How to Connect Claude Code to Linear, Jira, Slack, and Your Entire PM Stack

MCP (Model Context Protocol) lets Claude Code connect directly to your PM tools — Linear, Notion, Slack, PostHog, HubSpot — so AI can pull live data into your analysis, create tickets from your PRDs, and summarize what happened in Slack while you were in back-to-back meetings. This guide covers the five highest-impact integrations for product managers and how to set each one up.

If you've been using Claude Code for PM work, you've probably hit a wall. Claude can write a solid PRD. It can analyze interview transcripts. It can build competitive profiles. But at some point, you still have to copy data out of Linear, paste metrics from PostHog, or manually summarize a Slack thread before Claude can do anything useful with it.

MCP removes that wall. Instead of you being the middleman between your tools and your AI, Claude connects directly.

What Is MCP and Why Should PMs Care?

MCP stands for Model Context Protocol. In plain terms, it's a standard that lets Claude Code talk to other software — your project management tool, your documentation wiki, your analytics platform, your CRM.

Think of it like this: CLAUDE.md gives Claude your product context (who your users are, what you're building, your competitive landscape). MCP gives Claude access to your live operational data — the sprint board, the analytics dashboard, the Slack channel where your engineering lead just flagged a blocker.

Without MCP, Claude is smart but isolated. It knows your product context but can't see what's actually happening across your tools. With MCP, Claude can pull a sprint velocity report from Linear, check your conversion funnel in PostHog, and draft a stakeholder update — in one conversation.

Why this matters for PMs specifically:

Product managers live across more tools than almost any other role. Your day touches project management, documentation, communication, analytics, and CRM. Every time you context-switch between tools to gather information, you lose time. MCP collapses that.

"Context spread across different tools and systems — which takes a lot of time and effort to address. I want to be able to pull in all related content about an issue — with context on prior work done — to convert into an insight and actionable next steps."

That's the exact problem MCP solves. Not by replacing your tools, but by giving Claude a direct line to them.

The 5 MCP Integrations Every PM Should Set Up

Not every MCP integration is equally valuable for PM work. Here are the five that unlock the most leverage, ranked by impact.

Let's break down what each one actually does for your PM workflow.

1. Project Management (Linear or Jira)

This is the highest-impact integration for most PMs. Your project management tool is where execution lives — and it's where PMs spend a disproportionate amount of time on manual work.

What it unlocks:

• Pull sprint data into reports. Ask Claude to summarize what shipped this sprint, what's blocked, and what rolled over. No more manually compiling status updates.
• Auto-create tickets from PRDs. Write a PRD with Claude, then tell it to create the corresponding Linear issues — with descriptions, labels, and estimates already filled in.
• Track velocity trends. Ask Claude to pull the last 4 sprints of data and identify patterns. Are you consistently overcommitting? Which types of work take longer than estimated?
• Query your backlog. "What are the open bugs tagged 'payments'?" or "Show me all P1 issues assigned to the mobile team." Claude pulls the answer directly from Linear.

"I need to automate the creation of my Sprint and track metrics using agents."

Example prompt after connecting Linear: "Pull all issues completed in the last sprint, group them by epic, and draft a sprint summary I can share with the engineering team."

2. Documentation (Notion or Confluence)

Your documentation tool is your product's institutional memory. The problem is that most PMs can't search it effectively — and Claude definitely can't access it without MCP.

What it unlocks:

• Search across all docs. Ask Claude to find every document that mentions a specific feature, customer segment, or technical decision. It searches Notion directly.
• Pull specs into context. Writing a PRD for v2 of a feature? Claude can pull the v1 spec, the original design rationale, and the post-launch retro — then use all of it as context.
• Reference past decisions. "What did we decide about pricing tiers in Q3?" Claude searches your Notion workspace and finds the decision doc.
• Sync decisions to docs. After a planning session, Claude can create or update a Notion page with the outcomes.

Example prompt after connecting Notion: "Search our product specs for everything related to the onboarding flow. Summarize the current design, any known issues documented in retros, and what users said about it in the last round of interviews."

