Files
linux-scripts/add-fail2ban-ai-bots.sh
T
chiefgeek a1a17e81a1 Sync all scripts from website downloads — 352 scripts total
Includes updated JS challenge scripts with Claude-User whitelist,
same-site referer bypass, Blackbox-Exporter allowed bot, and all
new exporters, cheat sheets, and automation scripts.
2026-05-25 03:31:08 +02:00

396 lines
12 KiB
Bash
Executable File

#!/bin/bash
################################################################################
# Script Name: add-fail2ban-ai-bots.sh
# Version: 1.1
# Description: Adds a Fail2ban jail to block AI scrapers and unwanted bots
# that ignore robots.txt. Installs filter + jail config and
# reloads Fail2ban.
#
# Author: Phil Connor
# Contact: contact@mylinux.work
# Website: https://mylinux.work
# License: MIT
#
# Usage:
# sudo ./add-fail2ban-ai-bots.sh
# sudo ./add-fail2ban-ai-bots.sh --logpath /var/log/nginx/access.log
# sudo ./add-fail2ban-ai-bots.sh --bantime 604800
# sudo ./add-fail2ban-ai-bots.sh --dry-run
#
# Changelog:
# 1.1 — 2026-05-04: Removed Claude-Web, Perplexity-User, ChatGPT-User, and
# OAI-SearchBot from blocklist. These are user-facing fetcher bots that
# retrieve content when someone pastes a URL into an AI chat or search.
# Blocking them prevents your content from being cited in AI answers.
# Training crawlers (ClaudeBot, PerplexityBot, GPTBot) remain blocked.
#
################################################################################
set -euo pipefail
# ============================================================================
# DEFAULTS
# ============================================================================
readonly VERSION="1.1"
readonly SCRIPT_NAME="${0##*/}"
LOGPATH="auto"
BANTIME="86400"
MAXRETRY="1"
DRY_RUN=false
# Colors
RED='\033[0;31m'
GREEN='\033[0;32m'
YELLOW='\033[1;33m'
CYAN='\033[0;36m'
NC='\033[0m'
# ============================================================================
# HELPER FUNCTIONS
# ============================================================================
log_info() { echo -e "${GREEN}[INFO]${NC} $*"; }
log_warn() { echo -e "${YELLOW}[WARN]${NC} $*"; }
log_error() { echo -e "${RED}[ERROR]${NC} $*" >&2; }
log_step() { echo -e "${CYAN}[STEP]${NC} $*"; }
show_usage() {
cat <<EOF
Usage: sudo $SCRIPT_NAME [OPTIONS]
Adds an AI bot blocking jail to an existing Fail2ban installation.
OPTIONS:
--logpath PATH Nginx/Apache access log path (default: auto-detect)
--bantime SECS Ban duration in seconds (default: 86400 / 24 hours)
--maxretry NUM Hits before ban (default: 1)
--dry-run Show what would be done without making changes
-h, --help Show this help message
BLOCKED BOTS:
ABEvalBot Unknown operator web scraper
GPTBot OpenAI training crawler
CCBot Common Crawl (AI training datasets)
ClaudeBot Anthropic training crawler
anthropic-ai Anthropic legacy crawler
Bytespider ByteDance/TikTok crawler
TikTokSpider ByteDance/TikTok content spider
cohere-ai Cohere AI crawler
meta-externalagent Meta/Facebook content scraper
Meta-ExternalFetcher Meta content fetcher
PetalBot Huawei search/AI crawler
Amazonbot Amazon AI crawler
AI2Bot Allen Institute crawler
Ai2Bot-Dolma Allen Institute dataset crawler
YouBot You.com AI search crawler
PerplexityBot Perplexity AI indexing
Diffbot Content extraction bot
Applebot-Extended Apple AI training (not search)
Google-Extended Google AI training (not search)
MistralBot Mistral AI crawler
YandexGPTBot Yandex AI crawler
MJ12bot Majestic SEO crawler
Scrapy Scrapy-based scrapers
DataForSeoBot SEO data scraper
Timpibot AI training crawler
img2dataset Image dataset scraper
