Documentation Index

Fetch the complete documentation index at: https://docs.meetsquad.ai/llms.txt Use this file to discover all available pages before exploring further.

Using Squad via MCP

Connect Squad to Claude Desktop and other MCP-compatible tools

The Model Context Protocol (MCP) lets you access Squad's AI capabilities directly within Claude and other compatible applications. Query your product data and get strategic insights without leaving your workflow.

Overview

Squad's MCP server enables AI assistants to:

• Query your product strategy and roadmap
• Analyze opportunities and solutions
• Search through user feedback and insights
• Generate reports from your Squad workspace

This guide shows how to set up and use Squad through MCP.

Why this matters for product teams

* Copy-pasting specs between tools * Manual PRD templates * Context gaps in conversations * Ask Claude **"show me every open opportunity linked to our 'Improve retention' goal"** and get the live list back * **"Create a solution for reducing onboarding drop-off and draft the PRD"** – Claude calls `create_solution`, fills the doc, and drafts requirements for you * Claude can pull the full mission statement, knowledge-base docs and recent feedback on demand, so every reply is grounded in real product data **Bottom line**: MCP turns Claude into an extension of Squad, giving PMs instant, conversational access to strategy artifacts while developers stay in sync with up-to-date requirements.

Under the hood

Squad's remote MCP server provides 30+ tools across these categories:

Tool prefix	Purpose	Typical actions
opportunity_*	Discover and refine product opportunities	create, list, get, update
solution_*	Generate and iterate on solutions	create, list, get, update
goal_*	Track desired business or user goals	create, list, get, update
knowledge_*	Store useful references / research	create, list, get, delete
feedback_*	Manage raw customer feedback	create, list, get, delete
insight_*	Processed insights from feedback	create, list, get, delete
workspace_*	Manage Squad workspaces	get, update, list, select

All tools conform to the MCP JSON schema format so that agents can introspect inputs and outputs automatically. Authentication is handled via OAuth - you'll sign in with your Squad account on first use.

Prerequisites

Before starting, ensure you have:

Active Squad workspace with data Claude Code CLI, Claude Desktop 0.7.2+, or OpenAI Codex

Quick start

Run this command in your terminal:

```bash theme={null}
claude mcp add squad --transport http https://mcp.meetsquad.ai/mcp
```

On first use, you'll be prompted to authenticate via OAuth in your browser. Sign in with your Squad account to authorize access.

Verify the connection:

```bash theme={null}
claude mcp list
```

You can also use the `/mcp` command inside Claude Code to check connection status.

<Note>
  No restart needed - Squad is ready to use immediately after authentication.
</Note>

Remote MCP servers are configured via the Claude Desktop UI, not JSON config files.

<Steps>
  <Step title="Open Settings">
    In Claude Desktop, go to **Settings** (gear icon)
  </Step>

  <Step title="Navigate to Connectors">
    Click on **Connectors** in the settings menu
  </Step>

  <Step title="Add Squad">
    Click **Add Connector** and enter the Squad MCP server URL:

    ```
    https://mcp.meetsquad.ai/mcp
    ```
  </Step>

  <Step title="Authenticate">
    You'll be prompted to sign in with your Squad account via OAuth
  </Step>
</Steps>

Once connected, look for the 🔌 icon in the chat interface to verify Squad is available.

<Note>
  Remote MCP servers require Claude Desktop on Pro, Max, Team, or Enterprise plans.
</Note>

Run this command in your terminal:

```bash theme={null}
codex mcp add squad --url=https://mcp.meetsquad.ai/mcp
```

<Warning>
  There is a known issue with Codex and MCP OAuth authentication. The OAuth flow will fail during authorization. This is caused by a missing `resource` parameter in Codex's OAuth implementation and is expected to be resolved in an upcoming Codex release. See [openai/codex#11292](https://github.com/openai/codex/issues/11292) for status.
</Warning>

Connect using the remote server URL:

```
https://mcp.meetsquad.ai/mcp
```

OAuth configuration is automatically discovered via the server's `/.well-known/oauth-authorization-server` endpoint.

Example prompts

Once connected, you can query Squad data using natural language. Here are some examples to get you started:

Strategy queries

```text Mission and Goals theme={null} "What is our product mission?" "Show me our current goals and progress" "Which goals are we behind on?" ```

"What are our top opportunities?"
"Show opportunities related to user retention"
"Which opportunities have the highest impact?"

"List solutions in development"
"What solutions address checkout friction?"
"Show me the roadmap for Q2"

Analytics queries

```text User Feedback theme={null} "What are users complaining about most?" "Show recent feedback about delivery times" "What's our current NPS score?" ```

"How is our conversion rate trending?"
"Compare this month's metrics to last month"
"Which features have the highest adoption?"

