AI‑WAF

A self‑learning, Django‑friendly Web Application Firewall
with enhanced context-aware protection, rate‑limiting, anomaly detection, honeypots, UUID‑tamper protection, smart keyword learning, file‑extension probing detection, exempt path awareness, and daily retraining.

🆕 Latest Enhancements:

• ✅ Smart Keyword Filtering - Prevents blocking legitimate pages like /profile/
• ✅ Granular Reset Commands - Clear specific data types (--blacklist, --keywords, --exemptions)
• ✅ Context-Aware Learning - Only learns from suspicious requests, not legitimate site functionality
• ✅ Enhanced Configuration - AIWAF_ALLOWED_PATH_KEYWORDS and AIWAF_EXEMPT_KEYWORDS
• ✅ Comprehensive HTTP Method Validation - Blocks GET→POST-only, POST→GET-only, unsupported REST methods
• ✅ Enhanced Honeypot Protection - POST validation & 4-minute page timeout with smart reload detection
• ✅ HTTP Header Validation - Comprehensive bot detection via header analysis and quality scoring

---

🚀 Quick Installation

pip install aiwaf

⚠️ Important: Add 'aiwaf' to your Django INSTALLED_APPS to avoid setup errors.

📋 Complete Setup Guide: See INSTALLATION.md for detailed installation instructions and troubleshooting.

---

System Requirements

No GPU needed—AI-WAF runs entirely on CPU with just Python 3.8+, Django 3.2+, a single vCPU and ~512 MB RAM for small sites; for moderate production traffic you can bump to 2–4 vCPUs and 2–4 GB RAM, offload the daily detect-and-train job to a worker, and rotate logs to keep memory use bounded.

📁 Package Structure

aiwaf/
├── __init__.py
├── blacklist_manager.py
├── middleware.py
├── trainer.py                   # exposes train()
├── utils.py
├── template_tags/
│   └── aiwaf_tags.py
├── resources/
│   ├── model.pkl                # pre‑trained base model
│   └── dynamic_keywords.json    # evolves daily
├── management/
│   └── commands/
│       ├── detect_and_train.py      # `python manage.py detect_and_train`
│       ├── add_ipexemption.py       # `python manage.py add_ipexemption`
│       ├── aiwaf_reset.py           # `python manage.py aiwaf_reset`
│       └── aiwaf_logging.py         # `python manage.py aiwaf_logging`
└── LICENSE

---

🚀 Features

• IP Blocklist
  Instantly blocks suspicious IPs using Django models with real-time performance.

• Rate Limiting
  Sliding‑window blocks flooders (> AIWAF_RATE_MAX per AIWAF_RATE_WINDOW), then blacklists them.

• AI Anomaly Detection
  IsolationForest trained on:

  • Path length

• GeoIP Support
  AIWAF supports optional geo-blocking and country-level traffic statistics using a local GeoIP database.

  • Keyword hits (static + dynamic)
  • Response time
  • Status‑code index
  • Burst count
  • Total 404s

• Enhanced Dynamic Keyword Learning with Django Route Protection

  • Smart Context-Aware Learning: Only learns keywords from suspicious requests on non-existent paths
  • Automatic Django Route Extraction: Automatically excludes keywords from:
    • Valid Django URL patterns (/profile/, /admin/, /api/, etc.)
    • Django app names and model names (users, posts, categories)
    • View function names and URL namespaces
  • Unified Logic: Both trainer and middleware use identical legitimate keyword detection
  • Configuration Options:
    • AIWAF_ALLOWED_PATH_KEYWORDS - Explicitly allow certain keywords in legitimate paths
    • AIWAF_EXEMPT_KEYWORDS - Keywords that should never trigger blocking
  • Automatic Cleanup: Keywords from AIWAF_EXEMPT_PATHS are automatically removed from the database
  • False Positive Prevention: Stops learning legitimate site functionality as "malicious"
  • Inherent Malicious Detection: Middleware also blocks obviously malicious keywords (hack, exploit, attack) even if not yet learned

• File‑Extension Probing Detection
  Tracks repeated 404s on common extensions (e.g. .php, .asp) and blocks IPs.

