# ------------------------------------------------------
# sitemap entries
# ------------------------------------------------------

User-agent: *
Allow: https://www.cybersmile.org/sitemap-index.xml
Allow: https://www.cybersmile.org/sitemap-post.xml
Allow: https://www.cybersmile.org/sitemap-page.xml
Allow: https://www.cybersmile.org/sitemap-partner.xml
Allow: https://www.cybersmile.org/sitemap-ambassador.xml
Allow: https://www.cybersmile.org/sitemap-advisor.xml
Allow: https://www.cybersmile.org/sitemap-news.xml

# ------------------------------------------------------
# global rules
# ------------------------------------------------------

User-agent: *
Disallow: /wp-admin/
Disallow: /wp-login.php
Disallow: /wp-register.php
Disallow: /wp-includes/
Disallow: /xmlrpc.php
Disallow: /wp-json/
Disallow: /feed/
Disallow: /*/feed/
Disallow: /comments/feed/

# ------------------------------------------------------
# file types and params
# ------------------------------------------------------

Disallow: /*?author=
Disallow: /*?page=
Disallow: /*?sort=
Disallow: /*?filter=

Disallow: /*.pdf$
Disallow: /*.zip$
Disallow: /*.doc$
Disallow: /*.csv$
Disallow: /*.sql$
Disallow: /*.tar$
Disallow: /*.gz$
Disallow: /*.rar$
Disallow: /*.txt$
Disallow: /*.md$

Disallow: /*LICENSE$
Disallow: /*LICENSE.txt$
Disallow: /*README$
Disallow: /*README.md$
Disallow: /*README.txt$
Disallow: /*license$
Disallow: /*license.txt$
Disallow: /*readme$
Disallow: /*readme.html$
Disallow: /*readme.md$
Disallow: /*readme.txt$

Disallow: /*package.json$
Disallow: /*composer.json$
Disallow: /*composer.lock$
Disallow: /*yarn.lock$

Disallow: /*debug.log$
Disallow: /*error_log$
Disallow: /*error$

# ------------------------------------------------------
# content areas
# ------------------------------------------------------

Disallow: /author/
Disallow: /blog/
Disallow: /campaignformat/
Disallow: /campaignindustry/
Disallow: /campaigntags/
Disallow: /campaigntype/
Disallow: /faq/
Disallow: /faqcats/
Disallow: /helpcenter_topics/
Disallow: /news_category/
Disallow: /org/
Disallow: /orgtags/
Disallow: /orgterritory/
Disallow: /press/
Disallow: /presstags/
Disallow: /resourcetype/
Disallow: /resourceage/
Disallow: /resourceperiod/
Disallow: /resourceterritory/
Disallow: /stream/
Disallow: /streamcats/
Disallow: /streamtags/
Disallow: /tag/
Disallow: /tags/
Disallow: /testimonial/
Disallow: /testimonial_countries/

# ------------------------------------------------------
# low-value urls
# ------------------------------------------------------

Disallow: /*?utm_*
Disallow: /*&utm_*
Disallow: /*?fbclid=*
Disallow: /*?gclid=*
Disallow: /*?mc_cid=*
Disallow: /*?mc_eid=*
Disallow: /*?msclkid=*
Disallow: /*?ref=*
Disallow: /*?source=*
Disallow: /*//*
Disallow: /?s=
Disallow: /*&s=
Disallow: /*?order=
Disallow: /*&order=
Disallow: /*?orderby=
Disallow: /*&orderby=
Disallow: /*?dir=
Disallow: /*&dir=

# ------------------------------------------------------
# unwanted bots
# ------------------------------------------------------

User-agent: AI2Bot
User-agent: AdClickerBot
User-agent: Ai2Bot-Dolma
User-agent: Amazonbot
User-agent: AwarioRssBot
User-agent: AwarioSmartBot
User-agent: Brightbot
User-agent: Brightbot 1.0
User-agent: BruteForceBot
User-agent: Bytespider
User-agent: CCBot
User-agent: ChatGPT-User
User-agent: ContentScraperBot
User-agent: Cotoyogi
User-agent: Crawlspace
User-agent: DataForSeoBot
User-agent: Diffbot
User-agent: FacebookBot
User-agent: FakeBingbot
User-agent: FakeGooglebot
User-agent: FriendlyCrawler
User-agent: GPTBot
User-agent: ICC-Crawler
User-agent: ImagesiftBot
User-agent: Kangaroo Bot
User-agent: LinerBot
User-agent: MachineLearningForPeaceBot
User-agent: Meltwater
User-agent: NewsNow
User-agent: Nicecrawler
User-agent: NovaAct
User-agent: Operator
User-agent: PanguBot
User-agent: PetalBot
User-agent: PiplBot
User-agent: QualifiedBot
User-agent: SEO-Bot
User-agent: Scoop.it
User-agent: Scrapy
User-agent: Seekr
User-agent: SemrushBot-OCOB
User-agent: Sidetrade indexer bot
User-agent: SpamBot
User-agent: Sqlmap
User-agent: TikTokSpider
User-agent: Timpibot
User-agent: TurnitinBot
User-agent: VelenPublicWebCrawler
User-agent: WebReaper
User-agent: Webzio-Extended
User-agent: YouBot
User-agent: amazon-kendra
User-agent: cohere-ai
User-agent: cohere-training-data-crawler
User-agent: iaskspider/2.0
User-agent: img2dataset
User-agent: magpie-crawler
User-agent: news-please
User-agent: omgili
User-agent: omgilibot
Disallow: /