Post To X

Post to X (Twitter)

Cross-post LinkedIn content to X, adapting text and attaching media.

Prerequisites

X API credentials must be set in ~/.claude/skills/post-to-x/.env:

X_API_KEY=your_api_key
X_API_SECRET=your_api_secret
X_ACCESS_TOKEN=your_access_token
X_ACCESS_TOKEN_SECRET=your_access_token_secret

These come from the X Developer Portal (developer.x.com). You need a project with "Read and Write" permissions. The access token must be generated with read+write scope.

Workflow

Step 1: Find the LinkedIn post content

Look for the post text in the current directory or the directory the user references:

• post-copy-final.txt or post-copy*.txt for text
• *.mp4, *.mov, *.webm for video
• *.jpg, *.jpeg, *.png for images
• thumbnail*.jpg, thumbnail*.png for thumbnail images

If the user provides text directly, use that instead.

Step 2: Adapt the text for X

LinkedIn posts are long-form. X has character limits:

• 280 characters for text-only posts
• 280 characters for posts with media (text portion)
• Threads: up to 25 tweets, each 280 chars

Strategy:

1. Extract the core hook/message from the LinkedIn post
2. Create a concise X version (aim for 200-250 chars to leave room for links)
3. Remove LinkedIn-specific CTAs ("DM me", "Comment below")
4. Replace with X-appropriate CTAs if needed ("Link in replies", thread format)
5. Keep hashtags minimal (0-2 max, only if truly relevant)
6. No emojis unless the user explicitly requests them

If the post is too rich for a single tweet, propose a thread (2-4 tweets max).

Always show the adapted text to the user for approval before posting.

Step 3: Upload media (if any)

python3 ~/.claude/skills/post-to-x/scripts/post_to_x.py upload-media 

This returns a media_id to use in Step 4.

For video: the script handles chunked upload (INIT, APPEND, FINALIZE, STATUS polling). For images: simple upload.

Size limits:

• Images: 5 MB (JPEG, PNG, GIF, WEBP)
• Video: 512 MB, up to 2:20 duration (MP4 with H.264 video, AAC audio)
• GIF: 15 MB

Step 4: Post the tweet

# Text only
python3 ~/.claude/skills/post-to-x/scripts/post_to_x.py post --text "Your tweet text"

# With media
python3 ~/.claude/skills/post-to-x/scripts/post_to_x.py post --text "Your tweet text" --media-ids "media_id_1,media_id_2"

# Thread
python3 ~/.claude/skills/post-to-x/scripts/post_to_x.py thread --texts "Tweet 1|||Tweet 2|||Tweet 3"

# Thread with media on first tweet
python3 ~/.claude/skills/post-to-x/scripts/post_to_x.py thread --texts "Tweet 1|||Tweet 2" --media-ids "media_id_1"

# Dry run (show what would be posted without actually posting)
python3 ~/.claude/skills/post-to-x/scripts/post_to_x.py post --text "Your tweet" --dry-run

Step 5: Verify

The script prints the tweet URL on success. Confirm the post looks correct.

Text Adaptation Examples

LinkedIn (long):

I genuinely stopped using PowerPoint.
Spending days aligning boxes feels insane when you can build a whole app in the same time.
So I made an app for building slides.
Three prompts. Full deck. 2 minutes.
No Figma. No templates. Just prompts.
Open source. Free forever.
DM "DECK" if you want early access.

X (short):

I stopped using PowerPoint and built an app that generates full slide decks from prompts.

3 prompts. Full deck. 2 minutes.

Open source, free forever.

github.com/yourorg/yourrepo

Error Handling

• 401 Unauthorized: Check API credentials in .env
• 403 Forbidden: App needs "Read and Write" permissions, not just "Read"
• 413 Payload Too Large: Video exceeds 512 MB or image exceeds 5 MB
• 429 Rate Limited: Wait and retry (script handles this automatically)
• Media processing failed: Video codec/format issue, try re-encoding with ffmpeg

Rules

1. Always show the adapted text to the user before posting. Never auto-post.
2. No emojis unless user requests them.
3. No em dashes. Use commas, semicolons, colons.
4. Keep it punchy. X rewards brevity. Cut ruthlessly.
5. Dry run first if unsure. Use --dry-run flag.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: floomhq
• Source: floomhq/moto
• License: MIT
• Homepage: https://github.com/floomhq/moto#readme

Install and usage instructions live in the source repository linked above.