3. Communication (Slack)

Slack is where context goes to die. Important decisions get buried in threads. Blockers get flagged and forgotten. Customer feedback shows up in #support and never makes it to the product team.

What it unlocks:

• Summarize channel activity. "What happened in #product-engineering this week?" Claude reads the channel and gives you the highlights.
• Pull thread context. Working on a decision? Claude can pull the full Slack thread where the team debated the options, so you don't have to scroll back through 47 messages.
• Surface blockers. Ask Claude to scan engineering channels for anything flagged as blocked or at risk.
• Post updates. After generating a sprint report or stakeholder update, Claude can post it directly to the relevant channel.

"Currently I am trying to build a good approach to consolidate all the user feedback because right now it is spread across multiple tools — customer calls in Chorus, Slack threads, support tickets in Salesforce, internal feedback in Jira. It is difficult to find the patterns and put them all together."

Slack MCP won't solve all of that — but it eliminates one of the biggest fragmentation points. Customer feedback surfaced in Slack threads becomes searchable and synthesizable.

Example prompt after connecting Slack: "Pull all messages from #customer-feedback in the last 2 weeks. Group the feedback by theme and identify the top 3 patterns."

4. Analytics (PostHog or Amplitude)

PMs who can pull live metrics into their analysis — without opening a dashboard, exporting a CSV, or asking a data analyst — move faster. Period.

What it unlocks:

• Pull live metrics. "What's our 7-day activation rate?" Claude queries PostHog directly and returns the number.
• Funnel analysis. "Show me the signup-to-first-action funnel for the last 30 days." Claude pulls the funnel data and identifies where users are dropping off.
• Experiment results. Running an A/B test? Claude can pull the current results, calculate significance, and recommend whether to ship or kill it.
• Metric trends for reviews. Preparing for a product review? Claude pulls the key metrics, compares them to the previous period, and flags anything that moved significantly.

"Unified platform metrics to drive faster discovery and enable agility to deliver features for customers."

Example prompt after connecting PostHog: "Pull our weekly active users, activation rate, and feature adoption for the dashboard redesign. Compare this week to the prior 4-week average and flag anything that moved more than 10%."

5. CRM (HubSpot or Salesforce)

Most PMs don't touch the CRM directly. But the CRM contains some of the most valuable product intelligence in the company — win/loss data, pipeline signals, and customer feedback from sales calls.

What it unlocks:

• Win/loss patterns. "What are the top 3 reasons we lost deals last quarter?" Claude searches closed-lost opportunities and identifies patterns in the loss notes.
• Pipeline signals for roadmap. "Which feature requests are showing up most in active deals?" Claude scans deal notes and surfaces what sales is hearing.
• Customer feedback loop. Sales reps log feedback in HubSpot that never reaches the product team. MCP makes it searchable.
• Account context for customer calls. Before a customer call, Claude can pull the account's deal history, support tickets, and feature requests — so you walk in prepared.

Example prompt after connecting HubSpot: "Search closed-lost deals from last quarter. Group the loss reasons by category and tell me which product gaps are costing us the most revenue."

How to Connect Your First MCP Server

Setting up MCP is simpler than it sounds. You don't need to write code. Here's the step-by-step for connecting your first integration.

Step 1: Add the MCP server

In Claude Code, use the claude mcp add command to register a new MCP server. For example, to add Linear:

claude mcp add linear -- npx -y @mseep/linear-mcp

Claude Code will prompt you for any required environment variables (like API keys).

Step 2: Get your API key

Most PM tools require an API key for MCP access:

• Linear: Settings > API > Personal API Keys
• PostHog: Settings > Personal API Keys
• HubSpot: Settings > Integrations > Private Apps
• Notion: Settings > Integrations > Create new integration

Step 3: Test the connection

Start a new Claude Code session and ask a simple question that requires the integration:

• Linear: "List my open issues"
• Notion: "Search for pages about onboarding"
• PostHog: "What are my top events by volume this week?"

If Claude returns real data from your tool, you're connected.