Report generation

```text Status Reports theme={null} "Generate a weekly status update" "Create an executive summary of our strategy" "What did we ship last month?" ```

"Analyze our product-market fit"
"Compare our roadmap to original goals"
"What should we prioritize next?"

Example workflows

These examples show how Claude can analyze your Squad data and provide actionable insights.

**Prompt:** "Can you review our current product strategy and identify gaps?"

Example response:

Based on your data:

CURRENT STRATEGY:
- Mission: "Make food delivery faster and more affordable"
- Active goals: 3 (1 on track, 2 behind)
- Opportunities in backlog: 12
- Solutions in development: 4

KEY GAPS IDENTIFIED:
1. Delivery time goal is 7 minutes behind target
   - Root cause: Restaurant prep delays (8,234 mentions)
   - No active solution addressing this

2. Low user retention (2.1 orders/week vs 3.0 target)
   - Missing lunch-time availability
   - Limited restaurant variety in suburbs

3. Unaddressed opportunities:
   - Group ordering (3,892 requests, no solution)
   - Dietary filters (2,156 requests, no solution)

RECOMMENDATIONS:
1. Prioritize restaurant prep time optimization
2. Expand lunch partnerships
3. Add group ordering to Q3 roadmap

**Prompt:** "Analyze user feedback from the last 30 days and identify trends"

Example response:

FEEDBACK ANALYSIS (Last 30 days):
Total feedback items: 4,827

TOP THEMES:
1. Delivery delays (1,243 mentions, ↑23% from previous)
   - "Food arrives cold"
   - "Waited over an hour"
   - Concentrated in downtown areas

2. Payment issues (892 mentions, ↑45% from previous)
   - "Payment failed but still charged"
   - "Can't save payment method"
   - Spike started after latest app update

3. Restaurant variety (634 mentions, →0% change)
   - "Same restaurants every time"
   - "No healthy options"
   - Consistent complaint, not growing

EMERGING ISSUES:
- App crashes on Android 14 (127 mentions, new)
- Promo codes not applying (89 mentions, ↑200%)

SENTIMENT TREND:
- Overall: 3.8/5 (↓ from 4.1/5)
- Critical driver: Payment failures

**Prompt:** "Based on our data, what should we build next quarter?"

Example response:

Q3 ROADMAP RECOMMENDATIONS:

MUST DO (High impact, addresses critical issues):
1. Fix payment processing bugs
   - Impact: 892 users affected monthly
   - Effort: 1 sprint
   - Reduces churn by ~15%

2. Restaurant prep time prediction
   - Impact: 7-minute delivery reduction
   - Effort: 3 sprints
   - Addresses top complaint

DO NEXT (High value, strategic):
3. Group ordering
   - Impact: $45 AOV increase
   - Effort: 2 sprints
   - 3,892 requests

4. Lunch partnerships
   - Impact: +0.8 orders/week/user
   - Effort: 2 sprints
   - Unlocks new daypart

CONSIDER (Nice to have):
5. Dietary filters
6. Advanced routing v3
7. Loyalty program expansion

Detailed usage examples

These examples show real interactions with the Squad MCP server, including which tools get called and what data is returned.

**User Prompt:**

"What opportunities are in my Squad workspace? Show me the top 3."

What happens:

• Claude calls list_opportunities to retrieve all opportunities
• Analyzes and presents the top opportunities based on status and priority

Example output:

Your Squad workspace has 5 opportunities. Here are the top 3:

1. **Advanced AI Agent Customization Engine** (New)
   - Build customizable AI agent behaviors and predictive prioritization
     capabilities for enterprise workflows
   - Status: New, with solutions already generated

2. **Enterprise CRM Integration Ecosystem** (New)
   - Enable seamless integration with major CRM platforms

3. **Regulatory Compliance Automation Suite** (New)
   - Automate compliance workflows for regulated industries

Tools called: select_workspace (if needed), list_opportunities

**User Prompt:**

"Show me the solutions for the AI Agent Customization opportunity,
including which goals it supports."

What happens:

• Claude calls get_opportunity with the opportunity ID
• Requests relationships: solutions, outcomes (goals), and insights
• Parses the detailed response including PRDs, pros/cons, and goal alignments

Example output:

The **Advanced AI Agent Customization Engine** opportunity has 3 solutions:

**Solution 1: Workflow Designer**
- Visual interface for customizing AI agent decision-making processes
- Pros: Enables enterprise customization without technical expertise,
  adapts to diverse methodologies (SAFe, Scrum)
- Cons: Complex interface may extend time-to-market

**Related Goals:**
1. Accelerate Market Penetration in Enterprise Segment (Priority: 5/5)
2. Strengthen Product Differentiation Through AI Innovation (Priority: 5/5)

Tools called: get_opportunity with relationships

**User Prompt:**

"Show me my top business goals and which opportunities support them."

What happens:

• Claude calls list_goals to get all workspace outcomes
• For high-priority goals, calls get_goal with relationships
• Constructs a hierarchy showing goal → opportunities → solutions

Example output:

**Goal 1: Accelerate Market Penetration** ⭐ Priority 5/5
├─ Opportunity: Advanced AI Agent Customization Engine
│  ├─ Solution: Workflow Designer
│  ├─ Solution: Methodology Templates
│  └─ Solution: Impact Predictor
├─ Opportunity: Enterprise Multi-Knowledge Base Platform
└─ Opportunity: Regulatory Compliance Automation Suite

**Key Insight:** The "Advanced AI Agent Customization Engine"
opportunity supports BOTH of your top-priority goals.

Tools called: list_goals, get_goal with relationships

This demonstrates the core Opportunity Solution Tree (OST) methodology - showing how solutions ladder up to opportunities and ultimately to goals. **User Prompt:**

"Find everything in my workspace related to 'compliance' or 'regulatory'."

What happens:

• Claude calls similarity_search with the query terms
• Squad's vector search finds semantically similar content
• Results ranked by relevance across all entity types

Example output:

Found 8 items related to compliance/regulatory:

**Opportunities:**
- Regulatory Compliance Automation Suite

**Solutions:**
- Methodology Templates - Healthcare compliance features
- Workflow Designer - Compliance guardrails

**Knowledge Documents:**
- SOC 2 Requirements Overview
- GDPR Data Residency Guidelines

Tools called: similarity_search

**User Prompt:**

"The 'AI Agent Customization' opportunity looks interesting.
Can you generate some fresh solution ideas for it?"

What happens:

• Claude calls generate_solutions with the opportunity ID
• Squad's AI Ideation agent analyzes opportunity context
• Generates 3-5 new solution approaches with PRDs, pros/cons

Example output:

✨ Generating solutions for "Advanced AI Agent Customization Engine"...

Squad's Ideation agent is analyzing:
- Opportunity context and supporting insights
- Workspace knowledge base and industry trends
- Related outcomes and success criteria

✅ Complete! 3 new solutions have been added with:
- Detailed descriptions and rationale
- Pros and cons for each approach
- Preliminary PRD with requirements
- Alignment to your strategic goals

Tools called: generate_solutions, optionally get_opportunity to show results

Best practices

Ask for particular metrics or date ranges for better results Combine multiple data sources for comprehensive analysis Schedule weekly strategy reviews with Claude + Squad Always end with "What should we do about this?"

Troubleshooting

**Check connection status:**