• 🆕 HTTP Header Validation Advanced header analysis to detect bots and malicious requests:

  • Missing Required Headers - Blocks requests without User-Agent or Accept headers
  • Suspicious User-Agents - Detects curl, wget, python-requests, automated tools
  • Header Quality Scoring - Calculates realism score based on browser-standard headers
  • Legitimate Bot Whitelist - Allows Googlebot, Bingbot, and other search engines
  • Header Combination Analysis - Detects impossible combinations (HTTP/2 + old browsers)
  • Static File Exemption - Skips validation for CSS, JS, images

🛡️ Header Validation Middleware Features

The HeaderValidationMiddleware provides advanced bot detection through HTTP header analysis:

What it detects:

• Missing Headers: Requests without standard browser headers
• Suspicious User-Agents: WordPress scanners, exploit tools, basic scrapers
• Bot-like Patterns: Low header diversity, missing Accept headers
• Quality Scoring: 0-11 point system based on header completeness

What it allows:

• Legitimate Browsers: Chrome, Firefox, Safari, Edge with full headers
• Search Engine Bots: Google, Bing, DuckDuckGo, Yandex crawlers
• API Clients: Properly identified with good headers
• Static Files: CSS, JS, images (automatically exempted)

Real-world effectiveness:

✅ Blocks: WordPress scanners, exploit bots, basic scrapers
✅ Allows: Real browsers, legitimate bots, API clients
✅ Quality Score: 10/11 = Legitimate, 2/11 = Suspicious bot

Testing header validation:

# Test with curl (will be blocked - low quality headers)
curl http://yoursite.com/

# Test with browser (will be allowed - high quality headers)
# Visit site normally in Chrome/Firefox

# Check logs for header validation blocks
python manage.py aiwaf_logging --recent

• Enhanced Timing-Based Honeypot
  Advanced GET→POST timing analysis with comprehensive HTTP method validation:

  • Submit forms faster than AIWAF_MIN_FORM_TIME seconds (default: 1 second)
  • 🆕 Smart HTTP Method Validation - Comprehensive protection against method misuse:
    • Blocks GET requests to POST-only views (form endpoints, API creates)
    • Blocks POST requests to GET-only views (list pages, read-only content)
    • Blocks unsupported REST methods (PUT/DELETE to non-REST views)
    • Uses Django view analysis: class-based views, method handlers, URL patterns
  • 🆕 Page expiration after AIWAF_MAX_PAGE_TIME (4 minutes) with smart reload

• UUID Tampering Protection
  Blocks guessed or invalid UUIDs that don't resolve to real models.

• Built-in Request Logger
  Optional middleware logger that captures requests to Django models:

  • Automatic fallback when main access logs unavailable
  • Real-time storage in database for instant access
  • Captures response times for better anomaly detection
  • Zero configuration - works out of the box

• Blocked Request Debug Logging
  Optional debug logs that explain why a request was blocked:

  • Reason included (keyword, flood pattern, AI anomaly, header validation, etc.)
  • Request context (IP, method, path, user agent)
  • Disabled by default - enable via Django LOGGING

  Example settings.py:

  LOGGING = {
      "version": 1,
      "disable_existing_loggers": False,
      "handlers": {
          "console": {"class": "logging.StreamHandler"},
      },
      "loggers": {
          "aiwaf.middleware": {"handlers": ["console"], "level": "DEBUG"},
      },
  }

• Blocked Request Responses By default, AI‑WAF raises PermissionDenied("blocked") when a request is blocked, allowing Django to render a standard 403 page. For API clients that need JSON, add JsonExceptionMiddleware near the top of your MIDDLEWARE list; it will translate PermissionDenied into a JSON 403 response when request.content_type == "application/json".

• Smart Training System
  AI trainer automatically uses the best available data source:

  • Primary: Configured access log files (AIWAF_ACCESS_LOG)
  • Fallback: Database RequestLog model when files unavailable
  • Seamless switching between data sources
  • Enhanced compatibility with exemption system
  • Minimum log thresholds: AI training requires AIWAF_MIN_AI_LOGS (default 10,000); fewer logs falls back to keyword-only, which still requires AIWAF_MIN_TRAIN_LOGS (default 50)

Exempt Path & IP Awareness

Exempt Paths: AI‑WAF automatically exempts common login paths (/admin/, /login/, /accounts/login/, etc.) from all blocking mechanisms. You can add additional exempt paths in your Django settings.py:

