Search skills/

LinkedIn Marketing Skills Claude skills

12 skills that draft LinkedIn posts, comments, replies and profile copy in your voice, read engagement data via Apify, and publish through Publora after you approve.

sergebulaev/linkedin-skills 26 skills MarketingWriting & editing
Our verdict Recommended with caveats

A well-built, approval-gated bundle with real tests, but it depends on two or three paid third-party services and needs API keys in a local .env file.

3/5 Depth expertise shipped Note Security Review before installing 2d Maintenance Updated 2 days ago 5.3K Adoption stars OK License MIT, commercial OK

What LinkedIn Marketing Skills does

This bundle covers LinkedIn content operations end to end. linkedin-post-writer, linkedin-comment-drafter and linkedin-reply-handler draft content from a URL or topic; linkedin-humanizer rewrites drafts to remove AI tells and can audit a draft with multiple detectors; linkedin-hook-extractor pulls hook formulas from viral posts; linkedin-content-planner plans a week of posts; linkedin-thread-monitor and linkedin-engager-analytics read who replied to or engaged with a post; linkedin-profile-optimizer rewrites a profile; linkedin-employee-advocacy plans a team posting program; linkedin-repurposer adapts content from another platform; and linkedin-interviewer builds a story bank of your own material for the others to draw on so drafts do not invent facts.

Every skill follows draft, then your approval, then publish. Publishing and scheduling go through the Publora API (needs a PUBLORA_API_KEY). Reading posts, comments and engagers (comment drafter, reply handler, hook extractor, engagement monitor) needs an Apify token, or the Apify MCP connector on claude.ai instead of a key. An optional Pixfaro token covers image generation. The repo ships Python client libraries (lib/apify_client.py, lib/publora_client.py, lib/url_parser.py), 29 scripts and 18 tests at the root, is also distributed for Codex and OpenClaw, and the same team publishes matching bundles for X, Instagram, YouTube, TikTok, Threads and Facebook.

Good for

  • Individuals and marketers who want consistent-voice LinkedIn content with a human review step
  • Founders building a content cadence from their own material
  • Teams running employee advocacy on LinkedIn

Not for

  • Anyone who wants fully unattended posting
  • People unwilling to hold Apify, Publora or Pixfaro API keys

What's inside depth 3/5

12 skills (plus platform mirrors) averaging about 9.5 KB, with 29 scripts, 18 tests and reference files such as the voice profile and story bank at the repo root; one eval file.

Skills
26
Reference docs
136
Scripts
47
Tests / evals
22
Templates
4
Avg SKILL.md
9.3 KB

Counted from the repository file tree. How we score depth.

Dependencies and costs

  • Publora API key for publishing/scheduling
  • Apify token (or Apify MCP connector on claude.ai) for reading posts, comments and engagers
  • Optional Pixfaro token for image generation
  • A paid Claude plan with code execution for the plugin route
  • Telemetry: none found beyond the third-party API calls the skills make on your behalf (Apify, Publora, Pixfaro) using your own tokens.

Agent skills, not the skills section on your LinkedIn profile

Most people searching "linkedin skills" want a skills list to add to their LinkedIn profile: top skills, hard and soft skills, and which ones help recruiters find them. This repo is something else. "Skills" here means agent skills, the SKILL.md instruction files that Claude Code, Codex and OpenClaw load to handle a task. The closest thing to profile work in the bundle is linkedin-profile-optimizer, which audits and rewrites a LinkedIn profile. The repo material does not say it picks entries for the skills section, so a career guide will serve you better there. This bundle is for writing, commenting and content marketing on social media, not for job search profile tuning.

A first-run workflow

  1. If you have never posted, start with linkedin-interviewer. It asks you questions and saves the answers in references/story-bank.md, which every writing skill reads. The README says the voice profile needs posts you already wrote, while the story bank only needs a career.
  2. If you have a few past posts, run linkedin-humanizer --mode profile so it learns your voice and fills references/voice-profile.md. Without it, drafts fall back to generic voice rules.
  3. Ask linkedin-post-writer for a draft, then run linkedin-humanizer --mode audit before you approve it.
  4. To comment on someone else's post, paste the post URL into a request like "draft me a comment on this". linkedin-comment-drafter returns 1 to 3 variants of 200 to 350 characters with no hashtags. Each variant comes with a reaction type such as INTEREST or PRAISE and a label for which of its 7 templates it used. It publishes through Publora only after you say "post".

Two skills are useful beyond solo posting. linkedin-repurposer takes a tweet, thread, YouTube video, blog post or newsletter, rewrites the hook for the fold, and expands it to 900 to 1300 characters. linkedin-employee-advocacy plans a team program with a 14-day launch, posting cadence, brand governance and ROI tracking.

Setup gotchas from the README

  • In Claude Desktop, syncing the marketplace only adds the plugin to the Discover tab. You still have to click the + on the plugin card and then confirm under Yours.
  • The .env file must sit in the linkedin-skills/ root. A "Publora API key not provided" error usually means it is missing or in the wrong folder.
  • A 401 from Publora means the key is wrong or expired. Create a new one under Settings > API.
  • On claude.ai, the Python Apify client cannot run, so use the Apify MCP connector with OAuth and keep the ?actors= filter the README specifies.
  • For OpenClaw, add the system prompt from the README that points the agent at lib/url_parser.py, lib/apify_client.py and lib/publora_client.py.

