Write LinkedIn Posts That Sound Human

Turns Claude (or any LLM) into a LinkedIn ghostwriter that sounds like a real person instead of corporate AI. The system eliminates 40+ AI vocabulary tells, breaks AI's default structure, and includes a pre-publish checklist that catches the subtle patterns readers identify as bot-written content. Users can apply this directly to Claude Projects or adapt to any LLM, with voice customization so posts sound like you, not the author.

1. Build your voice profile by analyzing 5-10 of your own best LinkedIn posts, then paste the analysis into your skill file. 2. Upload the 4 skill files to your Claude Project and add custom instructions. 3. Write posts using the system: open with a specific hook, structure for saves, run the pre-publish anti-AI checklist. 4. Refine your voice profile every 5 posts based on what Claude got right or wrong. Output: LinkedIn posts ready to copy-paste that sound like you wrote them from experience, with no AI smell.

Every post passes the AI detection checklist before publishing, because the system eliminates the vocabulary, structure, and tone patterns readers use to spot AI-written content.

Requires Claude Pro and Claude Projects feature to upload knowledge files and use custom instructions. The system is plain markdown, so files can be pasted into any LLM's system prompt or context window. Works with ChatGPT, other LLMs, and local AI tools. The voice-building methodology is LLM-agnostic. The weekly system integrates with Notion but works with any content backlog tool (spreadsheet, plain text file, or even handwritten notes).

These files are calibrated to the author's voice (direct, confrontational, coaching-oriented). If used as-is without customization, posts will sound like him, not you. The anti-AI guide is universal, but all voice elements (banned words, phrases, tone) must become yours. Requires Claude Pro ($20/month) and 3 hours initial setup time to adapt the system to your voice.