Accessibility Scanner – Update Notes

Version 4.0.1

High-Impact Architectural Improvements

Minor update: to explore strange new features, seek out new optimisations...

Robust Sitemap Parsing

• Replaced fragile regex-based sitemap parsing with the fast-xml-parser library.
• Sitemap discovery is now much more reliable and resilient to different XML formats and edge cases.

Advanced Anti-Bot & Stealth Capabilities

• Dynamic Fingerprint Spoofing: The getFingerprintOverrides function now randomizes key browser properties (hardwareConcurrency, deviceMemory, WebGL renderer) on each run. This makes automated browsing harder to detect and improves evasion of sophisticated bot-detection systems.
• Proxy Rotation: The new createProxyRotator function enables the scanner to cycle through a list of proxy servers, distributing requests across multiple IP addresses. Critical for large-scale crawls, this avoids rate limiting and IP bans.

Intelligent Report Generation with Fallback

• A new fallback mechanism ensures reports are generated even when embedding many Base64 screenshots would previously cause memory errors.
• The scanner now:
  • Saves screenshots to a dedicated folder (e.g., report_name_screenshots).
  • Generates the HTML report with relative links to the saved images.
  • Includes a note explaining how to view the screenshots.
• Users can now generate usable reports for large or complex scans that would previously fail.

Structured Error Handling and Reporting

• Introduced a centralized addError function with categorized error types (CRAWL_ERROR, SCAN_ERROR, SYSTEM_ERROR).
• The HTML report now includes a much cleaner, organized “Errors” section, allowing users to quickly distinguish between crawling issues, scanning failures, and critical system problems.

New Features and Usability Enhancements

• Interactive Runtime Commands: The setupRuntimeCommands function adds a simple command-line interface. Includes the update_npm command for easy dependency updates without restarting the server.
• Improved User Feedback: Replaced the old custom spinner with ora for a more professional CLI experience during scans.
• Enhanced Cancellation Logic: executeWithRetry and crawlAndScan now properly respect scanSignal. Cancelled scans terminate retry loops and asynchronous tasks gracefully.
• Better Input Validation: JSON parsing for complex parameters like userJourneySteps and manualIssues now logs descriptive errors to simplify debugging.

Code and Dependency Refinements

• Dependencies: Added fast-xml-parser and ora, aligning with best practices.
• User Agent Randomization: In stealth mode, a random user agent from STEALTH_USER_AGENTS is selected for each scan, reducing predictability.
• Cleaner Error Sections: Reports now feature distinct sections for different error types, resulting from the improved error handling logic.