- Add headless UCI option to use mitmdump instead of mitmweb
- Enable headless by default for WAF (mitmproxy-in) instance
- Increase default memory limit from 256MB to 2GB
- Fix LXC config generation to always recreate on service start
- Fix rootfs check path (/usr/local/bin not /usr/bin)
- Use exec in startup script for proper foreground execution
Headless mode runs mitmdump (CLI) instead of mitmweb (web UI),
saving ~3.3GB RAM while maintaining full WAF functionality.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
WAF Dashboard:
- Use cached bans from cron (waf-stats-update) instead of slow cscli
- Fixes "Failed to load bans" timeout issue
DPI Dual-Stream:
- Add LAN Flow Analysis card showing active clients, destinations, protocols
- LAN passive flow analysis was working but not displayed
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Port 8889 conflicts with avatar-tap Streamlit service.
Updated mitmproxy-in instance to use port 8890 for HAProxy WAF routing.
Changes:
- UCI config: proxy_port and listen_port now default to 8890
- mitmproxyctl: Updated fallback defaults and documentation
- README: Updated architecture diagrams with correct port
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
secubox-users:
- Add gitea and jellyfin to supported services list
- Add create/update/delete handlers for gitea (via API) and jellyfin
- Update CLI help and status display to include new services
luci-app-secubox-users:
- Add jellyfin service checkbox and badge in frontend
- Update RPCD handler to check jellyfin service status
mitmproxy routing fix:
- nextcloudctl: Use host LAN IP instead of 127.0.0.1 for WAF routes
(mitmproxy runs in container, can't reach host's localhost)
- metablogizerctl: Same fix for mitmproxy route registration
- mitmproxyctl: Fix sync_metablogizer_routes to use host IP
This fixes 502/403 errors when accessing services through HAProxy->mitmproxy
because the mitmproxy container couldn't route to 127.0.0.1 on the host.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Changes:
- Enable WAF auto-ban by default (sensitivity: moderate, min_severity: high)
- Add whitelist for common safe IPs (localhost, router)
- Add browser cache busting via version parameter in CSS loads
- Document deployment scripts in secubox-tools/README.md
- Create CVE Layer 7 architecture documentation
WAF auto-ban now active with:
- 3 threats within 5 minutes triggers ban
- 4-hour ban duration
- Critical CVEs (Log4Shell, SQLi, CMDi) ban immediately
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add centralized route registry (`secubox-route`) in secubox-core to eliminate
route management duplication across metablogizerctl, streamlitctl, and
mitmproxyctl.
New features:
- `/etc/config/secubox-routes` - UCI config for central route registry
- `/usr/sbin/secubox-route` - CLI for route management (add, remove, sync)
- Import routes from HAProxy, MetaBlogizer, Streamlit with source tracking
- Auto-sync to all mitmproxy instances on route changes
- Skip wildcard domains and LuCI (port 8081) routes
Updated services to use centralized registry:
- metablogizerctl: Use secubox-route add instead of mitmproxyctl sync
- streamlitctl: Use secubox-route add with domain/port params
- peertubectl: Use secubox-route add for emancipation
- vhost-manager/mitmproxy.sh: Prefer secubox-route when available
- mitmproxyctl: Delegate to secubox-route import-all for sync-routes
This prevents route mixups between services and provides a single
source of truth for all WAF routing configuration.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Skip wildcard domains (starting with '.') in sync-routes to allow
unknown subdomains to show "WAF SAYS NO" 404 page instead of blog
- Fix log_info() to output to stderr to prevent JSON corruption in
sync-routes when log messages mixed with JSON fragments
- Escape CSS curly braces in NOT_FOUND_HTML for Python .format()
compatibility (fixes KeyError: 'box-sizing')
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- secubox_analytics.py: Add 'strict' as alias for 'aggressive' in autoban
- Fix waf_bypass false positives on LuCI static resources
- Root cause: different analytics versions across mitmproxy instances
- Update HISTORY.md with OpenClaw Gemini fix and WAF tuning
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
mitmproxy haproxy_router.py:
- Return 404 instead of routing to LuCI (8081) for missing routes
- Block any routes that point to port 8081
- Add "WAF Says NO" themed 404 page with multi-layer WAF visual
HAProxy (deployed on router):
- Configure end_of_internet backend with custom errorfiles
- Add "End of Internet" themed error pages for 5xx errors
- Patched haproxyctl to include errorfile directives
New package: secubox-app-openclaw
- Personal AI assistant integration for SecuBox
- Supports Anthropic Claude, OpenAI, and Ollama providers
- Chat integrations (Telegram, Discord, Slack)
- Email/calendar automation support
- CLI tool: openclawctl
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add route management commands:
- `mitmproxyctl route list` - List all routes
- `mitmproxyctl route add <domain> <ip> <port>` - Add manual route
- `mitmproxyctl route remove <domain>` - Remove route
- `mitmproxyctl route check` - Check for missing routes
- Improve sync-routes to scan MetaBlogizer and Streamlit services:
- Auto-detect enabled MetaBlogizer sites and add routes
- Auto-detect enabled Streamlit instances with matching vhosts
- Warn about mitmproxy_inspector vhosts with missing routes
- Update routes config with 188 routes
This fixes the issue where services using mitmproxy_inspector backend
would fall back to default because their routes were not configured.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Changed _check_interval from 10 to 1 to ensure new routes are picked up
immediately when the haproxy-routes.json file is updated.
