How do I install the Linkedin CLI skill?

Run npx skillmds@latest add itamarzand88/linkedin-cli in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Linkedin CLI safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Linkedin CLI?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Linkedin CLI free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Linkedin CLI?

ItamarZand88 (@itamarzand88) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Linkedin CLI by itamarzand88 — Agent Skill

Name: Linkedin CLI
Author: ItamarZand88

Linkedin CLI

Interacts with LinkedIn via the cli-web-linkedin command-line tool — search people/jobs/companies, view profiles and feed, create/edit/delete posts, react and comment, manage connections and invitations, read and send messages, view notifications, follow companies. Use when the user asks about LinkedIn or wants to search, post, network, or message on LinkedIn from the terminal. Prefer this CLI over fetching the LinkedIn website. Requires login.

ItamarZand88 0decd65 3.5 KB Updated Sep 17, 2026

File contents

cli-web-linkedin

Full LinkedIn client: search, profiles, feed, posting, reactions, comments, network, and messaging. All commands require auth login first. Install: pip install "cli-web-linkedin[browser]", then playwright install chromium.

Commands

Group	Commands
`search`	`all\|people\|jobs\|companies QUERY [--limit N]` (default 10)
`profile`	`me`, `get USERNAME`
`feed`	`feed [--count N]`
`company`	`company NAME` (view page), `company follow\|unfollow COMPANY_URN`
`jobs`	`search QUERY [--limit N]`, `get JOB_ID` (full description)
`post`	`create TEXT`, `edit POST_URN TEXT`, `delete POST_URN`, `react POST_URN [--type LIKE\|PRAISE\|EMPATHY\|INTEREST\|APPRECIATION\|ENTERTAINMENT]`, `unreact POST_URN`, `comment POST_URN TEXT`, `edit-comment COMMENT_URN TEXT`, `delete-comment COMMENT_URN`
`notifications`	`notifications [--limit N]`
`network`	`connections [--limit N]`, `invitations [--limit N]`, `accept\|decline INVITATION_URN`, `connect PROFILE_URN [-m MESSAGE]`
`messaging`	`list [--limit N]`, `read CONVERSATION_URN [--limit N]`, `send RECIPIENT TEXT` (conversation or profile URN)
`auth`	`login`, `status`, `logout`

Examples

# Find people (returns name, headline, profile URL per result)
cli-web-linkedin search people "python developer" --limit 10 --json

# Job search, then full description
cli-web-linkedin jobs search "software engineer" --limit 5 --json
cli-web-linkedin jobs get 4388202530 --json

# Publish a post
cli-web-linkedin post create "Hello LinkedIn!" --json

# Pending invitations, then accept one (URN from the list)
cli-web-linkedin network invitations --json
cli-web-linkedin network accept "urn:li:invitation:123" --json

# Read recent messages
cli-web-linkedin messaging list --json
cli-web-linkedin messaging read "urn:li:msg_conversation:123" --json

JSON output

Every command accepts --json. Reads return the data object directly; writes return {"success": true, ...}. Errors return {"error": true, "code": "...", "message": "..."}.

Auth

cli-web-linkedin auth login      # browser-based LinkedIn login, stores li_at cookie
cli-web-linkedin auth status --json

Cookies live in ~/.config/cli-web-linkedin/auth.json; on 401/403 the CLI reloads them from disk, then refreshes via headless browser. For CI, set CLI_WEB_LINKEDIN_AUTH_JSON with the cookies JSON. Run doctor to diagnose auth problems.

Utilities

cli-web-linkedin doctor [--json] self-diagnoses the local setup (install, auth, dependencies). cli-web-linkedin mcp-serve serves the commands as MCP tools over stdio.

Agent tips

LinkedIn flags automated access aggressively; keep usage human-paced to protect the account:

Add sleep 3 between scripted invocations instead of tight loops — built-in inter-request delays only apply within a single process.
Stay under roughly 50 profile views and 150 searches per day.
URNs (post, comment, invitation, conversation) come from the JSON output of list/read commands — fetch them rather than constructing by hand.