AIWAF_EXEMPT_PATHS = [
    "/api/webhooks/",
    "/health/",
    "/special-endpoint/",
]

You can also store exempt paths in the database (no deploy needed):

python manage.py aiwaf_pathshell

Or add directly:

python manage.py add_pathexemption /myapp/api/ --reason "API traffic"

AIWAF Path Shell Commands:

ls                     # list routes at current level
cd <index|name>        # enter a route segment
up / cd ..             # go up one level
pwd                    # show current path prefix
exempt <index|name|.>  # add exemption for selection or current path
exit                   # quit

Exempt Path & IP Awareness

Exempt Paths: AI‑WAF automatically exempts common login paths (/admin/, /login/, /accounts/login/, etc.) from all blocking mechanisms. You can add additional exempt paths in your Django settings.py:

AIWAF_EXEMPT_PATHS = [
    "/api/webhooks/",
    "/health/",
    "/special-endpoint/",
]

You can also store exempt paths in the database (no deploy needed):

python manage.py aiwaf_pathshell

Or add directly:

python manage.py add_pathexemption /myapp/api/ --reason "API traffic"

AIWAF Path Shell Commands:

ls                     # list routes at current level
cd <index|name>        # enter a route segment
up / cd ..             # go up one level
pwd                    # show current path prefix
exempt <index|name|.>  # add exemption for selection or current path
exit                   # quit

Exempt Views (Decorator): Use the @aiwaf_exempt decorator to exempt specific views from all AI-WAF protection:

from aiwaf.decorators import aiwaf_exempt
from django.http import JsonResponse

@aiwaf_exempt
def my_api_view(request):
    """This view will be exempt from all AI-WAF protection"""
    return JsonResponse({"status": "success"})

# Works with class-based views too
@aiwaf_exempt
class MyAPIView(View):
    def get(self, request):
        return JsonResponse({"method": "GET"})

All exempt paths and views are:

• Skipped from keyword learning
• Immune to AI blocking
• Ignored in log training
• Cleaned from DynamicKeyword model automatically

Exempt IPs: You can exempt specific IP addresses from all blocking and blacklisting logic. Exempted IPs will:

• Never be added to the blacklist (even if they trigger rules)
• Be automatically removed from the blacklist during retraining
• Bypass all block/deny logic in middleware

Managing Exempt IPs

Add an IP to the exemption list using the management command:

python manage.py add_ipexemption <ip-address> --reason "optional reason"

Resetting AI-WAF

The aiwaf_reset command provides granular control for clearing different types of data:

# Clear everything (default - backward compatible)
python manage.py aiwaf_reset

# Clear everything without confirmation prompt
python manage.py aiwaf_reset --confirm

# 🆕 GRANULAR CONTROL - Clear specific data types
python manage.py aiwaf_reset --blacklist      # Clear only blocked IPs
python manage.py aiwaf_reset --exemptions     # Clear only exempted IPs  
python manage.py aiwaf_reset --keywords       # Clear only learned keywords

# 🔧 COMBINE OPTIONS - Mix and match as needed
python manage.py aiwaf_reset --blacklist --keywords      # Keep exemptions
python manage.py aiwaf_reset --exemptions --keywords     # Keep blacklist
python manage.py aiwaf_reset --blacklist --exemptions    # Keep keywords

# 🚀 COMMON USE CASES
# Fix false positive keywords (like "profile" blocking legitimate pages)
python manage.py aiwaf_reset --keywords --confirm
python manage.py detect_and_train  # Retrain with enhanced filtering

# Clear blocked IPs but preserve exemptions and learning
python manage.py aiwaf_reset --blacklist --confirm

# Legacy support (still works for backward compatibility)
python manage.py aiwaf_reset --blacklist-only    # Legacy: blacklist only
python manage.py aiwaf_reset --exemptions-only   # Legacy: exemptions only

Enhanced Feedback:

$ python manage.py aiwaf_reset --keywords
🔧 AI-WAF Reset: Clear 15 learned keywords
Are you sure you want to proceed? [y/N]: y
✅ Reset complete: Deleted 15 learned keywords

This will ensure the IP is never blocked by AI‑WAF. You can also manage exemptions via the Django admin interface.

