# Threadwriter > Topic in, optimized X thread out. Takes your rough idea and turns it into a high-engagement Twitter thread with hooks, pacing, and CTAs. Use when: sharing insights on X/Twitter, building thought leadership, or creating viral-ready content. NOT for: automated posting bots, spam generation, or mass-scheduled content campaigns. - Skill: `majiayu000/threadwriter` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add majiayu000/threadwriter` - Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/threadwriter/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: majiayu000 (https://skillmd.com/u/majiayu000) - Updated: 2026-09-09 - Page: https://skillmd.com/skills/majiayu000/threadwriter --- # ThreadWriter ✍️ Turn rough ideas into polished X threads. Paw writes the hooks, nails the pacing, and adds CTAs. You just approve and post. ## When to Use ✅ **USE this skill when:** - You have insights but hate formatting threads - Building thought leadership on X/Twitter - Sharing technical content in digestible chunks - Announcing product launches - Breaking down complex topics - Creating viral-ready content - Repurposing blog posts into threads - Maximizing engagement on X ## When NOT to Use ❌ **DON'T use this skill when:** - Building automated posting bots → violates platform TOS - Generating spam or low-quality content - Mass-scheduling content campaigns → use Buffer, Hootsuite - Creating threads without actual insights (garbage in = garbage out) - Impersonating others - Generating misleading content ## Setup ```bash npm install -g @pawhub/threadwriter threadwriter setup ``` ## Example Interactions ### Quick Thread **You:** Write a thread about why AI agents are the next big thing **Paw:** *starts drafting* **Working on:** "Why AI agents are the next big thing" **Style:** Thought leadership **Target length:** 7-10 tweets *15 seconds later...* ✅ **Thread ready** (9 tweets) **Preview:** 1/ AI agents aren't just better chatbots. They're the first software that gets things done *for* you, not *with* you. And most people still don't see what's coming. --- 2/ Here's the thing: every previous AI breakthrough made you faster at doing tasks. GPT-4 made you write faster. DALL-E made you design faster. But you were still doing the work. --- 3/ Agents are different. You say "book me a flight to NYC, cheapest option under $400, window seat." And it's done. No tabs. No comparison shopping. No booking forms. The computer did the task. --- 4/ This is the shift from *tools* to *delegation*. Tools make you more productive. Agents make you more powerful. Different game entirely. (continues...) **Engagement optimizations:** - ✅ Hook in tweet 1 (grabs attention) - ✅ Pattern interrupt (tweet 2) - ✅ Concrete example (tweet 3) - ✅ Clear thesis (tweet 4) - ✅ CTA at end (tweet 9) **Want to:** - Review full thread? (yes/edit/rewrite) - Post now? (requires X auth) - Save draft? (for later posting) --- **You:** Review full thread **Paw:** **Full thread:** 1/ AI agents aren't just better chatbots. They're the first software that gets things done *for* you, not *with* you. And most people still don't see what's coming. --- 2/ Here's the thing: every previous AI breakthrough made you faster at doing tasks. GPT-4 made you write faster. DALL-E made you design faster. But you were still doing the work. --- 3/ Agents are different. You say "book me a flight to NYC, cheapest option under $400, window seat." And it's done. No tabs. No comparison shopping. No booking forms. The computer did the task. --- 4/ This is the shift from *tools* to *delegation*. Tools make you more productive. Agents make you more powerful. Different game entirely. --- 5/ Why now? Three things converged: • LLMs got good enough to understand complex instructions • APIs became ubiquitous (every service has one now) • Context windows got huge (remember entire workflows) The pieces were always there. Now they fit. --- 6/ What this means practically: In 2024, you had an assistant that answered questions. In 2026, you have an employee that executes tasks. By 2028? You'll have a team. --- 7/ The companies winning here aren't building better chatbots. They're building: • Reliable execution (it does what you ask) • Tool ecosystems (integrates with everything) • Memory systems (learns your preferences) Product =! Interface anymore. --- 8/ And this is where it gets wild: Once agents can reliably execute tasks, they can chain tasks together. Book flight → reserve hotel → add to calendar → notify team. Zero clicks. Just intent. --- 9/ We're at the "1995 of AI agents." Most people see chatbots. A few see what's really happening. What are you building with them? Drop a comment. I'm curious what use cases people are exploring. 👇 --- **Analysis:** - 📊 Thread length: 9 tweets (optimal for engagement) - 🎯 Hook strength: Strong (tweet 1) - 📈 Pacing: Good (builds logically) - 💡 Value: High (actionable insights) - 🔚 CTA: Clear (tweet 9 asks for engagement) **Estimated engagement:** Medium-High (if your audience is tech/startup-focused) **Edits needed?** Or ready to post? ✍️ --- **You:** Change tweet 7, remove the product ≠ interface line, too jargony **Paw:** *updates tweet 7* ✅ **Updated tweet 7:** 7/ The companies winning here aren't building better chatbots. They're building: • Reliable execution (it does what you ask) • Tool ecosystems (integrates with everything) • Memory systems (learns your preferences) The interface is just the starting point. --- Better? (yes/keep editing) --- **You:** Yes, post it **Paw:** *Connects to X API* 🚀 **Posted!