules/wfdownloads/visit.php
Disallow: /modules/wflinks/visit.php
Disallow: /modules/xdirectory/visit.php


# ONLY FOR SHORT URLS HACK
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

Disallow: /register.html
Disallow: /*brokenfile
Disallow: /*brokenlink
Disallow: /*brokenmovie
Disallow: /mydownloads+visit
Disallow: /newbb+report
Disallow: /wordbook+print


# FOR ANY OTHERS WEB BOTS CRAWLERS...
# =-=-=-==-=-=-==-=-=-==-=-=-=-=-==-=-=-==-=-=-==-=-*

User-agent: *

Disallow: /cache/
Disallow: /cgi-bin/
Disallow: /class/
Disallow: /core/
Disallow: /install/
Disallow: /kernel/
Disallow: /language/
Disallow: /preload/
Disallow: /settings/
Disallow: /temp/
Disallow: /templates_c/
Disallow: /tmp/


# ROOT FOLDERS - NON OFFICIAL
Disallow: /Frameworks/
Disallow: /common/
Disallow: /include/

# DUPLICATE CONTENT FROM APACHE

Disallow: /*?D=A
Disallow: /*?D=D
Disallow: /*?M=A
Disallow: /*?M=D
Disallow: /*?N=A
Disallow: /*?N=D
Disallow: /*?S=A
Disallow: /*?S=D

Marcadores: seo, webmasters, xoopscube.com.br