• Daily Retraining
  Reads rotated logs, auto‑blocks 404 floods, retrains the IsolationForest, updates model.pkl, and evolves the keyword DB. If GeoIP is enabled, it also prints a country summary for anomalous IPs.

---

⚙️ Configuration (settings.py)

INSTALLED_APPS += ["aiwaf"]

Database Setup

After adding aiwaf to your INSTALLED_APPS, run the following to create the necessary tables:

python manage.py makemigrations aiwaf
python manage.py migrate

---

Required

AIWAF_ACCESS_LOG = "/var/log/nginx/access.log"

---

Database Models

AI-WAF uses Django models for real-time, high-performance storage:

# All data is stored in Django models - no configuration needed
# Tables created automatically with migrations:
# - aiwaf_blacklistentry     # Blocked IP addresses
# - aiwaf_ipexemption        # Exempt IP addresses  
# - aiwaf_exemptpath         # Exempt path prefixes
# - aiwaf_dynamickeyword     # Dynamic keywords with counts
# - aiwaf_featuresample      # Feature samples for ML training
# - aiwaf_requestlog         # Request logs (if middleware logging enabled)

Benefits of Django Models:

• ⚡ Real-time performance - No file I/O bottlenecks
• 🔄 Instant updates - Changes visible immediately across all processes
• 🚀 Better concurrency - No file locking issues
• 📊 Rich querying - Use Django ORM for complex operations
• 🔍 Admin integration - View/manage data through Django admin

Database Setup:

# Create and apply migrations
python manage.py makemigrations aiwaf
python manage.py migrate aiwaf

---

Built-in Request Logger (Optional)

Enable AI-WAF's built-in request logger as a fallback when main access logs aren't available:

# Enable middleware logging
AIWAF_MIDDLEWARE_LOGGING = True                    # Enable/disable logging
AIWAF_MIDDLEWARE_LOG = "aiwaf_requests.log"        # Optional log file name
AIWAF_MIDDLEWARE_CSV = True                        # Write CSV log file (default: True)
AIWAF_MIDDLEWARE_DB = True                         # Write RequestLog entries (default: True)
AIWAF_USE_RUST = False                             # Use Rust backend for header validation

Then add middleware to MIDDLEWARE list:

MIDDLEWARE = [
    # ... your existing middleware ...
    'aiwaf.middleware_logger.AIWAFLoggerMiddleware',  # Add near the end
]

Manage middleware logging:

python manage.py aiwaf_logging --status    # Check logging status
python manage.py aiwaf_logging --enable    # Show setup instructions  
python manage.py aiwaf_logging --clear     # Clear log files

Benefits:

• Automatic fallback when AIWAF_ACCESS_LOG unavailable
• CSV or database storage with precise timestamps and response times
• Zero configuration - trainer automatically detects and uses model logs
• Lightweight - fails silently to avoid breaking your application

If you want the trainer to use the CSV log file, point AIWAF_ACCESS_LOG at the CSV path (e.g., aiwaf_requests.csv).

---

Optional Rust Backend (Header Validation)

When AIWAF_USE_RUST = True, AI-WAF uses a Rust backend (pyo3/maturin) for header validation. If the Rust module is not available, it automatically falls back to the Python implementation.

Default install does not require Rust:

pip install aiwaf

Install the Rust extension from the separate package:

pip install "aiwaf[rust]"

This pulls the separately-released Rust extension package (aiwaf-rust).

Enable in settings:

AIWAF_MIDDLEWARE_CSV = True
AIWAF_USE_RUST = True

---

Optional (defaults shown)

