MCP Integrations¶

~/Library/Application Support/Claude/mcp_servers.json

%APPDATA%\Claude\mcp_servers.json

~/.config/claude/mcp_servers.json

You: Can you analyze the performance of my user queries?

Claude: I'll analyze your MySQL database performance. Let me start by examining the current query patterns:

[MCP executes: SHOW PROCESSLIST and EXPLAIN for common queries]

Based on the analysis of your database:

**Current Performance Issues:**
1. **Slow Query on users table**: The login query is doing a full table scan
   - Missing index on email column
   - Query time: 2.3 seconds average

2. **High Connection Count**: 47/50 connections in use
   - Several long-running queries blocking others

**Recommended Optimizations:**
```sql
-- Add index for login queries
CREATE INDEX idx_users_email ON users(email);

-- Add composite index for user search
CREATE INDEX idx_users_search ON users(email, status, created_at);

### PostgreSQL Integration

**Configuration:**
```json
{
  "mcpServers": {
    "postgres": {
      "command": "npx", 
      "args": ["-y", "@anthropic-ai/claude-mcp-postgres"],
      "env": {
        "POSTGRES_URL": "postgresql://username:password@localhost:5432/dbname"
      }
    }
  }
}

You: My PostgreSQL database is running slowly. Can you help diagnose the issue?

Claude: I'll perform a comprehensive PostgreSQL performance analysis:

[Executes diagnostic queries through MCP]

**Database Health Summary:**
- **Size**: 15.3 GB with 2.1M users, 8.4M posts
- **Connections**: 23/100 active connections
- **Slow Queries**: 12 queries averaging >1 second

**Critical Issues Found:**

1. **Bloated Tables** (High Impact):
   ```sql
   Table: user_sessions (89% bloat, 2.1GB wasted space)
   Recommendation: VACUUM FULL user_sessions;
   ```

2. **Missing Indexes** (High Impact):
   ```sql
   -- Query: SELECT * FROM posts WHERE user_id = ? ORDER BY created_at DESC
   -- Execution time: 1,847ms
   -- Recommendation: Add composite index
   CREATE INDEX idx_posts_user_created ON posts(user_id, created_at DESC);
   ```

3. **Lock Contention** (Medium Impact):
   - Long-running transactions holding exclusive locks
   - Average lock wait time: 340ms

**Immediate Actions:**
1. Create the missing indexes (5 min downtime)
2. Schedule VACUUM FULL during maintenance window
3. Optimize the top 3 slowest queries

Should I show you the specific query optimizations and help create a maintenance plan?

You: I'm building a local analytics tool with SQLite. Help me design the schema and test some queries.

Claude: I'd be happy to help design your SQLite analytics schema! Let me examine your current database structure:

[Analyzes existing tables through MCP]

**Current Schema Analysis:**
Your database has 3 tables but I see some opportunities for optimization:

**Proposed Analytics Schema:**
```sql
-- Events table (fact table)
CREATE TABLE events (
    id INTEGER PRIMARY KEY,
    user_id TEXT NOT NULL,
    event_type TEXT NOT NULL,
    timestamp INTEGER NOT NULL,  -- Unix timestamp for efficient queries
    properties JSON,             -- Flexible event data
    session_id TEXT
);

-- Users table (dimension table) 
CREATE TABLE users (
    id TEXT PRIMARY KEY,
    created_at INTEGER NOT NULL,
    properties JSON
);

-- Sessions table (for session analytics)
CREATE TABLE sessions (
    id TEXT PRIMARY KEY,
    user_id TEXT NOT NULL,
    started_at INTEGER NOT NULL,
    ended_at INTEGER,
    properties JSON,
    FOREIGN KEY (user_id) REFERENCES users(id)
);

## 📁 File System Integration

### File Operations MCP

**Configuration:**
```json
{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@anthropic-ai/claude-mcp-filesystem"],
      "args": ["/Users/yourname/projects", "/Users/yourname/documents"]
    }
  }
}

You: Can you analyze my project structure and suggest improvements?

