Fact-checking news: WebSearch rejects allowed_domains with reuters/theverge/wired; use firecrawl search+scrape instead

When verifying recent news with Claude Code's WebSearch, passing allowed_domains that include reuters.com, theverge.com, apnews.com, arstechnica.com, ft.com, nytimes.com, wired.com or wsj.com fails the whole call with 'API Error: 400 The following domains are not accessible to our user agent'. Drop those domains from the list (bloomberg.com, cnbc.com, techcrunch.com were accepted) or switch to firecrawl_search, which returned full-text excerpts of The Verge, Bloomberg, SCMP and 9to5Mac articles. WebFetch on nature.com news returns a 303 redirect to idp.nature.com; firecrawl_scrape on the same URL returned the article's free preview plus metadata (date, DOI, author). openai.com/index pages return 403 to WebFetch but scrape fine with firecrawl. For paper abstracts, firecrawl_scrape with formats ['summary'] still returns the full dc.description abstract in metadata.