AIWAF_MODEL_PATH         = BASE_DIR / "aiwaf" / "resources" / "model.pkl"
AIWAF_MODEL_STORAGE      = "file"    # file | db | cache
AIWAF_MODEL_CACHE_KEY    = "aiwaf:model"
AIWAF_MODEL_CACHE_TIMEOUT = None     # seconds; None for no expiry
AIWAF_MODEL_STORAGE_FALLBACK = True  # fallback to file when db/cache unavailable
AIWAF_MIN_FORM_TIME      = 1.0        # minimum seconds between GET and POST
AIWAF_MAX_PAGE_TIME      = 240        # maximum page age before requiring reload (4 minutes)
AIWAF_AI_CONTAMINATION   = 0.05       # AI anomaly detection sensitivity (5%)
AIWAF_MIN_AI_LOGS        = 10000      # minimum log lines for AI training
AIWAF_MIN_TRAIN_LOGS     = 50         # minimum log lines for keyword training
AIWAF_FORCE_AI_TRAINING  = False      # override AIWAF_MIN_AI_LOGS gate
AIWAF_RATE_WINDOW        = 10         # seconds
AIWAF_RATE_MAX           = 20         # max requests per window
AIWAF_RATE_FLOOD         = 10         # flood threshold
AIWAF_WINDOW_SECONDS     = 60         # anomaly detection window
AIWAF_FILE_EXTENSIONS    = [".php", ".asp", ".jsp"]

# Geo-blocking (optional, requires aiwaf[geoblock])
AIWAF_GEO_BLOCK_ENABLED  = False
AIWAF_GEOIP_DB_PATH      = "aiwaf/geolock/ipinfo_lite.mmdb"
AIWAF_GEO_BLOCK_COUNTRIES = ["CN", "RU"]
AIWAF_GEO_ALLOW_COUNTRIES = []        # If set, only these countries are allowed
AIWAF_GEO_CACHE_SECONDS  = 3600
AIWAF_GEO_CACHE_PREFIX   = "aiwaf:geo:"
AIWAF_EXEMPT_PATHS = [          # optional but highly recommended
    "/favicon.ico",
    "/robots.txt",
    "/static/",
    "/media/",
    "/health/",
]

# 🆕 ENHANCED KEYWORD FILTERING OPTIONS
AIWAF_ALLOWED_PATH_KEYWORDS = [  # Keywords allowed in legitimate paths
    "profile", "user", "account", "settings", "dashboard",
    "admin", "api", "auth", "search", "contact", "about",
    # Add your site-specific legitimate keywords
    "buddycraft", "sc2", "starcraft",  # Example: gaming site keywords
]

AIWAF_EXEMPT_KEYWORDS = [        # Keywords that never trigger blocking
    "api", "webhook", "health", "static", "media",
    "upload", "download", "backup", "profile"
]

AIWAF_DYNAMIC_TOP_N = 10        # Number of dynamic keywords to learn (default: 10)

Note: You no longer need to define AIWAF_MALICIOUS_KEYWORDS or AIWAF_STATUS_CODES — they evolve dynamically.

Model storage options:

• file (default) writes to AIWAF_MODEL_PATH
• db stores the model in the AIModelArtifact table (run migrations)
• cache stores the model in your Django cache backend

Installation Modes

Full install (default) includes AI training and GeoIP support:

pip install aiwaf

Light install (manual deps only):

pip install aiwaf --no-deps
pip install "Django>=3.2" "requests>=2.25.0"

Geo-blocking uses the bundled .mmdb file by default. Set AIWAF_GEOIP_DB_PATH to override.

GeoBlock Middleware: Enable the middleware and the feature flag:

AIWAF_GEO_BLOCK_ENABLED = True

MIDDLEWARE = [
    "aiwaf.middleware.JsonExceptionMiddleware",   # Optional: JSON error responses for API clients
    "aiwaf.middleware.GeoBlockMiddleware",
    # ... other AI-WAF middleware ...
]

Acknowledgements

Geo-blocking functionality in AIWAF relies on the IPinfo MMDB for IP-to-country mapping.
Thanks to IPinfo for providing a reliable GeoIP database.

Dynamic country blocking (database-backed):

python manage.py geo_block_country list
python manage.py geo_block_country add US
python manage.py geo_block_country remove US

Path-Specific Rules

Use path rules to selectively disable middleware or override settings without full exemptions:

AIWAF_SETTINGS = {
  "PATH_RULES": [
    {
      "PREFIX": "/myapp/api/",
      "DISABLE": ["HeaderValidationMiddleware"],
      "RATE_LIMIT": {"WINDOW": 60, "MAX": 2000},
    },
    {
      "PREFIX": "/myapp/",
      "RATE_LIMIT": {"WINDOW": 60, "MAX": 200},
    },
  ]
}

Each middleware checks request.path, computes the effective policy, then applies or skips accordingly.

Define PATH_RULES in your Django settings file (e.g. settings.py) under AIWAF_SETTINGS.