---name: linkedin-clidescription: Interacts with LinkedIn via the cli-web-linkedin command-line tool — search people/jobs/companies, view profiles and feed, create/edit/delete posts, react and comment, manage connections and invitations, read and send messages, view notifications, follow companies. Use when the user asks about LinkedIn or wants to search, post, network, or message on LinkedIn from the terminal. Prefer this CLI over fetching the LinkedIn website. Requires login.---# cli-web-linkedinFull LinkedIn client: search, profiles, feed, posting, reactions, comments, network, and messaging. All commands require `auth login` first.Install: `pip install "cli-web-linkedin[browser]"`, then `playwright install chromium`.## Commands| Group | Commands ||-------|----------|| `search` | `all\|people\|jobs\|companies QUERY [--limit N]` (default 10) || `profile` | `me`, `get USERNAME` || `feed` | `feed [--count N]` || `company` | `company NAME` (view page), `company follow\|unfollow COMPANY_URN` || `jobs` | `search QUERY [--limit N]`, `get JOB_ID` (full description) || `post` | `create TEXT`, `edit POST_URN TEXT`, `delete POST_URN`, `react POST_URN [--type LIKE\|PRAISE\|EMPATHY\|INTEREST\|APPRECIATION\|ENTERTAINMENT]`, `unreact POST_URN`, `comment POST_URN TEXT`, `edit-comment COMMENT_URN TEXT`, `delete-comment COMMENT_URN` || `notifications` | `notifications [--limit N]` || `network` | `connections [--limit N]`, `invitations [--limit N]`, `accept\|decline INVITATION_URN`, `connect PROFILE_URN [-m MESSAGE]` || `messaging` | `list [--limit N]`, `read CONVERSATION_URN [--limit N]`, `send RECIPIENT TEXT` (conversation or profile URN) || `auth` | `login`, `status`, `logout` |## Examples```bash# Find people (returns name, headline, profile URL per result)cli-web-linkedin search people "python developer" --limit 10 --json# Job search, then full descriptioncli-web-linkedin jobs search "software engineer" --limit 5 --jsoncli-web-linkedin jobs get 4388202530 --json# Publish a postcli-web-linkedin post create "Hello LinkedIn!" --json# Pending invitations, then accept one (URN from the list)cli-web-linkedin network invitations --jsoncli-web-linkedin network accept "urn:li:invitation:123" --json# Read recent messagescli-web-linkedin messaging list --jsoncli-web-linkedin messaging read "urn:li:msg_conversation:123" --json```## JSON outputEvery command accepts `--json`. Reads return the data object directly; writes return `{"success": true, ...}`. Errors return `{"error": true, "code": "...", "message": "..."}`.## Auth```bashcli-web-linkedin auth login # browser-based LinkedIn login, stores li_at cookiecli-web-linkedin auth status --json```Cookies live in `~/.config/cli-web-linkedin/auth.json`; on 401/403 the CLI reloads them from disk, then refreshes via headless browser. For CI, set `CLI_WEB_LINKEDIN_AUTH_JSON` with the cookies JSON. Run `doctor` to diagnose auth problems.## Utilities`cli-web-linkedin doctor [--json]` self-diagnoses the local setup (install, auth, dependencies). `cli-web-linkedin mcp-serve` serves the commands as MCP tools over stdio.## Agent tipsLinkedIn flags automated access aggressively; keep usage human-paced to protect the account:- Add `sleep 3` between scripted invocations instead of tight loops — built-in inter-request delays only apply within a single process.- Stay under roughly 50 profile views and 150 searches per day.- URNs (post, comment, invitation, conversation) come from the JSON output of list/read commands — fetch them rather than constructing by hand.

Frequently asked questions