HanaleiBot Unidentified crawler (AWS, beta-stage)
SemrushBot Semrush SEO crawler
AhrefsBot Ahrefs SEO crawler
DotBot Moz SEO crawler
SERanking SE Ranking backlink/SEO crawler
trafilatura Web content extraction tool
EXAMPLES:
# Auto-detect and install
sudo $SCRIPT_NAME
# Custom log path and ban time (7 days)
sudo $SCRIPT_NAME --logpath /var/log/nginx/access.log --bantime 604800
# Preview without changes
sudo $SCRIPT_NAME --dry-run
EOF
exit 0
}
parse_args() {
while [[ $# -gt 0 ]]; do
case "$1" in
--logpath) LOGPATH="$2"; shift 2 ;;
--bantime) BANTIME="$2"; shift 2 ;;
--maxretry) MAXRETRY="$2"; shift 2 ;;
--dry-run) DRY_RUN=true; shift ;;
-h|--help) show_usage ;;
*) log_error "Unknown option: $1"; show_usage ;;
esac
done
}
check_root() {
if [[ $EUID -ne 0 ]]; then
log_error "This script must be run as root (sudo)"
exit 1
fi
}
# ============================================================================
# CHECKS
# ============================================================================
check_fail2ban() {
if ! command -v fail2ban-client &>/dev/null; then
log_error "Fail2ban is not installed"
log_error "Install it first: https://mylinux.work/guides/fail2ban-setup/"
exit 1
fi
if ! systemctl is-active --quiet fail2ban; then
log_error "Fail2ban is not running"
exit 1
fi
log_info "Fail2ban is installed and running"
}
detect_logpath() {
if [[ "$LOGPATH" != "auto" ]]; then
# Support glob patterns (e.g. /var/log/apache2/domains/*.log)
# shellcheck disable=SC2086,SC2206
local matches=( $LOGPATH )
if [[ ${#matches[@]} -eq 0 || ! -f "${matches[0]}" ]]; then
log_error "Log file not found: $LOGPATH"
exit 1
fi
log_info "Using specified log path: $LOGPATH (${#matches[@]} file(s))"
return
fi
log_step "Auto-detecting web server access log..."
# HestiaCP / VestaCP — apache domains (check first: has full access logs with user agents)
local hestia_apache=( /var/log/apache2/domains/*.log )
if [[ -f "${hestia_apache[0]:-}" ]]; then
LOGPATH="/var/log/apache2/domains/*.log"
log_info "Detected HestiaCP/VestaCP apache: $LOGPATH (${#hestia_apache[@]} file(s))"
return
fi
# HestiaCP / VestaCP — nginx domains (proxy logs only in nginx+apache mode)
local hestia_nginx=( /var/log/nginx/domains/*.log )
if [[ -f "${hestia_nginx[0]:-}" ]]; then
LOGPATH="/var/log/nginx/domains/*.log"
log_info "Detected HestiaCP/VestaCP nginx: $LOGPATH (${#hestia_nginx[@]} file(s))"
return
fi
# Nginx (standard)
if [[ -f /var/log/nginx/access.log ]]; then
LOGPATH="/var/log/nginx/access.log"
log_info "Detected nginx: $LOGPATH"
return
fi
# Apache (Debian/Ubuntu)
if [[ -f /var/log/apache2/access.log ]]; then
LOGPATH="/var/log/apache2/access.log"
log_info "Detected apache2: $LOGPATH"
return
fi
# Apache (RHEL/Rocky)
if [[ -f /var/log/httpd/access_log ]]; then
LOGPATH="/var/log/httpd/access_log"
log_info "Detected httpd: $LOGPATH"
return
fi
log_error "Could not auto-detect access log. Use --logpath to specify."
exit 1
}
# ============================================================================
# INSTALL FILTER
# ============================================================================
install_filter() {
local filter_file="/etc/fail2ban/filter.d/ai-bots.conf"
log_step "Installing filter: $filter_file"
if $DRY_RUN; then
log_info "[DRY RUN] Would create $filter_file"
echo ""
generate_filter
echo ""
return
fi
if [[ -f "$filter_file" ]]; then
log_warn "Filter already exists — backing up to ${filter_file}.bak"
cp "$filter_file" "${filter_file}.bak"
fi
generate_filter > "$filter_file"
log_info "Filter installed: $filter_file"
}
generate_filter() {
cat <<'EOF'