Legacy AIWAF_SETTINGS Compatibility

If you already use the nested AIWAF_SETTINGS dict, AI-WAF will map common keys into the flat AIWAF_* settings at startup (without overriding explicit AIWAF_* values). Supported mappings include RATE_LIMITING, EXEMPTIONS.PATHS, IP_BLOCKING.ENABLED, KEYWORD_DETECTION (custom patterns + sensitivity), and LOGGING.ENABLED.

---

🧱 Middleware Setup

Add in this order to your MIDDLEWARE list:

MIDDLEWARE = [
    "aiwaf.middleware.JsonExceptionMiddleware",   # Optional: JSON error responses for API clients
    "aiwaf.middleware.GeoBlockMiddleware",
    "aiwaf.middleware.IPAndKeywordBlockMiddleware",
    "aiwaf.middleware.RateLimitMiddleware", 
    "aiwaf.middleware.AIAnomalyMiddleware",
    "aiwaf.middleware.HoneypotTimingMiddleware",
    "aiwaf.middleware.UUIDTamperMiddleware",
    # ... other middleware ...
    "aiwaf.middleware_logger.AIWAFLoggerMiddleware",  # Optional: Add if using built-in logger
]

⚠️ Order matters! AI-WAF protection middleware should come early. The logger middleware should come near the end to capture final response data. JSON APIs: If you want JSON error bodies on PermissionDenied, add JsonExceptionMiddleware near the top so it runs last during exception handling.

UUIDTamperMiddleware behavior:

• Only checks models in the view's app that have UUID primary keys or unique UUID fields.
• If an app has no such models, the middleware is a no-op for that request.

Troubleshooting Middleware Errors

Error: Module "aiwaf.middleware" does not define a "UUIDTamperMiddleware" attribute/class

Solutions:

1. Update AI-WAF to latest version:

   pip install --upgrade aiwaf

2. Run diagnostic commands:

   # Quick debug script (from AI-WAF directory)
   python debug_aiwaf.py
   
   # Django management command  
   python manage.py aiwaf_diagnose

3. Check available middleware classes:

   # In Django shell: python manage.py shell
   import aiwaf.middleware
   print(dir(aiwaf.middleware))

4. Verify AI-WAF is in INSTALLED_APPS:

   # In settings.py
   INSTALLED_APPS = [
       # ... other apps ...
       'aiwaf',  # Must be included
   ]

5. Use minimal middleware setup if needed:

MIDDLEWARE = [ # ... your existing middleware ... "aiwaf.middleware.JsonExceptionMiddleware", # Optional: JSON error responses for API clients "aiwaf.middleware.IPAndKeywordBlockMiddleware", # Core protection "aiwaf.middleware.RateLimitMiddleware", # Rate limiting
"aiwaf.middleware.AIAnomalyMiddleware", # AI detection ]

**Common Issues:**
- **AppRegistryNotReady Error**: Fixed in v0.1.9.0.1 - update with `pip install --upgrade aiwaf`
- **Scikit-learn Version Warnings**: Fixed in v0.1.9.0.3 - regenerate model with `python manage.py regenerate_model`
- Missing Django: `pip install Django`
- Old AI-WAF version: `pip install --upgrade aiwaf`
- Missing migrations: `python manage.py migrate`
- Import errors: Check `INSTALLED_APPS` includes `'aiwaf'`


---

##  Running Detection & Training

```bash
python manage.py detect_and_train

What happens:

1. Read access logs (incl. rotated or gzipped) OR AI-WAF middleware model logs
2. Auto‑block IPs with ≥ 6 total 404s
3. Extract features & train IsolationForest
4. Save model.pkl with current scikit-learn version

Model Regeneration

If you see scikit-learn version warnings, regenerate the model:

# Quick model regeneration (recommended)
python manage.py regenerate_model

# Full retraining with fresh data
python manage.py detect_and_train

Benefits:

• ✅ Eliminates version compatibility warnings
• ✅ Uses current scikit-learn optimizations
• ✅ Maintains same protection level

4. Save model.pkl
5. Extract top 10 dynamic keywords from 4xx/5xx
6. Remove any keywords associated with newly exempt paths

Note: If main access log (AIWAF_ACCESS_LOG) is unavailable, trainer automatically falls back to AI-WAF middleware model logs.

---

🧠 How It Works

---

##  Running Detection & Training

```bash
python manage.py detect_and_train

What happens:

1. Read access logs (incl. rotated or gzipped)
2. Auto‑block IPs with ≥ 6 total 404s
3. Extract features & train IsolationForest
4. Save model.pkl
5. Extract top 10 dynamic keywords from 4xx/5xx
6. Remove any keywords associated with newly exempt paths

---

🔧 Troubleshooting

Legitimate Pages Being Blocked

Problem: Users can't access legitimate pages like /en/profile/ due to keyword blocking.

