Canlah Skills Radar
Pheme · KOL 运营专员

LinkedIn Tools

devag7/linkedin-mcp
✓ I/O · 22认领这个员工

LinkedIn Tools is an unofficial community MCP server enabling AI agents to interact with LinkedIn. It supports searching profiles and jobs, reading the feed, accessing profile data, messaging, and gated write operations. Published as an npm package for stdio-based use.

该 server 含写操作工具 (send_message, create_post) — 接入你自己的账号前,先确认它的授权范围。

真实 I/O · 已连通 ✓

启动 server · tools/list · 22 个工具 · linkedin-mcp

下面每个工具的字段、类型、必填项,都是我们真的把 server 启动、tools/list 拉回来的原始 JSON Schema —— 真实的工具契约。注意:下方「示例调用」是按 schema 自动生成的演示,不是真实调用的返回。

whoami

Report server version, browser/login status, and capabilities.

Input(真实 schema)
无参数
示例调用
{
  "tool": "whoami",
  "arguments": {}
}
health_check

Deep health check: cookie login state, a LIVE Voyager probe (confirms the API actually answers, not just that a cookie exists), and today's safety-budget headroom (per-action used/cap/remaining + pending invites).

Input(真实 schema)
无参数
示例调用
{
  "tool": "health_check",
  "arguments": {}
}
close_session

Close the browser context and release resources (kills the Chrome process).

Input(真实 schema)
无参数
示例调用
{
  "tool": "close_session",
  "arguments": {}
}
get_my_profile

Get the authenticated user's own LinkedIn profile (experience, education, headline, summary).

Input(真实 schema)
无参数
示例调用
{
  "tool": "get_my_profile",
  "arguments": {}
}
get_profile

Get a LinkedIn profile by public identifier (the slug in the profile URL, e.g. "satyanadella").

Input(真实 schema)
usernamestringrequired
LinkedIn public identifier / vanity slug, e.g. "williamhgates"
示例调用
{
  "tool": "get_profile",
  "arguments": {
    "username": "…"
  }
}
get_feed

Get recent posts from your LinkedIn home feed (author + post text).

Input(真实 schema)
countinteger
Number of posts (default 10)
示例调用
{
  "tool": "get_feed",
  "arguments": {
    "count": 0
  }
}
get_notifications

Get your recent LinkedIn notifications (headline, time, read state).

Input(真实 schema)
countinteger
Number of notifications (default 20)
示例调用
{
  "tool": "get_notifications",
  "arguments": {
    "count": 0
  }
}
search_people

Search LinkedIn people by keywords. Returns name, headline, location, and public identifier (pass that to get_profile for full details).

Input(真实 schema)
keywordsstringrequired
Search keywords, e.g. "recruiter at Google"
countinteger
Results (default 10)
示例调用
{
  "tool": "search_people",
  "arguments": {
    "keywords": "…"
  }
}
search_jobs

Search LinkedIn jobs by keywords (and optional location). Returns title, location, posted time.

Input(真实 schema)
keywordsstringrequired
Job search keywords, e.g. "software engineer"
location_geo_idstring
Optional LinkedIn geo URN id to scope the location
countinteger
Results (default 10)
示例调用
{
  "tool": "search_jobs",
  "arguments": {
    "keywords": "…"
  }
}
get_inbox

List your recent LinkedIn messaging conversations (title, last activity, unread count).

Input(真实 schema)
无参数
示例调用
{
  "tool": "get_inbox",
  "arguments": {}
}
get_job_details

Get full details for a job posting by its numeric id (the digits in /jobs/view/<id> or from search_jobs jobUrn).

Input(真实 schema)
job_idstringrequired
Numeric job id, e.g. "4423697734"
示例调用
{
  "tool": "get_job_details",
  "arguments": {
    "job_id": "…"
  }
}
search_companies

Search LinkedIn companies by keywords. Returns name + universalName slug (pass that to get_company for full details).

Input(真实 schema)
keywordsstringrequired
Search keywords, e.g. "fintech bangalore"
countinteger
Results (default 10)
示例调用
{
  "tool": "search_companies",
  "arguments": {
    "keywords": "…"
  }
}
get_company

Get a company by its LinkedIn URL slug (e.g. "google", "microsoft"). Returns name, description, website, industry, size, HQ.

Input(真实 schema)
universal_namestringrequired
Company URL slug, e.g. "google"
示例调用
{
  "tool": "get_company",
  "arguments": {
    "universal_name": "…"
  }
}
get_company_posts

Get a company's recent posts by its LinkedIn URL slug (e.g. "google"). Returns post text + a short meta line.

Input(真实 schema)
universal_namestringrequired
Company URL slug, e.g. "google"
countinteger
Posts to return (default 10)
示例调用
{
  "tool": "get_company_posts",
  "arguments": {
    "universal_name": "…"
  }
}
get_company_employees

List employees LinkedIn surfaces for a company (by URL slug). Returns name, headline, and public identifier (feed the slug to get_profile). Prospecting core.

Input(真实 schema)
universal_namestringrequired
Company URL slug, e.g. "anthropicresearch"
countinteger
Employees to return (default 10)
示例调用
{
  "tool": "get_company_employees",
  "arguments": {
    "universal_name": "…"
  }
}
get_pending_invitations

List your pending connection invitations — received (inbound, with the urn/sharedSecret to accept later) and sent (outbound). Read-only.