** Thread is live: [X.com link] **What I'll track:** - Engagement (likes, retweets, replies) - Peak engagement time - Which tweet got the most interaction - Reply sentiment I'll send you a performance summary in 24 hours. Want real-time alerts for viral tweets? 📊 --- ### Repurposing Content **You:** Turn this blog post into a thread: [URL] **Paw:** *fetches and analyzes* **Blog post:** "How We Scaled to 1M Users" **Length:** 2,400 words **Key points:** 7 main sections *analyzing...* This is dense. I can turn it into: 1. **Long thread** (15-20 tweets, keeps all details) 2. **Medium thread** (10-12 tweets, highlights only) 3. **Teaser thread** (5-7 tweets, drives traffic to blog) Which approach? Most people do #2 (highlights thread). 🎯 --- **You:** Highlights thread **Paw:** *extracts key points* ✅ **Thread created** (11 tweets) **Hook (tweet 1):** "We scaled from 0 to 1M users in 18 months. Here's what worked (and what almost killed us) 🧵" **Structure:** - Tweet 1: Hook - Tweets 2-4: Early growth tactics - Tweets 5-7: Scaling challenges - Tweets 8-10: Key lessons - Tweet 11: CTA (link to full blog post) **Link placement:** Tweet 11 (best practice for engagement) Review? (yes/edit) --- ### Voice-to-Thread **You:** *[voice message]* "I just realized something about product-market fit. Everyone talks about finding it, but nobody talks about maintaining it. Your market shifts. Your product evolves. PMF isn't a destination, it's a moving target. And if you stop moving, you lose it. We learned this the hard way when our core user base changed but our product didn't adapt fast enough." **Paw:** *transcribes and analyzes* Great insight. Converting to thread format... ✅ **Thread ready** (6 tweets) **Preview:** 1/ Everyone talks about *finding* product-market fit. Nobody talks about *maintaining* it. And that's where most companies fail. --- 2/ PMF isn't a destination. It's a moving target. Your market shifts. Your product evolves. Competitors emerge. If you stop moving, you lose fit. --- 3/ We learned this the hard way. Our core user base changed, but our product didn't adapt fast enough. We had PMF in 2024. By 2025, we didn't. Same product. Different market. --- (continues...) **Style:** Personal/authentic (based on your voice) **Engagement potential:** High (relatable struggle) Want to post this? 🎯 --- ### Thread Analytics **Paw:** 📊 **Thread Performance (24h)** **Thread:** "Why AI agents are the next big thing" **Posted:** Yesterday, 9:14 AM **Engagement:** - 🔄 Retweets: 342 - ❤️ Likes: 1,847 - 💬 Replies: 94 - 👁️ Impressions: 87,230 **Top-performing tweet:** Tweet 3 (the flight booking example) — 2,103 likes **Engagement pattern:** - Peak: 2-4 hours after posting (typical) - Drop-off: tweet 7 (might be too technical) - Recovery: tweet 9 (CTA worked) **Reply sentiment:** - 87% positive - 8% questions (mostly "what tools?" ) - 5% critical (healthy skepticism) **Compared to your average thread:** - +127% retweets 🔥 - +89% likes - +65% replies This one hit. The flight booking example is resonating. **Action:** Want me to write a follow-up thread expanding on that example? 📈 ## Commands ```bash # Quick thread from prompt threadwriter create "topic here" # Thread from blog post URL threadwriter from-url https://blog.example.com/post # Thread from voice memo threadwriter from-voice audio.m4a # Thread from existing text file threadwriter from-file notes.txt # Post thread directly (requires X auth) threadwriter post thread-draft-123 # Save thread as draft threadwriter save thread-draft-123 # Analytics for posted thread threadwriter stats TWEET_ID # List all drafts threadwriter drafts list ``` ## Style Configuration ```bash # Set default tone threadwriter config set tone "authentic" # casual, professional, authentic, technical # Set default length threadwriter config set length "medium" # short (5-7), medium (8-12), long (13-20) # Enable/disable CTAs threadwriter config set auto-cta true # Set hook style threadwriter config set hook-style "pattern-interrupt" # direct, pattern-interrupt, question, story ``` ## Telegram Integration **Quick thread:** **You:** `/thread "AI agents topic"` **Paw:** *drafts thread* Thread ready (9 tweets). Preview first 2: 1/ AI agents aren't just better chatbots... 2/ Here's the thing: every previous AI... React with ✅ to post, ✏️ to edit, 💾 to save draft. ## Tips from Paw > "The hook is everything. If tweet 1 doesn't grab attention, the rest doesn't matter. I optimize for pattern interrupts." > "Threads under 10 tweets perform better. People's attention is finite. Say more with less." > "Always end with a CTA. Ask a question, request engagement, link to something. Don't just trail off." > "The flight booking example in that AI agents thread? That's specificity. Concrete examples > abstract concepts." > "Post between 9-11 AM EST or 2-4 PM EST. That's when tech Twitter is most active." ## Pricing - **Free tier:** 5 threads/month, basic optimization - **Pro:** $10/month — unlimited threads, voice-to-thread, analytics, priority generation - **Team:** $29/month — multiple accounts, brand voice training, A/B testing Install from PawHub: [pawhub.ai/threadwriter](https://pawhub.ai/threadwriter) ## Notes - Threads are optimized for X/Twitter (280 char limit) - Analytics require X API access (read-only) - Posting requires X OAuth (secure, revocable) - Voice transcription uses Whisper API - Thread drafts stored locally (encrypted) - All content remains yours (no rights transfer) --- Built for people with insights but no patience for thread formatting. ✍️🐾