📦 NEW: TweetClaw Social Automation - Add a new skill

README.md

@@ -101,6 +101,7 @@ Skills work seamlessly with **Command Code**  providing consistent, high-quality
 - **[File Organizer](skills/file-organizer/)** - Intelligently organizes files and folders by understanding context, finding duplicates, and suggesting better organizational structures.
 - **[Invoice Organizer](skills/invoice-organizer/)** - Automatically organizes invoices and receipts for tax preparation by reading files, extracting information, and renaming consistently.
 - **[Raffle Winner Picker](skills/raffle-winner-picker/)** - Randomly selects winners from lists, spreadsheets, or Google Sheets for giveaways and contests with cryptographically secure randomness.
+- **[TweetClaw Social Automation](skills/tweetclaw-social-automation/)** - Guides safe X/Twitter workflows with TweetClaw, including search tweets, search tweet replies, posting, follower export, media workflows, monitors, webhooks, and giveaway draws.
 
 
 ## Getting Started

skills/tweetclaw-social-automation/SKILL.md

@@ -0,0 +1,201 @@
+---
+name: tweetclaw-social-automation
+description: Guides safe X/Twitter workflows through TweetClaw. Use when you need to search tweets, search tweet replies, post tweets, post tweet replies, export followers, manage media, configure monitors or webhooks, or run giveaway draws with explicit approval boundaries.
+---
+
+# TweetClaw Social Automation
+
+This skill helps an agent plan, verify, and operate TweetClaw workflows for X/Twitter jobs. It is designed for cases where the operator wants concrete social automation outcomes, but still needs clear setup checks, narrow scope control, and explicit approval before any visible action.
+
+## When to Use This Skill
+
+- Search tweets for keywords, accounts, hashtags, or topics
+- Search tweet replies for support, sentiment, or community research
+- Post tweets or post tweet replies with preview-before-send review
+- Export followers, following, or other user lists for analysis
+- Upload or download tweet media during content workflows
+- Configure monitors or webhooks for ongoing tweet tracking
+- Run giveaway draws from replies or engagement lists
+- Verify a TweetClaw install path from GitHub, npm, ClawHub, or Context7 docs
+
+## What This Skill Does
+
+1. **Checks Setup Path**: Confirms whether TweetClaw is installed and which documentation surface should be used
+2. **Clarifies Intent**: Turns vague social requests into one concrete TweetClaw job
+3. **Narrows Scope**: Sets account, query, limit, date range, or target tweet before execution
+4. **Protects Writes**: Requires explicit confirmation before posting, replying, messaging, or editing anything public
+5. **Guards Bulk Actions**: Adds a second confirmation for exports, monitors, webhooks, and giveaway draws
+6. **Keeps Output Useful**: Uses concrete X/Twitter wording such as search tweets, tweet replies, post tweets, follower export, media upload, and media download
+
+## Prerequisites
+
+This skill assumes TweetClaw is already available to the operator through one of these surfaces:
+
+- GitHub: `https://github.com/Xquik-dev/tweetclaw`
+- npm: `@xquik/tweetclaw`
+- ClawHub package listing
+- Context7 library docs
+
+If TweetClaw is not installed yet, guide the user to the documented setup path before planning live actions.
+
+## How to Use
+
+### Setup Verification
+
+Ask:
+
+```
+Is TweetClaw already installed and configured, or should I walk you through the setup path first?
+```
+
+Then verify which surface the user wants:
+
+- GitHub README and docs
+- npm package install
+- ClawHub install path
+- Context7 documentation
+
+### Read Workflows
+
+Use prompts such as:
+
+```
+Search tweets for mentions of [topic] from the last 7 days, limit 25.
+```
+
+```
+Search tweet replies on this post and summarize the top complaint themes.
+```
+
+```
+Export followers for this account, limit 200, and group them by bio keywords.
+```
+
+### Write Workflows
+
+Use prompts such as:
+
+```
+Draft a reply to this tweet, then show me the final text before posting.
+```
+
+```
+Prepare a tweet with this image and CTA, but do not send until I approve.
+```
+
+### Monitor And Webhook Workflows
+
+Use prompts such as:
+
+```
+Set up a monitor for tweets mentioning [brand], but first show me the scope, filters, and notification behavior.
+```
+
+```
+Create a webhook for new tweets from this account, but confirm the destination and event rules first.
+```
+
+### Giveaway Workflows
+
+Use prompts such as:
+
+```
+Run a giveaway draw from replies on this tweet, exclude duplicates, and show me the entry count before picking winners.
+```
+
+## Instructions
+
+When a user requests help with TweetClaw:
+
+1. **Identify the job**
+
+   Ask for the exact task:
+   - Search tweets?
+   - Search tweet replies?
+   - Post tweets?
+   - Post tweet replies?
+   - Export followers?
+   - Media upload or media download?
+   - Direct messages?
+   - Monitor tweets?
+   - Webhooks?
+   - Giveaway draws?
+
+2. **Confirm setup state**
+
+   Determine whether the user:
+   - already has TweetClaw installed
+   - needs install guidance
+   - needs config verification
+   - only wants documentation or workflow planning
+
+   If setup is missing, stop execution planning and guide the user through the documented install path first.
+
+3. **Narrow the scope**
+
+   Before any live action, gather:
+   - target account
+   - tweet URL or search query
+   - result limit
+   - date range
+   - output format
+   - whether this is one-time or recurring
+
+   Do not proceed on vague requests such as "scrape Twitter" or "post something about this."
+
+4. **Protect public and paid actions**
+
+   For posting, replying, deleting, following, liking, messaging, monitors, webhooks, and giveaway draws:
+   - summarize the exact action
+   - show the final text or configuration
+   - mention the target account or tweet
+   - wait for explicit approval
+
+5. **Handle read workflows efficiently**
+
+   For search, reply analysis, user lookup, and follower export:
+   - keep default limits narrow
+   - prefer the smallest result set that answers the question
+   - offer expansion only after the first result is reviewed
+
+6. **Summarize results in user language**
+
+   Prefer concrete phrasing:
+   - "searched tweets"
+   - "searched tweet replies"
+   - "prepared a tweet draft"
+   - "exported followers"
+   - "prepared a media upload"
+   - "configured a monitor"
+
+   Avoid vague wording such as "handled social workflow" or "processed automation request."
+
+## Example Workflow
+
+**User**: "I need to find people complaining about our checkout flow on X and reply from our brand account."
+
+**Agent using this skill**:
+
+1. Confirms TweetClaw is installed and which account will be used
+2. Searches tweets for relevant complaint terms with a narrow limit
+3. Summarizes the complaint clusters
+4. Drafts candidate replies
+5. Shows the final reply text and target tweet URLs
+6. Waits for explicit approval before posting replies
+
+## Safety Rules
+
+- Never post, reply, message, follow, like, or delete without explicit approval
+- Never widen a monitor, webhook, export, or giveaway scope without telling the user exactly what changed
+- Never assume which X account should be used
+- Never expose private account data unless the user clearly asks for it
+- Never present a draft as already posted
+
+## Recommended References
+
+Prefer current public references over memory:
+
+- GitHub repository and README
+- npm package metadata
+- ClawHub listing
+- Context7 library docs and benchmark page