feat: add declarative site parser engine

STATE.md

@@ -1,6 +1,6 @@
 # N8 Parser: current state
 
-Last updated: 2026-08-10
+Last updated: 2026-08-11
 
 This file is the short handoff state for future Codex threads. Read this first before touching the project, so the whole chat history does not need to be carried forward.
 
@@ -107,20 +107,18 @@ and returns normalized JSON.
   `/raw/media/{media_id}` URLs backed by Telegram storage, not by the original
   Cloudflare-protected URL.
 
-Current RAA test source:
+Site configs now use the versioned declarative v1 contract. The worker supports
+RSS or HTML discovery, exact per-field selectors with ordered fallbacks,
+detail-page extraction, media normalization, bounded retries, and per-item
+errors. A failed detail does not discard successful siblings; the source is
+marked with a visible partial diagnostic.
 
-```json
-{"format":"rss","access":"cloudflare","max_items":10,"min_text_length":50}
-```
-
-- Source: `Popularairsoft`, `https://popularairsoft.com/rss.xml`, id `72`.
-- Do not add `follow_links` for the current feed: one of ten entries has a real
-  body and four photos; the other nine are title-plus-video entries and are
-  intentionally skipped by the body-length threshold.
-- Verified on 2026-08-10: source status `ok`, raw post `102` reached
-  `storage_ready`, and all four photos were uploaded. Its later rejection is
-  an AI non-target decision, not a parser/media failure.
-- Verified RAA deploy: commit `f866a13a8a83598e26a7a6a3406b8bc1c01c9a80`.
+PopularAirsoft source id `72` should use `https://popularairsoft.com/`, not its
+mixed-content RSS feed. The complete config is documented in
+`site_parser_worker/README.md`. It discovers the `The Latest News` cards and
+then extracts each `article.news.full` page. Verified live on 2026-08-11: nine
+items, nine full details, zero errors. Twitch embeds return as
+`external_video`; YouTube embeds are normalized for the existing downloader.
 
 LXC 108 operations: