firefox stealth engine for StealthyFetcher (the #291 idea, firefox side)

[feder-cr]

saw issue #291 (CloakBrowser stealth backend for cloudflare) got a "will take a look later" rather than a no, so opening this as the firefox-side version of the same idea, with an open-source MPL-2 binary.

three things in the repo make this a small lift:

1. executable_path already exists in the fetcher options (custom browser executable instead of bundled chromium, for custom builds), so the wiring to point at an alternate binary is already there.
2. camoufox was the StealthyFetcher engine until 0.3.13, then dropped for patchright over memory/stability per docs/fetching/stealthy.md. so a firefox stealth engine isn't a foreign idea here, it was the original one.
3. the recurring pain (Feature: CloakBrowser stealth backend for Cloudflare bypass #291, Unable to bypass Cloudflare challenge for Quora URL #191 cloudflare on quora, Ebay Bot Detection #146 ebay bot detection) is exactly what an engine swap targets.

the detection angle: patchright patches chromium at the js layer, still readable via native .toString() enumeration, the CDP attach signature, and chromium-shape risk weighting. a firefox build patched at the c++ source level (canvas readback, webgl getParameter, font metrics, audio, navigator, system colors) has no js shim and no CDP signature, and is a non-chrome engine some anti-bot stacks score more leniently. it's also a different build from camoufox: patches live in the compiled binary and it's fully pref-driven, which sidesteps the camoufox memory/stability issues that made you drop it.

proposed shape: a firefox branch in the StealthyFetcher engine that launches playwright.firefox.launch(executable_path=, firefox_user_prefs=) reusing the existing executable_path plumbing. binary is feder-cr/invisible_firefox (MPL-2), auto-downloads on first run. opened a draft PR with a proposal doc.

honest caveats: helps the fingerprint/engine layer only, not IP reputation or a Press & Hold once it fires; firefox via playwright has no CDP so the cdp_url path stays chromium-only. if in scope i'll wire the firefox engine branch + a docs example. if not, happy to close it.

[feder-cr]

opened a draft PR with the proposal doc: #342

[D4Vinci]

Thanks. I might look into changing the core a bit in the next versions, but adding multiple browsers by default is a bad idea. Anyway, users always have the freedom to create their own fetchers, but they don't have to be pushed to Scrapling.