🎯 What you get: the free, open-source skill that lets Claude run your entire LinkedIn — 11 commands, installed in about 2 minutes. This is the exact setup from the video.
The skill (free, open source): https://github.com/Jakeschincariol/linkedin-agent-skill
Install it in 2 minutes
Open Claude (Claude Code / the desktop app with skills enabled)
Paste this and hit enter:
npx skills add Jakeschincariol/linkedin-agent-skill (or clone the repo above and drop it in your skills folder)Say "install the skill" and let it finish
Type
/li-post to confirm it's liveThe 11 commands — what each one does
| Command | What it does | Reach for it when… |
|---|---|---|
/li-post | Writes a post: 3 hook options from 21 formulas • a full draft, humanized before you see it | You have an idea and want a post now |
/li-comment | Writes comments on other people's posts (9 contextual styles) | You're doing your daily engagement rounds |
/li-reply | Sorts your reply threads by priority and drafts responses | Your notifications are piling up |
/li-profile | Scores your profile against a 12-part rubric out of 100 and rewrites the weak parts | Your profile hasn't been touched in months |
/li-plan | Plans your week: what to post, when, and 10 people to engage | Sunday, setting up the week |
/li-human | The humanizer — strips AI tells and scores the draft (see below) | Before anything goes out |
/li-carousel | Builds a multi-slide document post with cover copy + PDF | You want a carousel, not a text post |
/li-repurpose | Turns one source (video, newsletter, transcript) into a week of posts | You just made a long piece of content |
/li-dm | Writes a 200-char invite note + 2 follow-ups | You're doing outreach |
/li-inbox | Triages incoming DMs into 5 categories | Your inbox is a mess |
/li-audit | Ranks your posts by engagement + reach, post-mortem style | End of month, figuring out what worked |
The humanizer (the part everyone asks about)
/li-human runs locally — two Python scripts, no dependencies, nothing leaves your machine. It automatically:
- Removes em dashes, curly quotes, and invisible Unicode characters (the hidden "watermark")
- Strips 113 stock AI-slop words (delve, leverage, seamless, robust…)
- Scores your draft on 5 checks — Burstiness, Specificity, Slop Density, Fingerprint, Voice — and flags rule-of-three patterns, rhetorical questions, and uniform sentence length
⚠️ Straight talk: these are local heuristics, not a call to ChatGPT-Zero or any detector API. It makes your writing sound like you and kills the obvious tells — it can't promise a specific detector's verdict. Anyone claiming a guarantee is selling you something.
Your 10-minute daily rhythm
/li-planonce a week → your content + engagement targets are decided/li-post(or/li-repurposeoff your latest video) → today's post, humanized/li-commenton your 10 targets → real engagement, not "great post!"/li-reply+/li-inbox→ clear the notifications/li-auditat month-end → double down on what worked
Nothing posts until you say yes. You stay in the driver's seat; Claude just does the typing.
💡 Want your whole content + lead system built for you — not just LinkedIn, but the engine that turns posts into booked calls? That's what I build for creators and practices. DM me BUILD on Instagram and I'll show you what it looks like.