Browser Extract

Extracts structured data from web pages with built-in security gates that scan for PII and prompt injection before anything reaches the model or gets stored. The template system is clever: successful extractions auto-save as reusable recipes you can run against similar pages on the same domain. It wraps the older browser-scrape skill but adds AIDefence checks at every handoff point and records sessions as RVF containers for replay. The host-scoped templates are practical for recurring scrapes but won't transfer between sites. If you need to pull data from rendered pages without leaking sensitive strings into your workflow, this handles the paranoid plumbing so you don't have to wire aidefence_has_pii calls yourself.

Install to Claude Code

npx -y skills add ruvnet/ruflo --skill browser-extract --agent claude-code