FxTwitter extractor: username regex rejects usernames starting with digits

[brettdavies]

Bug

The FxTwitter extractor in src/extractors/x-oembed.ts silently fails for X/Twitter usernames that start with a digit. The regex in tryExtractFxTwitter:

const match = this.url.match(/\/([a-zA-Z][a-zA-Z0-9_]{0,14})\/(status|article)\/(\d+)/);

requires the first character to be [a-zA-Z], but X allows usernames starting with digits (e.g., 0xRicker, 0xMovez, 0xwhrrari). Since match returns null, tryExtractFxTwitter returns null, and the extractor silently falls back to oEmbed, which only returns the tweet text, not the article content.

The same pattern exists in src/extractors/x-article.ts in getAuthorFromUrl.

Reproduction

# Works: username starts with a letter (1627 words = full article)
bunx defuddle parse "https://x.com/coreyganim/status/2036070952987988290" --json | jq .wordCount
# 1627

# Fails: username starts with digit (4 words = just the tweet text, not the article)
bunx defuddle parse "https://x.com/0xRicker/status/2035334040216113631" --json | jq .wordCount
# 4

Both tweets are X Articles. FxTwitter API returns full content for both:

curl -s "https://api.fxtwitter.com/0xRicker/status/2035334040216113631" | jq '.tweet.article.title'
# "I Simulated 10,000 Polymarket Seasons. Only 3 Strategies Survived"

curl -s "https://api.fxtwitter.com/0xRicker/status/2035334040216113631" | jq '.tweet.article.content.blocks | length'
# 59

Suggested fix

Change [a-zA-Z][a-zA-Z0-9_]{0,14} to [a-zA-Z0-9_]{1,15} in the username capture group. Same pattern appears in two files:

• src/extractors/x-oembed.ts#L183 (tryExtractFxTwitter)
• src/extractors/x-article.ts#L73 (getAuthorFromUrl)

- /\/([a-zA-Z][a-zA-Z0-9_]{0,14})\/(status|article)\/(\d+)/
+ /\/([a-zA-Z0-9_]{1,15})\/(status|article)\/(\d+)/

Also affects usernames starting with underscore (e.g. _akhaliq).

[brettdavies]

Thank you!