# Fail2ban filter to block AI scrapers and unwanted bots
# https://mylinux.work
#
# Matches common AI crawler user agents in web server access logs.
# These bots scrape content for AI model training and typically
# ignore robots.txt directives.
[Definition]
# Match AI and unwanted bot user agents in access logs
# Supports both combined and common log formats
failregex = ^<HOST> \S+ \S+ \[.*\] "\S+ \S+ \S+" \d+ \d+ "\S+" ".*(?:ABEvalBot|GPTBot|CCBot|ClaudeBot|anthropic-ai|Bytespider|TikTokSpider|cohere-ai|meta-externalagent|Meta-ExternalFetcher|PetalBot|Amazonbot|AI2Bot|Ai2Bot-Dolma|YouBot|PerplexityBot|Diffbot|Applebot-Extended|Google-Extended|MistralBot|YandexGPTBot|MJ12bot|Scrapy|DataForSeoBot|Timpibot|img2dataset|HanaleiBot|SemrushBot|AhrefsBot|DotBot|SERanking|trafilatura).*"
ignoreregex =
# Author: Phil Connor — https://mylinux.work
EOF
}
# ============================================================================
# INSTALL JAIL
# ============================================================================
install_jail() {
local jail_file="/etc/fail2ban/jail.d/ai-bots.conf"
log_step "Installing jail: $jail_file"
if $DRY_RUN; then
log_info "[DRY RUN] Would create $jail_file"
echo ""
generate_jail
echo ""
return
fi
if [[ -f "$jail_file" ]]; then
log_warn "Jail config already exists — backing up to ${jail_file}.bak"
cp "$jail_file" "${jail_file}.bak"
fi
generate_jail > "$jail_file"
log_info "Jail config installed: $jail_file"
}
generate_jail() {
cat <<EOF
# Fail2ban jail to block AI scrapers and unwanted bots
# https://mylinux.work
#
# Bans IPs on first request matching an AI bot user agent.
# These bots ignore robots.txt so we enforce it at the firewall level.
[ai-bots]
enabled = true
port = http,https
filter = ai-bots
logpath = $LOGPATH
maxretry = $MAXRETRY
bantime = $BANTIME
findtime = $BANTIME
# Author: Phil Connor — https://mylinux.work
EOF
}
# ============================================================================
# RELOAD AND VERIFY
# ============================================================================
reload_fail2ban() {
log_step "Reloading Fail2ban..."
if $DRY_RUN; then
log_info "[DRY RUN] Would reload fail2ban"
return
fi
# Test config before reloading
if ! fail2ban-client --test 2>/dev/null; then
log_warn "Config test not available — reloading directly"
fi
fail2ban-client reload
sleep 2
if systemctl is-active --quiet fail2ban; then
log_info "Fail2ban reloaded successfully"
else
log_error "Fail2ban failed to restart — check: journalctl -u fail2ban"
exit 1
fi
}
verify_jail() {
log_step "Verifying ai-bots jail..."
if $DRY_RUN; then
log_info "[DRY RUN] Would verify jail status"
return
fi
echo ""
if fail2ban-client status ai-bots 2>/dev/null; then
echo ""
log_info "AI bots jail is active and monitoring $LOGPATH"
else
log_error "Jail 'ai-bots' is not running — check: fail2ban-client status"
log_error "Debug with: fail2ban-regex $LOGPATH /etc/fail2ban/filter.d/ai-bots.conf"
exit 1
fi
}
# ============================================================================
# MAIN
# ============================================================================
main() {
parse_args "$@"
echo ""
echo "============================================"
echo " Fail2ban AI Bot Blocker v${VERSION}"
echo " https://mylinux.work"
echo "============================================"
echo ""
check_root
check_fail2ban
detect_logpath
install_filter
install_jail
reload_fail2ban
verify_jail
echo ""
echo "============================================"
echo " Setup Complete"
echo "============================================"
echo ""
echo " Jail: ai-bots"
echo " Log: $LOGPATH"
echo " Ban time: ${BANTIME}s ($(( BANTIME / 3600 ))h)"
echo " Max retry: $MAXRETRY"
echo ""
echo " Useful commands:"
echo " fail2ban-client status ai-bots"
echo " fail2ban-client set ai-bots unbanip <IP>"
echo " fail2ban-regex $LOGPATH /etc/fail2ban/filter.d/ai-bots.conf"
echo ""
}
main "$@"