Crimson Hunter
Overview
Recursive web crawler that discovers AI-facing attack surfaces by exploring web applications. Crawls same-origin links, parses sitemap.xml, respects robots.txt, and identifies AI endpoints through pattern matching and structural analysis.
Capabilities
- Recursive crawling — BFS traversal up to configurable depth (default: 3 levels)
- Sitemap.xml ingestion — discovers URLs from sitemap for comprehensive coverage
- robots.txt compliance — respects disallow rules with wildcard support
- AI endpoint detection — matches patterns in page content, script tags, WebSocket URLs
- Rate limiting — configurable delay between requests (default: 300ms)
- Page limits — prevents runaway crawls (default: 50 pages)
- Same-origin enforcement — only crawls within target domain
What it finds
- API endpoints matching AI patterns (
/chat,/completions,/generate, etc.) - WebSocket URLs for streaming AI responses
- Chat input fields with prompt indicators
- JavaScript references to AI service endpoints
- Server-sent events (SSE) for streaming responses
Permissions required
proxy.read— reads captured traffic for link extractionscanner.write— writes discovered surfaces as findingsui.tab— registers the Crimson Hunter tabhttp.fetch— makes outbound requests during crawling
UI type
Tab — provides a full-width discovery dashboard showing crawled surfaces, entry points, and coverage.
How to install
Available in the Xtension Store. Requires Pro or Enterprise.