This fixes the quick publish flow where new sites weren't accessible
immediately because mitmproxy only checked for route changes every
10 requests.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
haproxyctl:
- Copy generated config to /etc/haproxy/ inside container before reload
- HAProxy reads from /etc/haproxy/haproxy.cfg, not /opt/haproxy/config/
mitmproxy haproxy_router.py:
- Save original Host header before setting backend destination
- Restore Host header after routing to preserve it for backend validation
- Fixes PeerTube OAuth and other apps that validate Host header
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Support both "*.domain" and ".domain" wildcard formats in haproxy_router.py
- Sort wildcards by length (longest first) for correct specificity matching
- Add auto-reload: check routes file mtime every 10 requests
- Update metablogizerctl to use mitmproxyctl sync-routes
Also fix luci-app-wazuh api.js to use baseclass.extend
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The sync-routes command was only writing to the default /srv/mitmproxy
path. Now copies haproxy-routes.json to all configured instances
(mitmproxy-in, mitmproxy-out) so the HAProxy router addon can
properly route traffic.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The reload-autoban command was only writing to the default data path,
missing the mitmproxy-in and mitmproxy-out instances. Now iterates
over all configured instances to ensure autoban.json is updated
everywhere.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Remove 'mozilla/5.0' from BOT_SIGNATURES - was flagging ALL modern
browsers as bots since this is the standard UA prefix
- Fix suspicious UA detection - no longer flags normal browsers
- Increase CrowdSec bruteforce threshold from 5/30s to 10/60s to reduce
false positives from normal login flows
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
SSRF detection was triggering on any request to internal IPs
(192.168.x.x, 10.x.x.x, etc.) because it was checking the target
URL itself. Now only checks query parameters and request body for
SSRF patterns, which is where actual SSRF attacks occur.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Allow disabling HAProxy router mode per-instance via UCI option
`mitmproxy.<instance>.haproxy_router_enabled`. This prevents port
conflicts when running multiple mitmproxy instances (e.g., mitmproxy-out
on 8888 and mitmproxy-in on 8889) where only the inbound instance
needs HAProxy router mode.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add TRUSTED_PATH_PREFIXES for LuCI, ubus, and CGI paths
- Fix moderate mode to always require threshold (3 attempts in 5 min)
instead of immediate ban on critical threats
- Add WireGuard endpoint whitelist support to prevent VPN peer bans
- New script: mitmproxy-sync-wg-endpoints extracts peer IPs from UCI
- Bump version to v2.4
Prevents accidental bans from legitimate external LuCI login attempts.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Track requests, threats, protocols per subdomain
- Record HTTP methods, status codes, top URIs, countries
- New RPCD method: subdomain_metrics
- Metrics auto-saved to /tmp/secubox-subdomain-metrics.json
- Add wan_setup/wan_clear to ACL write permissions
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add extract_zip_flatten() to Streamlit RPCD for nested ZIP handling
- Add bot whitelist to mitmproxy WAF (Facebook, Google, Bing crawlers)
- Skip threat detection for whitelisted legitimate crawlers
- Track Fabricator app and stats evolution in HISTORY.md
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
WAF Auto-ban Features:
- Three sensitivity levels: aggressive, moderate, permissive
- Aggressive: Immediate ban on first critical threat
- Moderate: Ban after 3 attempts in 5 minutes (default)
- Permissive: Ban after 5 attempts in 1 hour
- Attempt tracking with configurable thresholds
Critical threats (immediate in aggressive/moderate):
- CVE exploits, SQL injection, Command injection
- XXE, Log4Shell, SSTI attacks
CrowdSec Integration:
- Auto-ban requests written to /srv/mitmproxy/autoban-requests.log
- Cron job processes bans every minute via mitmproxyctl
- Bans sent to CrowdSec for network-wide enforcement
New Commands:
- mitmproxyctl process-autoban: Process pending bans
- mitmproxyctl reload-autoban: Reload config after UCI changes
CrowdSec Dashboard:
- Added ban button to alerts page
- Modal confirmation with 24h ban duration
- Real-time banned IP tracking
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Content-Type based CVE detection must happen before SSRF patterns
to avoid false positives when routing through localhost.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
LAN transparent mode now requires explicit opt-in via transparent.enabled
to prevent HTTPS certificate errors for LAN clients.
