The setup that actually works
Heads up first: pasting the repo link into a Claude chat and asking it to "install this" does nothing — that's a known issue with the pack's own README. Here's what works:
- Claude app — one step, all 11. Customize → Plugins → Add → Add marketplace → enter Jakeschincariol/linkedin-agent-skill → Add. Then type / in any chat and pick linkedin-agent:li-post (plugin installs use the linkedin-agent: prefix). Keep "code execution and file creation" on in Settings — that's what lets the humanizer's scripts run.
- Claude Code — two commands.
/plugin marketplace add Jakeschincariol/linkedin-agent-skill, then /plugin install linkedin-agent.
- Phone or no install — paste one. Open any skill's SKILL.md on GitHub, copy it, paste it at the top of a new chat, then ask for what you want. Works on any plan anywhere; you lose the humanizer's scripts, so /li-human is mostly pointless this way.
On plans: per Anthropic's help center and pricing page, free Claude accounts can add plugins and upload skills. A couple of older Anthropic doc pages still say Pro+, so if your free account is missing the menu, use the paste method.
How it sounds like you (not the humanizer)
The "give it examples of your writing" feature is a file called voice.md, not /li-human. Paste three of your own posts and say “write my voice.md from these” — /li-post reads it before every draft and refuses to invent a voice. On the Claude app, keep that file in a Project so every chat can see it. The humanizer is separate: it runs automatically inside /li-post, removing stock phrases, em-dashes, and hidden characters, then flags structural tells like "It's not X, it's Y."
The student prompts
/LI-POST — CLASS PROJECT TO POST/li-post I'm a [MAJOR] student. Turn this class project into a LinkedIn post: [2-3 sentences on what you built, the problem, and one specific result with a number]. Audience: recruiters and people in [INDUSTRY]. Give me 3 first lines from the hook formulas, then the full post, plain language, no hashtags in the body.
/LI-PROFILE — THE REWRITE/li-profile Here's my current headline, About section, and experience: [PASTE]. I'm a [YEAR] [MAJOR] student targeting [ROLE] internships. Score it out of 100, then rewrite the weakest two sections so a recruiter understands what I can do in 5 seconds.
/LI-DM — THE ALUMNI MESSAGE/li-dm I want to message [NAME], a [SCHOOL] alum who's a [ROLE] at [COMPANY]. We've never spoken. Write a connection note under 200 characters, a first message that asks one specific question about their path, and two short follow-ups spaced a week apart. No "I'd love to pick your brain."
/LI-PLAN — THE WEEK/li-plan I'm a [MAJOR] student targeting [ROLE] internships for [SEASON]. I can post 3-4 times this week. Plan the week: which post ideas (pull from my class projects: [LIST 2-3]), what days and times, and 10 specific people or accounts in [INDUSTRY] I should be commenting on.
The weekly loop
- Once: your voice.md.
- Monday: /li-plan — it recommends four posts a week, not seven, with engage-only days.
- Draft days: /li-post (humanizer runs inside it), then you read it. Every time.
- Daily: /li-comment on 5 of the 10 people — comments are where recruiters notice students.
- Friday: /li-audit — rank the week, keep what worked.
Keeping it honest: this is a third-party MIT pack by Jake Schincariol, not an Anthropic feature — Anthropic doesn't review uploaded skills, so read the skill files before installing. The humanizer is a plain-language editor, not an AI detector; the pack itself says not to treat posts as undetectable. One of its rules ("put the link in the first comment") is reported outdated on LinkedIn. Skills free, Claude access separate. Independent guide, not affiliated with Anthropic or the authors. Verified October 15, 2026.