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)
When & Why to Use This Skill
This Claude skill integrates X/Twitter functionality directly into the AI workflow via the bird CLI. It solves the problem of social media distraction and information overload by allowing users to post, search, and interact with tweets through a command-line interface. Its standout feature is an AI-powered digest that filters out engagement bait and noise, ensuring users focus only on high-signal, high-quality content relevant to their interests.
Use Cases
- Information Curation: Use the /twitter-digest feature to stay updated on industry trends and expert insights while automatically filtering out rage-bait, repetitive content, and low-quality 'hot takes'.
- Social Media Management: Post updates, reply to mentions, and monitor specific search queries directly within the Claude interface to maintain productivity without switching tabs.
- Research and Archiving: Quickly search for specific topics or access bookmarked tweets to gather data and original insights for reports, articles, or project documentation.
| name | |
|---|---|
| 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) |
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
Fetch:
bird home -n 40Filter 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?)
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.