  ```bash theme={null}
  claude mcp list
  ```

  Or use `/mcp` inside Claude Code to see server status.

  **Common fixes:**

  1. **Re-add the server**: Remove and re-add
     ```bash theme={null}
     claude mcp remove squad
     claude mcp add squad --transport http https://mcp.meetsquad.ai/mcp
     ```
  2. **Re-authenticate**: If your session expired, you'll be prompted to sign in again
  3. **Check network**: Ensure you can reach `https://mcp.meetsquad.ai`
</Accordion>

<Accordion title="OAuth authentication fails">
  If the browser authentication doesn't complete:

  1. Make sure you're signing in with the correct Squad account
  2. Check that pop-ups aren't blocked in your browser
  3. Try signing into Squad directly at [meetsquad.ai](https://meetsquad.ai) first
</Accordion>

**Check these common issues:**

  1. **Plan requirements**: Remote MCP servers require Pro, Max, Team, or Enterprise plans
  2. **Connector status**: Go to Settings > Connectors to check if Squad is connected
  3. **Re-authenticate**: Remove and re-add the connector if the session expired
</Accordion>

<Accordion title="Can't add connector">
  If you don't see the Connectors option in Settings:

  1. Ensure you're on Claude Desktop 0.7.2 or later
  2. Check that you have a Pro, Max, Team, or Enterprise plan
  3. Try restarting Claude Desktop
</Accordion>

**If Squad queries fail in either client:**

1. Verify the MCP server shows as connected
2. Check that you've completed OAuth authentication
3. Ensure your Squad workspace has data to query
4. Try simpler queries first (e.g., "What is our mission?")

**If you see rate limit errors:**

• Squad MCP allows 100 requests per minute
• Complex queries may use multiple requests
• Wait 60 seconds if you hit limits
• Consider upgrading your Squad plan for higher limits

If Claude is accessing the wrong workspace:

• Ask Claude to "list my Squad workspaces"
• Then ask it to "switch to [workspace name]"
• The server remembers your workspace selection for the session

Security considerations

Squad MCP uses OAuth authentication - your credentials are never stored locally. When you authenticate:

• You sign in via Squad's secure login page
• Claude receives a token scoped to your account
• Sessions expire automatically for security
• You can revoke access anytime from your Squad account settings

Advanced usage

Custom queries

Squad MCP supports complex analytical queries:

"Compare our user retention to industry benchmarks and suggest 
improvements based on what's working for competitors"

"Generate a board presentation covering our Q2 performance, 
Q3 roadmap, and key risks to address"

"Analyze which features correlate with highest user LTV and 
recommend where to double down"

Workflow automation

Combine Squad MCP with other tools:

1. Daily standup prep: "What changed in our metrics since yesterday?"
2. Sprint planning: "Which backlog items align with this sprint's goal?"
3. Investor updates: "Generate a monthly progress report"

Removing Squad MCP

```bash theme={null} claude mcp remove squad ``` ```bash theme={null} codex mcp remove squad ``` Go to **Settings > Connectors**, find Squad, and click **Remove**.

Getting help

Report issues or contribute Learn more about MCP Contact the Squad team