Improved filtering for Amazon and Tumblr

[1.2.3] - 2026-02-27

Added

2257 Compliance Detection

• New detector module: js/detectors/2257-compliance-detector.js
• Detects 18 U.S.C. § 2257 compliance statements in anchor tags to identify adult content sites
• Rule-based matching with 6 detection patterns:
  • compliance_statement: 2257 + compliance + statement
  • custodian_records: 2257 + custodian + records
  • exempt_exemption: 2257 + exempt/exemption (union)
  • usc_reference: 18 + USC/U.S.C. (union) + 2257
  • record_keeping: 2257 + record-keeping/record keeping (union)
  • cfr_reference: 28 + CFR/C.F.R. (union) + 75
• Only evaluates same-domain links (link href hostname must match current page hostname) to eliminate cross-domain false positives
• Page-level whitelist skips scanning entirely on known safe domains: .gov, .edu, .law, wikipedia.org, justia.com, aclu.org, oyez.org, eff.org, duckduckgo.com, google.com, bing.com, yahoo.com, chanrobles.com, laws-info.com, law.com
• Regex-based skip for US state government domains matching *.XX.us pattern (e.g. courts.state.ny.us)
• Skips file: protocol pages
• MutationObserver with 250ms debounce for dynamic content detection (SPAs, infinite scroll)
• Runs indefinitely — never disconnects the observer
• Safety gate using data-2257-checked attribute to prevent re-checking links
• Logging via console.debug() for verbose browser output
• Signal integration: New ICRA:2257 signal maps to the Sexual/Nudity category
• Block reason: "Sexual/Nudity (2257 Compliance)" displayed in block overlay
• Block overlay "Why:" section: When a block is triggered by the 2257 detector, the overlay now shows the matching link text and URL below the Reason line to aid in diagnosis

Users must now opt-in to early unlock option when activating self-lock

Major New Feature: Safe Request Mode We have introduced "Safe Request Mode," which automatically enforces server-side safety filters for popular search engines and content platforms. This works alongside our existing on-page filtering to provide multi-layered protection.

Search Engines: Enforces "SafeSearch" or "Strict Mode" for Google, Bing, Yahoo, and DuckDuckGo.
YouTube: Enforces Restricted Mode (Strict or Moderate).
Tumblr Integration (New in 1.2.0): Seamlessly filters mature content from Tumblr feeds without breaking the page layout. This includes:
Filtering initial page content.
Intercepting background API requests (infinite scroll).
Removing posts labeled as "Potentially mature," "Adult," or "Explicit."
New Feature: Self-Lock Mode Designed for voluntary content restriction, Self-Lock allows users to lock the filter settings for a set duration.

Flexible Timers: Lock for 1 hour, 24 hours, 1 week, or a custom duration.
Anti-Tamper: Prevents disabling the extension or changing settings while active.
Strict Enforcement: Optionally enforces Safe Request Mode and overrides allow-lists during the lock period.
Recovery: Includes a secure recovery code system and a "cool-down" unlock mechanism to prevent impulsive decisions.
Enhancements:

Settings UI: Completely redesigned settings page with tabbed navigation (General, Self-Lock, Security, Safe Request Mode).
Security: Separate passphrases for Admin settings and Self-Lock mode.
Performance: Optimized content detection scripts for faster page loads.