tweet-composer
Score and optimize tweets based on X's real open-source ranking algorithm. Analyzes draft tweets against the actual ranking code — not generic tips. Use when: composing tweets, optimizing drafts for reach, planning threads, analyzing why a tweet performed well/poorly, or asking for posting strategy advice.
- License 许可
- MIT
- 最近更新
- 2026-11-10
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- tweet-composer
- author
- ERIC
- author_blog
- https://topdigg.com
- description
- Score and optimize tweets based on X's real open-source ranking algorithm. Analyzes draft tweets against the actual ranking code — not generic tips. Use when: composing tweets, optimizing drafts for reach, planning threads, analyzing why a tweet performed well/poorly, or asking for posting strategy advice.
- author_email
- gyc567@gmail.com
- author_twitter
- https://twitter.com/EricBlock2100
02 资源文件 4
- 资产文件assets/banner.svg789 B
- 许可证LICENSE1.0 KB
- 参考文档README.md6.3 KB
- 参考文档references/algorithm-rules.md8.3 KB
03 SKILL.md 正文
Tweet Composer
Score and optimize tweets using rules derived from X's open-source ranking algorithm.
How It Works
X's "For You" feed is ranked by a Grok-based transformer (Phoenix) that predicts 19 engagement actions for every candidate tweet. The final score is a weighted sum of these predictions. This skill encodes the structural rules from that pipeline into a scoring system.
For the full algorithm breakdown, read references/algorithm-rules.md.
Scoring a Draft Tweet
When a user asks to score or optimize a tweet draft:
- Read
references/algorithm-rules.mdfor the complete rules engine - Analyze the draft against all rules
- Output the score card in this format:
🐦 Tweet Composer — Score: XX/100
[Category scores with ✅ ⚠️ ❌ indicators]
📊 Predicted Action Boost:
├─ P(reply): [assessment]
├─ P(favorite): [assessment]
├─ P(share): [assessment]
├─ P(dwell): [assessment]
└─ P(not_interested): [assessment]
💡 Suggestions:
→ [actionable improvements]
✏️ Optimized version:
"[rewritten tweet]"
Scoring Rubric (Quick Reference)
Score 0-100 based on weighted categories:
| Category | Weight | What to check |
|---|---|---|
| Reply potential | 25 | Questions, opinions, CTAs that drive replies |
| Media | 20 | Native image/video attached (not link previews) |
| Shareability | 15 | Would someone DM this or copy the link? |
| Dwell time | 15 | Length that makes people stop scrolling |
| Content quality | 10 | Clear, original, not generic |
| Format | 10 | No links in body, no hashtags, good length |
| Negative signals | 5 | Risk of not_interested/mute/block |
Thread Optimization
When composing threads:
- First tweet = strongest hook (DedupConversationFilter keeps only the best per conversation)
- 3-6 tweets max (short threads > mega-threads)
- Each tweet self-contained (many see only the first)
- Media on tweet 1 or 2 for photo_expand boost
- CTA in last tweet
Quick Rules (No Reference File Needed)
- Links: Always in reply, never in body (learned penalty from lower engagement)
- Hashtags: Zero. The model learns they reduce engagement
- Length: 100-200 chars sweet spot for single tweets
- Media: Native image/video = separate P(photo_expand) and P(video_quality_view) predictions
- Video: Must exceed minimum duration threshold for VQV weight to apply
- Timing: Post when your audience is active — engagement velocity in first 30 min is critical
- Frequency: AuthorDiversityScorer penalizes exponentially: 2nd post ~55% score, 3rd ~33%. Max 3-4 strong tweets/day
- Quote tweets: P(quote) has dedicated weight — QTs with added value outperform plain retweets
- Engagement bait: Questions/polls drive P(reply). "What would you add?" > "Like if you agree"
<!-- ERIC-AUTHOR-INFO:START -->
📬 作者信息
ERIC — AI技术专家,专注于人工智能和自动化工具的研究与应用
| 渠道 | 链接 |
|---|---|
| 📧 邮箱 | gyc567@gmail.com |
| @EricBlock2100 | |
| 💬 微信 | 360369487 |
| ✈️ Telegram | @fatoshi_block |
| 📢 Telegram 频道 | @cryptochanneleric |
| 🔐 加密情报 TG 群 | @btcgogopen |
| ▶️ YouTube | @0XBitFinance |
| 🌐 个人技术博客 | topdigg.com |
本 Skill 由 ERIC 免费提供。想深入交流或获取更多同类 Skill,欢迎通过上表任一渠道联系。
<!-- ERIC-AUTHOR-INFO:END -->
04 同仓库更多 Skills
writing-dna-skill
从至少 20 篇完整文章中蒸馏可复用的写作 DNA,分析语言、文章结构、选题逻辑、素材策略、认知框架和视觉风格,并生成 Writing-DNA.md;按该 DNA 写作时会读取全部蒸馏产物和 5 篇相关原文。用于中英文作者、账号、品牌或出版物的风格分析与一致性写作。 Distill reusable Writing DNA from at least 20 complete articles for Chinese or English authors, publications, brands, and accounts; use for language, structure, topic logic, source strategy, cognitive-frame, and visual-style analysis.
ub-quality-gates
Uncle Bob 编程纪律与确定性质量门工具链。写代码或改代码时使用:先让它能跑,再让它正确,最后守纪律。 当用户要求 /ub-mode、需要代码质量门禁、CRAP 评分、重复代码检测、变异测试、 或抱怨代码难懂/难测/有重复时触发。覆盖 Clojure/Java/Go/TypeScript/Rust/Python。
short-video-scripter
Use when the user asks to "script this short video", "write a TikTok / Reels / Shorts script", "给这条抖音或视频号视频写脚本", or "fix the hook — viewers drop off in the first seconds"; produces timestamped beat-sheet scripts on the retention-gate model (0-2s hook / 2-5s confirmation / 5-15s payoff / loop-or-CTA) with per-scene script lines, on-screen text for muted viewing, asset keywords, and 9:16 format-param rows per platform norm card for TikTok, Reels, Shorts, 抖音, and 视频号 — plus 2-3 hook options from named hook families and an AI-content disclosure line by default on realistic synthetic media. Spec-only: rendering, TTS, and publishing stay with the user's own tools — no pipelines, no upload automation. Not for creator video briefs — use brief-generator; long-form video SEO belongs to content-writer. 短视频脚本/抖音分镜/开头钩子/竖屏9:16
remotion-video-master
Remotion 视频编程大师 — 用 React 组件以编程方式创建 MP4 视频。支持动画、字幕、音乐可视化、 AI 配音(TTS)与场景自动配时。触发词:用代码做视频、编程视频、React 视频、Remotion、 /remotion-video。适用场景:程序化视频(批量生成、数据驱动如年度总结、音乐可视化、自动字幕)、 教程讲解视频(技术概念可视化如 CNN 与算法、分层递进讲解、AI 配音教程)、 3D 视频(产品展示与模型动画、卡通角色讲解、3D 数据可视化、Logo 动画)。
short-videos
Produce 90-second universal short-video scripts (HeyGen / Instagram Reels / LinkedIn Video / YouTube Shorts) and standalone LinkedIn posts using native Hermes tools. No Claude Code dependency.
medium-story
Full Medium article pipeline using native Hermes tools. Research → write → 4 parallel output agents → HTML conversion → git commit/push. Includes pre-flight infrastructure verification step, revisor-methodology fact-checking, and an anti-AI-slop writing pass (banned vocabulary, structural variety, accuracy rules).