twitter

odysseus0 的头像作者 odysseus0

通过bird CLI使用X/Twitter。当用户提及Twitter、X、发推、发布、阅读推文或想要查看他们的动态时使用。包括用于精选动态过滤的/twitter-digest。(用户)

📖 中文导读:通过bird CLI使用X/Twitter。当用户提及Twitter、X、发推、发布、阅读推文或想要查看他们的动态时使用。包括用于精选动态过滤的/twitter-digest。(用户) 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。

name: twitter description: X/Twitter via bird CLI. Use when user mentions Twitter, X, tweeting, posting, reading tweets, or wants to check their feed. Includes /twitter-digest for curated feed filtering. (user)

Twitter

bird CLI for X/Twitter. Account: @odysseus0z

Runs from local ~/projects/bird (personal branch). Update: cd ~/projects/bird && git pull && npm run build

Quick Reference

bird tweet "text"              # post
bird reply <url> "text"        # reply
bird read <url>                # read tweet (full article body)
bird search "query" -n 10      # search
bird mentions -n 10            # mentions
bird bookmarks -n 10           # bookmarks
bird home -n 20                # home timeline
bird home --following -n 20    # following-only timeline

Run bird --help for full syntax.


/twitter-digest

Curated feed that filters the home timeline, protecting attention from algorithmic engagement bait.

Process

  1. Fetch: bird home -n 40

  2. Filter each tweet by:

    • Signal vs noise (insight vs engagement bait)
    • Relevance to user's interests
    • Quality of thought (original vs retweet farming)
    • Actionability (something to learn or do?)
  3. Present digest:

## Twitter Digest

### Worth Your Attention (N items)

**@username** — [why relevant]
> Tweet preview...
> [link]

---

### Skipped (N items)
<details>
<summary>Review what I filtered</summary>

- **@user**: "preview..." — *reason* [→ link]
- ...
</details>

---

### Feedback
Did I filter correctly?
- Anything I should have included?
- Anything I included that wasn't useful?

Include

  • Original insights
  • Threads with depth
  • Relevant news/updates
  • Interesting people, interesting thoughts

Skip

  • Engagement bait ("hot takes...")
  • Rage bait, dunking, drama
  • Repetitive content
  • Self-promotion without substance
  • Vague motivational fluff

Iteration

v1 uses AI judgment. As user provides feedback, patterns emerge → eventually capture in a personalization profile.