v1.1.0 — Security Hardening

Security Hardening

This release closes 22 vulnerabilities (4 CRITICAL, 7 HIGH, 8 MEDIUM, 4 LOW).

What's new

• SSRF Prevention — URL validation with DNS rebinding check
• Playwright Sandbox — restricted API surface for user scripts
• Rate Limiting — 30 req/min default, file-based cross-process state
• Path Traversal Protection — scripts must be within skill directory
• Supply Chain — exact version pinning, lockfile, .npmrc
• Docker Hardening — localhost binding, env-file, resource limits
• New flags — --unsafe, --verbose, --no-rate-limit
• 40 new tests — URL validation (28) + sandbox (12)
• Security audit script — bash scripts/audit.sh

Special thanks

Huge thanks to @HeDo88TH (Luca Di Leo) for the comprehensive security audit, implementation, and code review.

v1.0.0 — Initial release

browser-search v1.0.0

First public release of browser-search — a skill for AI agents that orchestrates SearXNG, Camofox, and CloakBrowser for web search and automated browsing.

What's included

• SKILL.md — instruction set for AI agents (OpenCode, Claude Code, Cursor, etc.)
• SearXNG integration — multi-engine metasearch
• Camofox integration — full browser via REST API
• CloakBrowser integration — stealth browser for protected sites (Cloudflare, Akamai, etc.)
• Challenge detection for 6 anti-bot systems
• Readability.js integration for clean article extraction
• Automated setup and health check scripts
• MIT License

Quick start

git clone https://github.com/Johell1NS/browser-search
cd browser-search
npm install

Ask your AI agent to configure the rest.