Add HTTP-first web fetch tool

[mikehostetler]

Summary

• add Jido.Browser.web_fetch/2 and a new Jido.Browser.Actions.WebFetch action for stateless HTTP-first retrieval
• wire web fetch into the browser plugin with seen_urls tracking and web_fetch_uses guardrails
• add README/changelog coverage and focused tests for fetch behavior, plugin state, and composite action flow

Notes

• HTML/text fetches support domain allow/block rules, approximate token caps, focused filtering, ETS caching, and citation-ready passages
• PDF fetches require pdftotext to be available or configured via config :jido_browser, :web_fetch, pdftotext_path: ...

Testing

• mix test