ingest-tweet

name	ingest-tweet
description	Extract full text from X/Twitter posts for wiki ingestion. Uses FXTwitter API, no auth required.
user-invocable	false
allowed-tools	Bash(curl *)

Ingest Tweet

Extract the full text of an X/Twitter post.

Extraction Method

1. Parse the URL to extract username and tweet ID:

   • Pattern: https://x.com/{user}/status/{id} or https://twitter.com/{user}/status/{id}
   • Strip query parameters (?s=46, ?t=...)

2. Fetch via FXTwitter API (free, no auth, returns full text including long "note tweets"):

curl -s "https://api.fxtwitter.com/{user}/status/{id}"

3. Parse JSON response:
   • tweet.text — full tweet text (with expanded URLs, not t.co links)
   • tweet.author.name / tweet.author.screen_name — author
   • tweet.created_at — date
   • tweet.likes, tweet.retweets, tweet.replies — engagement
   • tweet.quote — quoted tweet (if present)
   • tweet.media — images/videos (if present)

Post-Extraction

• Save to vault's raw/<author>-tweet-<id>.md with YAML header:

---
source-url: <tweet-url>
title: "<author> tweet on <topic>"
author: "<name> (@<screen_name>)"
date-fetched: <today>
source-type: tweet
---

• Include quoted tweet text if present
• Note engagement stats in the source-note

Fallback

If FXTwitter fails, use oEmbed (truncates long tweets):

curl -s "https://publish.x.com/oembed?url=<original-url>"

Dependencies

None — uses only curl.

See also

• ronan-skills/x-scan — generic sibling for browsing tweets (URL, user timeline, topic search) without vault persistence. Use this skill when you want to keep the content.