mirror of
https://github.com/hugo-fixit/FixIt.git
synced 2026-08-24 15:28:57 +00:00
Cell
120 lines
1.9 KiB
Plaintext
120 lines
1.9 KiB
Plaintext
User-agent: *
|
|
|
|
Disallow: /images/
|
|
Disallow: /js/
|
|
Disallow: /css/
|
|
|
|
Disallow: /*link/
|
|
Disallow: /*offline/
|
|
Disallow: /*404.html$
|
|
Disallow: /*.md$
|
|
|
|
{{ if .Site.Params.robots.allow_crawlers -}}
|
|
# Content-Signal (non-standard, for AI-aware crawlers)
|
|
Content-Signal: search=yes, ai-input=yes, ai-train=no
|
|
{{- end }}
|
|
|
|
User-agent: MJ12bot
|
|
Disallow: /
|
|
|
|
User-agent: AhrefsBot
|
|
Disallow: /
|
|
|
|
User-agent: BLEXBot
|
|
Disallow: /
|
|
|
|
# Block SISTRIX
|
|
User-agent: SISTRIX Crawler
|
|
Disallow: /
|
|
User-agent: sistrix
|
|
Disallow: /
|
|
User-agent: 007ac9
|
|
Disallow: /
|
|
User-agent: 007ac9 Crawler
|
|
Disallow: /
|
|
|
|
# Block Uptime robot
|
|
User-agent: UptimeRobot/2.0
|
|
Disallow: /
|
|
|
|
# Block Ezooms Robot
|
|
User-agent: Ezooms Robot
|
|
Disallow: /
|
|
|
|
# Block Perl LWP
|
|
User-agent: Perl LWP
|
|
Disallow: /
|
|
|
|
# Block netEstate NE Crawler (+http://www.website-datenbank.de/)
|
|
User-agent: netEstate NE Crawler (+http://www.website-datenbank.de/)
|
|
Disallow: /
|
|
|
|
# Block WiseGuys Robot
|
|
User-agent: WiseGuys Robot
|
|
Disallow: /
|
|
|
|
# Block Turnitin Robot
|
|
User-agent: Turnitin Robot
|
|
Disallow: /
|
|
|
|
# Block Heritrix
|
|
User-agent: Heritrix
|
|
Disallow: /
|
|
|
|
# Block pricepi
|
|
User-agent: pimonster
|
|
Disallow: /
|
|
|
|
User-agent: SurdotlyBot
|
|
Disallow: /
|
|
|
|
User-agent: ZoominfoBot
|
|
Disallow: /
|
|
|
|
{{ if not .Site.Params.robots.allow_crawlers -}}
|
|
# OpenAI, ChatGPT
|
|
User-agent: GPTBot
|
|
Disallow: /
|
|
|
|
# Google AI (Bard, etc)
|
|
User-agent: Google-Extended
|
|
Disallow: /
|
|
|
|
# Block common crawl
|
|
User-agent: CCBot
|
|
Disallow: /
|
|
|
|
# Facebook
|
|
User-agent: FacebookBot
|
|
Disallow: /
|
|
|
|
# Cohere.ai
|
|
User-agent: cohere-ai
|
|
Disallow: /
|
|
|
|
# Perplexity
|
|
User-agent: PerplexityBot
|
|
Disallow: /
|
|
|
|
# Anthropic
|
|
User-agent: anthropic-ai
|
|
Disallow: /
|
|
|
|
# Claudebot by Anthropic
|
|
User-agent: ClaudeBot
|
|
Disallow: /
|
|
{{- end -}}
|
|
|
|
{{- range .Site.Params.robots.rules }}
|
|
|
|
User-agent: {{ .user_agent }}
|
|
{{- with .disallow }}
|
|
Disallow: {{ . }}
|
|
{{- end }}
|
|
{{- with .allow }}
|
|
Allow: {{ . }}
|
|
{{- end -}}
|
|
{{- end -}}
|
|
|
|
Sitemap: {{ "/sitemap.xml" | absURL }}
|