Yadulink / Documentation / MCP OAuth

Limit MCP permissions before letting AI act on your LinkedIn data.

A remote MCP server should remain explicit: who connects, which data is accessible, which actions are allowed, and how tokens are controlled.

View MCP page All guides

MCP OAuth

Least privilege

Signal analysis does not need the same rights as a sequence action. Separate permissions by use case to reduce operational risk.

MCP OAuth

Action traceability

Sensitive actions should stay visible: tool called, connected account, date, action object, and observed result.

Verified facts

What this documentation should make explicit.

Purpose

A remote MCP server should remain explicit: who connects, which data is accessible, which actions are allowed, and how tokens are controlled.

Workflow proof

Signal analysis does not need the same rights as a sequence action. Separate permissions by use case to reduce operational risk.

Assistant prompt

Help me choose the minimum OAuth scopes for using Yadulink MCP with an AI assistant without over-granting permissions.

Verification method

How to turn the guide into an operational test.

  1. 01 Permission checklist List the actions truly needed for the workflow.
  2. 02 Permission checklist Choose the most restrictive scopes possible.
  3. 03 Permission checklist Test a read-only action before any sending or modification.

Ask your preferred AI.

ChatGPT Claude Perplexity Codex

Docs linking

Guides to read next.

MCP

Set up Yadulink MCP with Claude, ChatGPT, and AI assistants

Connect Yadulink to Claude, ChatGPT, or an MCP-compatible assistant.

API

Yadulink API authentication: keys, scopes, and best practices

Use Yadulink API keys without exposing your LinkedIn account.

Webhooks

Yadulink webhooks: sync replies, reads, and LinkedIn signals

Turn LinkedIn events into CRM actions and sales alerts.