Cause: AIWAF learned legitimate keywords (like "profile") as suspicious from previous traffic.

Solution:

# 1. Clear problematic learned keywords
python manage.py aiwaf_reset --keywords --confirm

# 2. Add legitimate keywords to settings
# In settings.py:
AIWAF_ALLOWED_PATH_KEYWORDS = [
    "profile", "user", "account", "dashboard",
    # Add your site-specific keywords
]

# 3. Retrain with enhanced filtering (won't learn legitimate keywords)
python manage.py detect_and_train

# 4. Test - legitimate pages should now work!

Preventing Future False Positives

Configure AIWAF to recognize your site's legitimate keywords:

# settings.py
AIWAF_ALLOWED_PATH_KEYWORDS = [
    # Common legitimate keywords
    "profile", "user", "account", "settings", "dashboard",
    "admin", "search", "contact", "about", "help",
    
    # Your site-specific keywords
    "buddycraft", "sc2", "starcraft",  # Gaming site example
    "shop", "cart", "checkout",        # E-commerce example  
    "blog", "article", "news",         # Content site example
]

Reset Command Options

# Clear everything (safest for troubleshooting)
python manage.py aiwaf_reset --confirm

# Clear only problematic keywords
python manage.py aiwaf_reset --keywords --confirm

# Clear blocked IPs but keep exemptions
python manage.py aiwaf_reset --blacklist --confirm

---

🧠 How It Works

Middleware	Purpose
GeoBlockMiddleware	Blocks traffic by country based on GeoIP database
IPAndKeywordBlockMiddleware	Blocks requests from known blacklisted IPs and Keywords
RateLimitMiddleware	Enforces burst & flood thresholds
AIAnomalyMiddleware	ML‑driven behavior analysis + block on anomaly
HoneypotTimingMiddleware	Enhanced bot detection: GET→POST timing, POST validation, page timeouts
UUIDTamperMiddleware	Blocks guessed/nonexistent UUIDs across models with UUID PKs or unique UUID fields in an app (no-op if none)
HeaderValidationMiddleware	Blocks suspicious header patterns and low‑quality user agents
AIWAFLoggerMiddleware	Optional request logger for model training and analysis

🍯 Enhanced Honeypot Protection

The HoneypotTimingMiddleware now includes advanced bot detection capabilities:

🚫 Smart POST Request Validation

• Analyzes Django views to determine actual allowed HTTP methods
• Intelligent detection of GET-only vs POST-capable views
• Example: POST to view with http_method_names = ['get'] → PermissionDenied (403)

⏰ Page Timeout with Smart Reload

• 4-minute page expiration prevents stale session attacks
• HTTP 409 response with reload instructions instead of immediate blocking
• CSRF protection by forcing fresh page loads for old sessions

👥 Authenticated Session Exemption

• Skips timing checks when the request belongs to an authenticated Django session
• Prevents false positives for offices or VPNs where multiple users share an IP
• Still enforces method validation and blacklist logic—only the timing rule is bypassed

# Configuration
AIWAF_MIN_FORM_TIME = 1.0     # Minimum form submission time
AIWAF_MAX_PAGE_TIME = 240     # Page timeout (4 minutes)

Timeline Example:

12:00:00 - GET /contact/   ✅ Page loaded
12:02:00 - POST /contact/  ✅ Valid submission (2 minutes)
12:04:30 - POST /contact/  ❌ 409 Conflict (page expired, reload required)

---

Sponsors

This project is proudly supported by:

DigitalOcean provides the cloud infrastructure that powers AIWAF development.

---

License

This project is licensed under the MIT License. See the LICENSE file for details.

---