Input(真实 schema)
directionstring
Which queue to return (default both)
countinteger
Max per direction (default 50)
示例调用
{
  "tool": "get_pending_invitations",
  "arguments": {
    "direction": "received",
    "count": 0
  }
}
get_conversation

Read messages in a LinkedIn conversation by its URN (get the URN from get_inbox).

Input(真实 schema)
conversation_urnstringrequired
Full urn:li:msg_conversation:(...) from a get_inbox result
示例调用
{
  "tool": "get_conversation",
  "arguments": {
    "conversation_urn": "…"
  }
}
connect_with_person

[ALPHA, write] Send a connection request. Gated: requires confirm:true. Returns a structured status (ok | duplicate | already_connected | restricted | quota_exhausted | failed). Counts against the daily connect cap.

Input(真实 schema)
profile_idstringrequired
The fsd_profile id (the ACoAA… part of the profile URN)
messagestring
Optional note (max 300 chars)
confirmboolean
Must be true to actually execute. Omit/false = refuse (safety).
示例调用
{
  "tool": "connect_with_person",
  "arguments": {
    "profile_id": "…"
  }
}
send_message

[ALPHA, write] Send a message. Pass thread_id / conversation_urn to REPLY into an existing conversation (verified); otherwise a new thread is started to recipient_urn (best-known). Gated: requires confirm:true. Returns a structured status. Counts against the daily message cap.

Input(真实 schema)
recipient_urnstring
Recipient member URN (urn:li:fsd_profile:ACoAA…). Required when starting a NEW thread.
thread_idstring
Existing thread id (2-…) or full msg_conversation urn to reply into (preferred over recipient_urn).
messagestringrequired
Message body (multiline supported)
confirmboolean
Must be true to actually execute. Omit/false = refuse (safety).
示例调用
{
  "tool": "send_message",
  "arguments": {
    "message": "…"
  }
}
create_post

[ALPHA, write] Publish a text post to your feed. Gated: requires confirm:true. Returns a structured status.

Input(真实 schema)
textstringrequired
Post text
visibilitystring
Audience
confirmboolean
Must be true to actually execute. Omit/false = refuse (safety).
示例调用
{
  "tool": "create_post",
  "arguments": {
    "text": "…"
  }
}

共 22 个工具,此处展示前 20 个。

综合分

Canlah Score28
归属岗位Pheme · KOL 运营专员
岗位分值×1.05
★ Stars2
↓ 下载量1.5k
工具数(实测)22
npm 包linkedin-mcp-tools
✓ 已连通:字段来自真实的 tools/list(工具清单)。工具本身尚未真实调用。

效果测评卡

进行中 · 待运行
任务成功率
输出质量 (LLM 评审)
延迟
成本 / token
稳定性

真实 I/O 已解决「输入输出是什么」;下一步接 agent 真跑,量化「效果好不好」。

取用 / 安装

npx -y linkedin-mcp-tools
同岗位其他 skill
Reddit Buddy
karanb192/reddit-mcp-buddy
✓ I/O · 5

This Reddit Buddy MCP server by Karan Bansal provides clean, LLM-optimized access to Reddit's public API through five specialized tools for browsing subreddits, searching content, retrieving post details with comments, analyzing user profiles, and explaining Reddit terminology. Built with TypeScript and featuring intelligent caching with adaptive TTL, rate limiting, and optional Reddit OAuth authentication for 10x higher rate limits, it offers robust error handling with detailed user-friendly messages, content filtering options for NSFW posts, and efficient data processing that extracts essential fields while avoiding Reddit's verbose response format. The implementation includes Docker deployment support, both stdio and HTTP transport modes, and comprehensive user analysis capabilities that aggregate posting patterns and subreddit activity, making it ideal for AI assistants that need reliable Reddit integration for content research, social media analysis, and community insights without the complexity of direct API management.

Pheme7241.8k
53
LinkedIn + LiGo
ertiqah/linkedin-mcp-runner
✓ I/O · 10

LinkedIn MCP Runner is a server implementation that enables Claude AI to interact with LinkedIn, allowing users to publish posts, schedule content, analyze chats, and retrieve profile information through a set of specialized tools. Developed by Ertiqah, LLC, this Node.js-based server communicates with LinkedIn's API via a backend service at ligo.ertiqah.com, handling authentication through API keys stored in Claude's configuration. The implementation supports various LinkedIn-specific operations including post generation with multiple variants, profile data retrieval, and post analytics - making it particularly valuable for professionals who want to manage their LinkedIn presence directly through Claude without switching contexts.

Pheme308.0k
43
TwitterAPI.io Docs
dorukardahan/twitterapi-io-mcp
✓ I/O · 7

Provides offline access to complete TwitterAPI.io documentation including 54 API endpoints across 10 categories, authentication guides, rate limiting information, and pricing details. Features fuzzy search with camelCase support and hybrid caching for fast lookups. TwitterAPI.io is a third-party service offering cost-effective Twitter/X data access at 96% lower cost than official Twitter API, with 1000 QPS support and no authentication complexity.

Pheme76.0k
36