AgriciDaniel/claude-repurpose / skills/repurpose-twitter
repurpose-twitter
VerifiedGenerates Twitter/X content from content atoms: threaded breakdowns (8-12 tweets), standalone tweets (3-5 variations), and engagement polls. Optimizes for reply-driven reach, curiosity-gap hooks, and algorithm-friendly formatting. Sub-skill of the Content Repurposing Engine. Use when user says "twitter", "tweet", "X post", "thread", "tweet thread", or "repurpose for twitter".
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 5.4 KB |
SKILL.md
View on GitHub---
name: repurpose-twitter
description: >
Generates Twitter/X content from content atoms: threaded breakdowns (8-12 tweets),
standalone tweets (3-5 variations), and engagement polls. Optimizes for reply-driven
reach, curiosity-gap hooks, and algorithm-friendly formatting. Sub-skill of the
Content Repurposing Engine. Use when user says "twitter", "tweet", "X post",
"thread", "tweet thread", or "repurpose for twitter".
user-invokable: true
argument-hint: "[url-or-atoms]"
license: MIT
metadata:
author: AgriciDaniel
version: "1.0.0"
category: content
---
# Twitter/X Content Generator
Generate thread, standalone tweets, and poll from content atoms.
## Inputs
You receive from the orchestrator:
- **Content atoms** (typed and impact-rated)
- **Main argument**, target audience, primary topic
- **Brand voice profile** (5-dimension scores or override flag)
- **Flags**: `--brief` (1 tweet only), `--images` (generate image prompts)
## References
Load before generating:
- `references/platform-specs.md` -- character limits, algorithm weights, posting times
- `references/hook-formulas.md` -- curiosity gap, contrarian, data-driven, question hooks
- `references/voice-adaptation.md` -- Twitter tone rules (concise, punchy, no hedging)
## Platform Rules
| Spec | Value |
|------|-------|
| Character limit | 280 (standard), 25,000 (Premium) |
| Sweet spot | 71-100 characters for highest engagement |
| Thread length | 8-12 tweets (+63% impressions vs single) |
| Hashtags | 0-2 max per tweet (more = lower reach) |
| Algorithm | Reply = 27x a like; Conversation = 150x a like |
| External links | Near-zero reach for non-Premium accounts |
| Images | 1600x900 (16:9), up to 4 per tweet |
| Best time | Tuesday 9AM, generally 8-10AM weekdays |
## Output 1: Thread (8-12 Tweets)
Write to `twitter/thread.md`.
### Thread Structure
1. **Hook tweet** (tweet 1)
- Must work as a standalone tweet even if nobody expands the thread
- Use a curiosity gap OR bold stat from the highest-impact atom
- End with `:` or `A thread:` to signal continuation
- Never open with "Thread:" or "1/"
- 71-100 characters is the sweet spot
2. **Insight tweets** (tweets 2-9)
- One key point per tweet, drawn from individual atoms
- Number each tweet for scannability (`2/` prefix)
- Use line breaks between ideas within a single tweet
- Alternate between stat, insight, contrarian, and howto atoms
- Insert a **visual break** every 3-4 tweets: a one-liner quote, a question, or an image suggestion
3. **Closing tweet** (final tweet)
- Summarize the main argument in one sentence
- Include a clear CTA: follow, repost, reply, or bookmark
- If a link is needed: suggest placing it in a **reply** to the final tweet (not in the tweet body) to avoid reach suppression
- For non-Premium accounts, alternatively suggest "link in bio"
### Thread Rules
- Each tweet must be under 280 characters
- Thread should flow logically even if individual tweets are read in isolation
- Frame at least 2 tweets as questions or contrarian takes to invite replies (Reply = 27x like)
- Do not repeat the same hook formula twice in one thread
## Output 2: Standalone Tweets (3-5)
Write to `twitter/standalone-tweet.md`.
Generate 3-5 standalone tweets. Each one:
- Captures a single atom and works completely independently
- Uses a different format from the others:
- **Stat tweet**: leads with a number or data point
- **Question tweet**: asks a thought-provoking question
- **Hot take tweet**: bold/contrarian claim that invites debate
- **Tip tweet**: one actionable piece of advice
- **Analogy tweet**: comparison that makes the concept click
- Stays within the 71-100 character sweet spot when possible (never exceed 280)
- Includes 0-1 hashtags maximum
### Engagement Optimization
- Frame for replies, not likes. Questions and contrarian takes generate 27x more algorithmic weight than passive likes.
- Tweets that invite disagreement or personal experience outperform agreement-bait.
- Avoid "Agree?" or "Thoughts?" -- ask a specific question instead.
## Output 3: Poll
Write to `twitter/poll.md`.
Derive one poll from the content's key decision point or debatable topic.
| Spec | Constraint |
|------|-----------|
| Options | 2-4 |
| Chars per option | 25 maximum |
| Duration | Suggest 24h or 7 days |
**Poll structure:**
- Question: specific, not generic (derived from the content's central tension or choice)
- Options: represent real positions; include one surprising/contrarian option
- Context tweet: 1-2 sentences above the poll that frame why this question matters
## External Link Warning
Links in tweet bodies receive near-zero reach for non-Premium accounts. Always:
1. Place links in a **reply** to the main tweet (not the tweet body)
2. Or suggest "Link in bio" with a directional cue
3. Note this in the output so the user understands the trade-off
## Image Prompts
If `--images` flag is set, suggest one hero image for the thread opener:
- Dimensions: 1600x900 (16:9)
- Purpose: visual hook for tweet 1 of the thread
- Include a prompt description for /banana (subject, context, style, composition)
## Output Format
Each output file should use this structure:
```markdown
# Twitter Thread: [Topic]
## Hook Tweet
[tweet text]
## Tweet 2/N
[tweet text]
...
## Closing Tweet
[tweet text]
---
**Posting notes:** [best time, link placement advice, image suggestions]
```
## Brief Mode
If `--brief` is active, produce only:
- 1 standalone tweet (highest-impact atom)
- Skip thread and poll
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |