← 返回 Skills 市场
10madh

LinkedIn Bulk Connect

作者 10Madh · GitHub ↗ · v1.1.0
cross-platform ⚠ suspicious
1199
总下载
1
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install linkedin-bulk-connect
功能描述
Send LinkedIn connection requests to a list of people via browser automation and track status in a CSV/TSV file. Use when the user wants to bulk-connect with a list of people on LinkedIn (founders, speakers, leads, etc.) from a spreadsheet or list containing LinkedIn profile URLs. Handles Connect button, Follow-mode profiles, already-connected detection, stale URL fallback via LinkedIn search and Google search, and incremental status tracking.
使用说明 (SKILL.md)

LinkedIn Connect

Automates sending LinkedIn connection requests from a list and tracks results in a data file.

⚠️ Pre-flight Checklist — Confirm Before Starting

Before doing anything else, confirm all of the following with the user. Do not proceed until each item is confirmed.

1. Data File

Ask the user to provide their spreadsheet/CSV/TSV file and confirm it has (or can have) these columns:

  • Person/Founder Name — full name of the person to connect with
  • Company/Brand Name — their company or brand (used for search fallback)
  • LinkedIn Profile URL — optional but highly recommended; reduces automation footprint

If the file lacks any column, tell the user which columns are missing and offer to add them.

2. Browser Setup

Ask which browser setup they're using:

Option A — Chrome Browser Relay (recommended for accounts flagged for automation)

  • User must have the OpenClaw Browser Relay Chrome extension installed
  • User opens LinkedIn in their regular Chrome browser and clicks the OpenClaw Relay toolbar icon on that tab (badge turns ON)
  • Use profile="chrome" for all browser tool calls in this mode

Option B — OpenClaw Isolated Browser (openclaw profile)

  • OpenClaw manages a separate Chrome instance
  • On first use, navigate to https://www.linkedin.com and let the user log in; cookies persist across sessions
  • Use profile="openclaw" for all browser tool calls in this mode

Confirm which option they've set up. Default to Option A (Chrome Relay) if the user's account has been flagged or warned about automation.

3. Ready Check

Only proceed once the user says:

  • ✅ File is ready and accessible
  • ✅ Browser is open with LinkedIn logged in (and relay is attached if Option A)

Browser Profile

Set the profile variable based on user's choice in the Pre-flight Checklist:

  • Option A: profile="chrome" — reuse the relay-attached tab; get targetId via browser action=tabs
  • Option B: profile="openclaw" — OpenClaw-managed isolated Chrome instance

Do not mix profiles mid-run. Pick one and use it consistently for every browser tool call.

Data File Setup

Ensure the tracking file has a Connection Status column. If missing, add it:

import csv
rows = []
with open('file.tsv', 'r') as f:
    reader = csv.DictReader(f, delimiter='	')
    fieldnames = reader.fieldnames + ['Connection Status']
    rows = list(reader)
with open('file.tsv', 'w', newline='') as f:
    writer = csv.DictWriter(f, fieldnames=fieldnames, delimiter='	')
    writer.writeheader()
    for row in rows:
        row['Connection Status'] = ''
        writer.writerow(row)

Three-Tier Profile Discovery (Priority Order)

Always try in this order. Move to the next tier only if the current one fails.

Tier 1 — Direct LinkedIn URL (fastest, zero ambiguity)

Navigate directly to the LinkedIn profile URL from the data file.

  • ✅ URL loads → correct profile, proceed to connect
  • ❌ Returns 404 → escalate to Tier 2
  • Skip Tier 1 if no URL is in the data file for this person

Tier 2 — Google Search (reliable fallback, preserves accuracy)

Search Google for "Founder Name" "Brand/Company" linkedin.

  • Navigate to: https://www.google.com/search?q=\x3CName>+\x3CCompany>+linkedin
  • Find the LinkedIn profile link in results (usually first result), click it
  • Once on the profile, proceed to Connect step
  • ⚠️ Only escalate to Tier 3 if Google can't find the right person or returns no LinkedIn result

Tier 3 — LinkedIn People Search (last resort)

Run a LinkedIn people search for the founder + brand directly inside LinkedIn.

  • Navigate to: https://www.linkedin.com/search/results/people/?keywords=\x3CName>+\x3CCompany>
  • Look for inline Connect buttons first; otherwise open the profile from search results
  • Confirm name + headline/company match before connecting
  • ❌ No trustworthy match → mark Profile Not Found

See references/browser-workflow.md for detailed browser steps for each tier.

Connecting on a Profile

Once on the correct profile, two patterns exist:

Pattern A - Direct Connect button visible on profile → click it → confirm dialog → Send without a note

Pattern B - Follow mode (no Connect button, only Follow + Message + More) → click More actions → use selector .artdeco-dropdown__content--is-open to get dropdown → click Invite [Name] to connect → confirm dialog → Send without a note

If neither Connect nor Invite is available → mark Follow Only.

Status Values

Status Meaning
Request Sent Connection request sent this session
Already Connected 1st degree - no action needed
Pending Request already sent previously
Follow Only No Connect option available on this profile
Profile Not Found All three tiers failed
Skipped Intentionally skipped

Multi-founder Rows

When a TSV row has multiple founders, track per-founder status separated by |:

Founder1Slug: Request Sent | Founder2Slug: Already Connected

Rate Limiting & Anti-Detection

⚠️ LinkedIn flags accounts that jump directly between profile URLs. Always visit the feed between profiles — no exceptions.

  • Navigate to /feed/ before every single profile, without exception. See references/browser-workflow.md for the exact call. This is the primary anti-detection measure.
  • Add a short natural pause (2–4 seconds) after loading the feed before navigating to the next profile.
  • If >3 consecutive clean URLs return 404, pause for 10 seconds on the feed before continuing (then fall back to Google/LinkedIn search).
  • Do not open new browser tabs — the relay breaks; reuse the same attached tab for every action.
  • Aim for no more than 20–25 connection requests per session. Stop and tell the user if you're approaching this limit.

Saving Progress

Use a linkedin_progress.json sidecar file:

{ "statuses": { "https://www.linkedin.com/in/username/": "Request Sent" } }

Update the TSV from this dict every 10 profiles or at the end.

References

  • references/browser-workflow.md - Detailed browser steps for all three tiers and both connect patterns
安全使用建议
Before installing/using: (1) Understand this will act using your logged-in LinkedIn session (no explicit API key is required) — the browser-relay extension or OpenClaw-managed browser must be trusted because it can access your pages and cookies. (2) Automating connection requests can trigger LinkedIn detection and may violate LinkedIn policies; test on a spare account and keep conservative rates. (3) Back up the CSV/TSV before running (the skill reads/writes it). (4) Review and consent to any outreach you send on behalf of your account; avoid scraping or sending to people who did not opt in. (5) The README includes a Python snippet to update files but does not declare a Python requirement — ensure your runtime can safely run any local helper code before executing it.
功能分析
Type: OpenClaw Skill Name: linkedin-bulk-connect Version: 1.1.0 The skill bundle automates LinkedIn connection requests, which inherently involves high-risk browser automation and interaction with user-provided data. While the stated purpose is benign, the skill instructs the agent to read/write to a user-provided TSV file (SKILL.md) and construct search URLs using user-provided names and company data (references/browser-workflow.md). This reliance on unsanitized user input for file operations and URL construction presents a vulnerability for potential local file manipulation or search query injection if the agent does not robustly sanitize inputs. The extensive use of browser automation commands also carries inherent risks, even with the included anti-detection measures.
能力评估
Purpose & Capability
The name/description match the runtime instructions: the SKILL.md and reference file give step-by-step browser automation to open profiles, click Connect/Invite, and write per-profile status into a CSV/TSV. There are minor mismatches: the README includes a short Python snippet for adding a column but the skill declares no required binaries or runtime — this is a small documentation gap (not a functional contradiction).
Instruction Scope
The instructions explicitly direct the agent to control the user's browser (navigate, click, take snapshots) and to read/write the user's TSV/CSV file. That is expected for this purpose, but it means page snapshots will capture profile content and the agent will act using the user's authenticated LinkedIn session. The workflow also instructs performing Google searches from the LinkedIn tab and repeatedly visiting the feed to evade LinkedIn detection — these are operational choices that have privacy and account-safety implications but are within the skill's stated scope.
Install Mechanism
This is an instruction-only skill with no install spec and no downloaded code; that is the lowest install risk. It does require the user to install/enable an OpenClaw Browser Relay Chrome extension for one recommended mode, which is a separate trust decision outside the skill bundle.
Credentials
The skill requests no environment variables or external credentials, which is proportionate. However, it relies on an already-authenticated browser tab (cookies/session) and optionally an extension (Browser Relay) that has access to browsing context — effectively the skill will act with your LinkedIn account privileges. This is expected for a browser-automation connector but is an important privacy/credential consideration.
Persistence & Privilege
The skill is not always-enabled and does not request system-wide persistence or modify other skills' configurations. It runs only when invoked and uses the browser tooling invoked during the session.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install linkedin-bulk-connect
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /linkedin-bulk-connect 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.1.0
v1.1.0: mandatory feed navigation before every profile (anti-detection), pre-flight checklist (data file + browser setup confirmation), fixed tier order (Direct URL → Google → LinkedIn search), raised session cap to 20-25 requests, both Chrome Relay and openclaw isolated browser supported.
v1.0.0
Initial release: bulk LinkedIn connection requests with anti-detection (feed-first navigation), pre-flight checklist, three-tier profile discovery (Direct URL → Google → LinkedIn search), Chrome relay + openclaw browser support, incremental CSV/TSV progress tracking.
元数据
Slug linkedin-bulk-connect
版本 1.1.0
许可证
累计安装 0
当前安装数 0
历史版本数 2
常见问题

LinkedIn Bulk Connect 是什么?

Send LinkedIn connection requests to a list of people via browser automation and track status in a CSV/TSV file. Use when the user wants to bulk-connect with a list of people on LinkedIn (founders, speakers, leads, etc.) from a spreadsheet or list containing LinkedIn profile URLs. Handles Connect button, Follow-mode profiles, already-connected detection, stale URL fallback via LinkedIn search and Google search, and incremental status tracking. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 1199 次。

如何安装 LinkedIn Bulk Connect?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install linkedin-bulk-connect」即可一键安装,无需额外配置。

LinkedIn Bulk Connect 是免费的吗?

是的,LinkedIn Bulk Connect 完全免费(开源免费),可自由下载、安装和使用。

LinkedIn Bulk Connect 支持哪些平台?

LinkedIn Bulk Connect 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 LinkedIn Bulk Connect?

由 10Madh(@10madh)开发并维护,当前版本 v1.1.0。

💬 留言讨论