Step 4: Add to your project's .mcp.json for team use

For shared team setups, add the MCP configuration to your project's .mcp.json file so every PM gets the same integrations:

{
  "mcpServers": {
    "linear": {
      "command": "npx",
      "args": ["-y", "@mseep/linear-mcp"],
      "env": {
        "LINEAR_API_KEY": "your_key_here"
      }
    }
  }
}

Each team member substitutes their own API key. The configuration stays consistent.

Advanced: Chaining MCP Integrations with Skills

MCP on its own is useful. MCP combined with PM skills becomes a force multiplier. Here's what chaining looks like in practice.

Sprint review workflow:

1. Claude pulls completed issues from Linear (MCP)
2. Claude pulls relevant metrics from PostHog (MCP)
3. Claude runs /weekly-metrics to format the data into a structured report
4. Claude posts the summary to Slack (MCP)

Competitive research workflow:

1. Claude pulls recent deal notes from HubSpot mentioning competitor names (MCP)
2. Claude runs /competitive-profile-builder to update competitor profiles
3. Claude saves the updated profiles to Notion (MCP)

Roadmap planning workflow:

1. Claude pulls backlog items and velocity data from Linear (MCP)
2. Claude pulls feature request patterns from HubSpot (MCP)
3. Claude runs /roadmap-builder with real data, not guesses
4. Claude creates the roadmap document in Notion (MCP)

This is the difference between AI as a writing tool and AI as a PM operating system. The writing tool takes your input and produces output. The operating system connects to your data, runs your workflows, and delivers results.

What Becomes Possible

Once your PM stack is connected through MCP, the nature of your AI interactions changes. You stop being the middleman who copies data between tools. Instead, you become the decision-maker who asks questions and gets answers grounded in real data.

Before MCP: "Let me export this sprint data from Linear, paste it into Claude, also grab the metrics from PostHog, and then ask Claude to write a stakeholder update."

After MCP: "Write a stakeholder update using this sprint's Linear data and our PostHog metrics. Post it to #product-updates in Slack."

That's not a marginal improvement. That's a workflow that used to take 45 minutes happening in 2 minutes.

The PMs who set up these integrations now — while MCP is still new and most teams haven't heard of it — will have a compounding advantage. Every week of live data flowing into their PM workflows is a week of better decisions, faster reports, and less time lost to tool-switching.

Frequently Asked Questions

What is MCP in the context of Claude Code and product management?

MCP (Model Context Protocol) is a standard that lets Claude Code connect directly to PM tools like Linear, Notion, Slack, PostHog, and HubSpot. Instead of copying data between tools and pasting it into AI, MCP gives Claude read and write access to your live operational data, so it can pull sprint data, query metrics, search documentation, and post updates — all within a single conversation.

Do I need to be a developer to set up MCP integrations?

No. The Claude Code connectors UI handles most setups without touching the terminal. You search for the integration, click connect, and authenticate with your account. Some tools require generating an API key from their settings page, but the process takes minutes, not hours. No coding is required.

Is my data safe when I connect tools through MCP?

MCP connections use the same authentication and permissions as the underlying tool's API. Claude can only access what your API key or OAuth token allows. If your Linear API key has read-only access, Claude can query issues but can't modify them. You control the permission scope for every integration.

Which MCP integration should I set up first?

Start with your project management tool (Linear or Jira). It has the highest immediate impact — pulling sprint data, creating tickets from PRDs, and querying your backlog are workflows every PM does daily. Analytics (PostHog or Amplitude) is the strong second choice if you spend significant time pulling metrics for reviews and reports.

Can I chain multiple MCP integrations in a single conversation?

Yes. This is where MCP becomes a force multiplier for PM work. In a single Claude Code session, you can pull sprint data from Linear, query conversion metrics from PostHog, and post the resulting stakeholder update to Slack — all without leaving the conversation. Chaining integrations with PM skills like /weekly-metrics or /roadmap-builder turns Claude from a writing tool into a PM operating system that works across your entire stack.

mySecond's 70+ PM skills are designed to work with MCP integrations — pulling live data from Linear, PostHog, and your CRM into every workflow. Browse the skills at mysecond.ai/skills.

Ron Yang is a product leader and the founder of mySecond, the PM Operating System built on Claude. He builds PM infrastructure for product teams at growing companies.