Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
This collection provides a suite of specialized tools for content marketers to manage their LinkedIn presence. It streamlines the process of drafting posts and carousels using established hook formulas, while organizing weekly content planning.
Beyond creation, the skill set includes dedicated workflows for managing inbox interactions, drafting direct messages, and handling comments to maintain consistent professional engagement.
WHO IT'S FOR
Content Marketers
automating LinkedIn content creation and planning
Social Media Managers
managing engagement and inbox interactions
Personal Brand Strategists
optimizing profile presence and content performance
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Follow the documented setup, then try a first task.
I’m using
Install the repository bundle through Claude
Installs: Complete eleven-skill LinkedIn agent bundle · Claude
Before you start
Claude
Enter this complete block in a Claude session. The URL identifies the repository, and the following sentence asks Claude to install it and confirm that li-post works.
https://github.com/Jakeschincariol/linkedin-agent-skill
Install this skill, then confirm /li-post works.
After installation, configure the shared voice file: copy `templates/voice.md` to `~/.claude/linkedin/voice.md` and fill it in, or paste three of your own posts into Claude and ask it to write `voice.md` from them.
Clone and install the bundle globally in Claude Code
Installs: Complete eleven-skill LinkedIn agent bundle in the user's global Claude skills directory · Claude Code
Before you start
Claude Code
Git already installed and available in the terminal
Run this complete two-line block in a terminal. It clones the repository and copies every `li-*` skill folder, including each folder's supporting files, into the global Claude skills directory.
Configure the shared voice file by copying `templates/voice.md` to `~/.claude/linkedin/voice.md` and filling it in, or by pasting three of your own posts into Claude and asking it to write `voice.md` from them.
Install the LinkedIn agent plugin from its marketplace
Installs: The `linkedin-agent` plugin containing the complete eleven-skill LinkedIn agent bundle · Claude Code
Before you start
Claude Code with plugin marketplace support
Enter this complete two-line block in a Claude Code session. The first command adds the repository as a marketplace; the second installs its `linkedin-agent` plugin.
Configure the shared voice file by copying `templates/voice.md` to `~/.claude/linkedin/voice.md` and filling it in, or by pasting three of your own posts into Claude and asking it to write `voice.md` from them.
The full text or a screenshot of one LinkedIn post
The author's name and role, if known
Any genuine data, experience, disagreement, or question the comment may draw on
Use the li-comment skill. I will paste one LinkedIn post, the author's name and role if known, and any relevant firsthand fact or opinion I can honestly contribute. Draft two comment options of different supported types, label each type, and recommend one with a one-line reason. Do not post either comment or use browser automation.
Score a LinkedIn profile and fix its highest-priority section
Uses li-profile
Current headline
About section
Current role and the last two experience entries
Whether the profile has a banner and a featured section
Optionally, a screenshot of the top card for a first pass
Use the li-profile skill. Score my pasted LinkedIn profile against its 12-part, 100-point rubric and show the score table. Then identify the section losing the most points and rewrite only that section first, giving three options if it is the headline. Do not log into LinkedIn or save any changes.
A pasted batch of LinkedIn messages or screenshots
Enough sender context to distinguish leads, recruiters, peers, asks, and spam
Any relevant details needed to answer genuine questions accurately
Use the li-inbox skill. I will paste a small batch of LinkedIn connection requests or DMs. Count and group them into LEAD, RECRUITER, PEER, ASK, and SPAM; briefly explain any detected automated sequence; and draft replies only for messages worth answering. Do not access my account, archive anything, or send any reply.
Claude