Changes:
- mitmproxyctl: Check transparent_enabled before setting up LAN firewall rules
- LuCI settings: Add warning about certificate requirements for LAN mode
- Default config already has transparent.enabled='0'
WAN protection mode remains active for incoming threat detection.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add WAF-like functionality to mitmproxy for protecting services exposed
to the internet. Incoming WAN traffic is redirected through mitmproxy
for threat detection before reaching backend services.
Features:
- WAN protection mode with nftables rules for incoming traffic
- Enhanced bot scanner detection with 50+ scanner signatures
- Behavioral detection for config/admin/backup/shell hunting
- CrowdSec integration with new scenarios for bot scanners
- LuCI interface for WAN protection configuration
- DPI mirror mode support (secondary feature)
New CrowdSec scenarios:
- secubox/mitmproxy-botscan: Detect automated reconnaissance
- secubox/mitmproxy-shell-hunter: Detect shell/backdoor hunting
- secubox/mitmproxy-config-hunter: Detect credential file hunting
- secubox/mitmproxy-suspicious-ua: Detect suspicious user agents
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The mitmproxy container mounts /srv/mitmproxy as /data, so the GeoIP
database path must use /data/GeoLite2-Country.mmdb for the addon to
find it. This enables proper country detection for external IPs in
threat logs, allowing CrowdSec SSRF scenarios to correctly identify
and ban foreign attackers.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Change analytics addon to write threats to /data/threats.log (bind-mounted to host)
- Add CrowdSec acquisition config to read from /srv/mitmproxy/threats.log
- Add parser for mitmproxy JSON threat logs with source_ip in Meta
- Add scenarios for web attacks, scanners, SSRF, and CVE exploits
- Update RPCD to read alerts from host-visible path without lxc-attach
This enables automatic IP banning when mitmproxy detects:
- SQL injection, XSS, command injection (capacity: 3, ban: 15m)
- Path traversal, XXE, LDAP injection, Log4Shell
- Aggressive web scanning (capacity: 10, ban: 10m)
- SSRF attempts from external IPs (capacity: 5, ban: 10m)
- Known CVE exploits (immediate trigger, ban: 30m)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The sync-routes command was failing to generate routes for most vhosts due to:
- Subshell bug: pipe in while loop caused variable changes to be lost
- Only supported old-style backends (inline .server field)
- Did not support new-style backends with separate =server sections
Changes:
- Rewrite sync-routes to avoid subshell by using temp file
- Add support for both backend styles (inline and separate server sections)
- Use original_backend field when vhosts are in inspection mode
- Skip luci/fallback/mitmproxy_inspector backends in route generation
Now properly generates 13+ routes for HAProxy backend inspection.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add HAProxy → mitmproxy → Backend inspection chain for filtering
all vhost traffic through mitmproxy with threat detection
- Add haproxy_router.py addon for Host-based request routing
- Add mitmproxyctl commands: sync-routes, haproxy-enable, haproxy-disable
- Add auth token to status response for Web UI auto-authentication
- Add HAProxy Backend Inspection section to LuCI status page with
enable/disable/sync controls
- Add HAProxy Router settings section to LuCI settings page
- LXC container now supports dual-port mode (8888 + 8889 for HAProxy)
- Token displayed with copy button in dashboard
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add get_web_token to RPCD ACL permissions (was missing, causing 403)
- Add fallback token retrieval from container via lxc-attach
- Improve token capture regex to support alphanumeric tokens
- Fix startup script with background process + tee for reliable capture
- Add IP forwarding enablement for transparent proxy mode
- Fix bypass rule for traffic destined to router itself
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add PYTHONUNBUFFERED=1 to ensure mitmweb output is not buffered
- Use inline while loop to capture authentication token from startup output
- Fix RPCD backend to read token from correct path ($DATA_DIR/.mitmproxy_token)
- Add proper shell detection and symlink creation in Docker rootfs extraction
- Remove unnecessary exec in pipeline that prevented output capture
The mitmweb authentication token is now properly captured and available
to the LuCI Web UI view for iframe embedding.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The previous pipe approach didn't work because the while loop
runs in a subshell. Now using a background job to poll the log
file for the token while tee outputs to both console and log.
Bump release to r13.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add get_web_token RPCD method to retrieve auth token
- Create webui.js view that embeds mitmweb in an iframe
- Capture auth token at startup and save to file
- Add Web UI navigation to all mitmproxy views
- Fix PATH for /usr/local/bin in Docker image
- Change default port from 8080 to 8888 (avoid CrowdSec conflict)
secubox-app-mitmproxy: bump to r12
luci-app-mitmproxy: bump to r2
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Recent mitmproxy versions require web authentication by default.
Disable it with --set web_password= for easier LAN access.
Bump release to r11.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