Security review

Review before installing

Twelve LinkedIn marketing skills with Python helpers that read posts via Apify, optionally generate images via Pixfaro and publish or schedule posts and comments via Publora with keys you supply. The docs and code present a draft and wait for your approval before posting; schedule_post.py itself does not prompt, so it relies on the agent having approval first. A custom-poster hook (LINKEDIN_SKILLS_CUSTOM_POSTER) can run a command you set in the environment. Third-party posts and comments the skills read are untrusted text.

Worth knowing

  • Needs an API key for its own serviceAPIFY_TOKEN, PUBLORA_API_KEY, PIXFARO_TOKEN and LINKEDIN_PLATFORM_ID are loaded from .env (only those names) and sent only to api.apify.com, api.publora.com and api.pixfaro.com respectively.lib/_env.py:9
  • Makes network requestsReads LinkedIn posts/comments through Apify actors and publishes, schedules, reacts and comments on your LinkedIn account through Publora once you approve a draft; the humanizer's test_detectors.py can post text to GPTZero, Originality.ai, ZeroGPT, Sapling and Copyleaks with your keys for those services.lib/publora_client.py:73
  • Installs packages or toolsOptional setup is pip install requests python-dotenv, unpinned.README.md:281

What we checked

The repository (228 files) on Oct 29, 2026. Not a full review: we read 9 of 51 code files; it relies on code outside the repository: Apify actors (apimaestro/*, scraping_solutions/* LinkedIn scrapers), Apify MCP at mcp.apify.com, Publora and Pixfaro services; lib/backend_selector.py, apify_client.py, pixfaro_client.py, publora_client.py, check_config.py, selftest.py, test_detectors.py, evals/run_evals.py and the tests were checked with pattern searches and excerpts, not read in full.

NVIDIA SkillSpector static scan: 259 pattern hits (Data Exfiltration 78, Privilege Escalation 66, Dangerous Code Execution 36, Data Flow 24, Supply Chain 23, Tool Misuse 10). The CRITICAL taint hits are API tokens read from env and sent in requests to the matching vendor (Apify, Publora, Pixfaro) or to the AI-detector services the humanizer tests against. 'Concealed executable artifact' hits are scanner noise on the .codex-marketplace mirror of the same files. The .env hits are the loader and docs.

A review, not a certification. Skills change after we check them — read the files before installing.

License

MIT — commercial use allowed.

Can it run unattended?

Not really. Every skill is explicitly draft-then-approve-then-publish; nothing posts without your sign-off.

Alternatives to LinkedIn Marketing Skills

Marketing Skills (Corey Haines)

coreyhaines31/marketingskills

52 marketing skills for CRO, copywriting, SEO, analytics, paid ads and growth, built around a shared product-marketing context file.

Recommended Check first Depth 4/5

Aaron Marketing Skills

aaron-he-zhu/aaron-marketing-skills

120 marketing skills across seven disciplines (narrative, SEO/GEO, social, email, paid ads, influencer, launch) with scored auditor gates and an optional 8-bot team mode.

With caveats Check first Depth 4/5

NotFair Plugin

nowork-studio/notfair-plugin

48 skills for SEO, GEO/AEO, paid media across six ad platforms, WordPress, GoHighLevel and GA4/Search Console analytics, backed by live-data workflows.

Recommended Check first Depth 3/5

FAQ

What is LinkedIn Marketing Skills?

It is a 12-skill bundle by sergebulaev for Claude Code and Codex that drafts LinkedIn posts, comments and replies in your voice, analyzes engagement, and publishes through the Publora API only after you approve each draft.

How do I install it?

In Claude Code run /plugin marketplace add sergebulaev/linkedin-skills then /plugin install linkedin-skills@linkedin-skills, or clone the repo so the bundled .claude/skills/ symlinks load automatically (Windows needs git symlink support enabled).

Do I need API keys?

Publishing needs a Publora API key. Reading posts, comments and engagers needs an Apify token, or the Apify MCP connector if you use claude.ai. Image generation through Pixfaro is optional.

Is it safe to use?

It never publishes without your approval and keeps API keys in a local .env file rather than hardcoded. The main risk is ordinary credential hygiene for the three optional third-party services.

Will this help me pick skills for my LinkedIn profile's skills section?

Not directly. These are agent skills for drafting and publishing LinkedIn content. linkedin-profile-optimizer audits and rewrites your profile, but the repo does not describe it choosing skills-section entries.

How much does the Apify read layer cost?

The README estimates that a typical creator running daily comment work plus a weekly engager-analytics sweep stays under $2 per month, which it says fits inside Apify's free tier.

Can I test a local clone in Codex before installing the published version?

Yes. Clone the repo, cd into linkedin-skills, then run codex plugin marketplace add . followed by codex plugin add linkedin-skills@linkedin-skills.

Sources: GitHub repository (stars, license, file tree, last push Oct 27, 2026). Review based on the README and SKILL.md files as published.