feed-wsj (WSJ Markets source adapter — headlines only)

Pure fetch + normalize adapter for a paywalled outlet. WSJ bodies are behind a paywall, so this adapter emits headline + url + published_at only and marks the body [UNAVAILABLE - paywall]. Dedup/ store/judge live downstream in [[crypto-news-store]] + [[narrative-news]].

Hard rule (paywall)

NEVER fabricate body text. Headline-only is acceptable; invented prose is a defect (PRD AC6). On any failure → [UNAVAILABLE]. Return ≥1 headline record or a clean [UNAVAILABLE] (AC5).

Retrieval recipe

• Endpoint (verified resolving, RSS 2.0 — WSJ Markets Main): https://feeds.a.dj.com/rss/RSSMarketsMain.xml Other WSJ feeds (same publisher, same paywall rule): https://feeds.a.dj.com/rss/RSSWorldNews.xml, https://feeds.a.dj.com/rss/WSJcomUSBusiness.xml.
• Parse channel > item: title→title, link→url (canonicalize, strip utm_*), pubDate(RFC-822)→published_at (ISO-8601 UTC). The RSS description is WSJ's own teaser — keep verbatim as summary if present (publisher-provided, not fabricated), else "[UNAVAILABLE - paywall]". Do NOT scrape the full body. category→tags, lang: en, source: wsj.

Reading the BODY (verified method, June 2026)