Claude: I'll analyze your project structure using the file system integration:

[Browses project files through MCP]

**Project Structure Analysis:**

**Current Structure:**

**Issues Identified:**

1. **Test Coverage Gap**: Only 3 test files for 82 source files
   - Recommendation: Add tests for each major component
   - Target: At least 80% test coverage

2. **Component Organization**: 47 files in flat components directory
   - Recommendation: Group by feature or function
   ```
   components/
   ├── common/          (buttons, inputs, layouts)
   ├── user/           (profile, settings, auth)
   ├── dashboard/      (widgets, charts, analytics)
   └── admin/          (user management, system config)
   ```

3. **Missing Configuration Files**:
   - No ESLint configuration
   - No TypeScript configuration
   - No CI/CD setup

**Proposed Improvements:**
Let me create a better structure and move files appropriately:

[Executes file operations through MCP]

**Restructured Project:**
- Created feature-based component organization
- Added missing configuration files
- Set up proper test directory structure
- Added documentation templates

**Next Steps:**
1. Review the new structure
2. Update import paths in components
3. Add ESLint and Prettier configurations
4. Set up automated testing pipeline

Would you like me to help update the import statements throughout your codebase to match the new structure?

You: Can you analyze our team's commit patterns and suggest improvements to our Git workflow?

Claude: I'll analyze your Git repository to understand commit patterns and workflow effectiveness:

[Analyzes Git history through MCP]

**Repository Analytics (Last 90 Days):**

**Commit Patterns:**
- **Total Commits**: 347
- **Contributors**: 5 developers
- **Average commits/day**: 3.9
- **Peak activity**: Tuesdays and Wednesdays

**Code Quality Indicators:**
1. **Commit Message Analysis**:
   - 23% follow conventional commits format
   - 41% are single-word messages (concerning)
   - Most common: "fix", "update", "changes"

2. **File Change Patterns**:
   - `src/utils.js` modified in 67% of commits (potential hotspot)
   - Large commits (>50 files): 12 instances
   - Binary file commits: 23 (should use Git LFS)

**Branch Strategy Analysis:**
- **Main branch**: 89 direct commits (should be protected)
- **Feature branches**: Average lifespan 4.3 days
- **Merge strategy**: Mix of merge commits (67%) and rebases (33%)

**Problematic Patterns:**
1. **Hotspot Files** (high change frequency):
   ```
   src/utils.js        - 234 changes
   src/config.js       - 189 changes  
   package.json        - 156 changes
   ```

2. **Large Commits** (refactoring candidates):
   - `commit abc123`: 127 files changed
   - `commit def456`: 89 files changed

3. **Merge Conflicts** (last 30 days): 18 instances
   - Most conflicts in: `src/components/`, `package-lock.json`

**Recommendations:**

1. **Improve Commit Quality**:
   ```bash
   # Add commit message template
   git config commit.template .gitmessage

   # Template content:
   # <type>(<scope>): <subject>
   # 
   # <body>
   # 
   # <footer>
   ```

2. **Protect Main Branch**:
   - Require PR reviews
   - Block direct pushes
   - Require status checks

3. **Address Hotspots**:
   - Refactor `utils.js` into smaller modules
   - Extract configuration to separate files
   - Consider architectural improvements

4. **Standardize Workflow**:
   - Use squash merges for feature branches
   - Implement conventional commits
   - Add pre-commit hooks for quality checks

Would you like me to help implement any of these improvements, or analyze specific aspects of your codebase in more detail?

You: Can you help me optimize my AWS costs? My bill has been increasing steadily.

Claude: I'll analyze your AWS infrastructure and identify cost optimization opportunities:

[Connects to AWS through MCP and analyzes resources]

**AWS Cost Analysis Summary:**

**Current Monthly Spending**: $2,847
**Trend**: +23% increase over last 3 months

**Top Cost Drivers:**
1. **EC2 Instances**: $1,245/month (44%)
   - 12 t3.large instances running 24/7
   - CPU utilization average: 23% (overprovisioned)

2. **RDS Databases**: $678/month (24%)
   - Production: db.r5.xlarge (necessary)
   - Staging: db.r5.large (can downsize)
   - Development: db.t3.medium (can shut down nights/weekends)

3. **S3 Storage**: $432/month (15%)
   - 15TB in Standard storage
   - 8TB hasn't been accessed in 90+ days (IA candidate)

4. **Data Transfer**: $289/month (10%)
   - High inter-AZ transfer costs
   - CloudFront might reduce costs

**Optimization Recommendations:**

**Immediate Actions** (Est. savings: $890/month):

1. **Right-size EC2 Instances**:
   ```bash
   # Current: 12 × t3.large ($67/month each)
   # Recommended: 8 × t3.medium + 4 × t3.small
   # Savings: $312/month
   ```

2. **Implement Auto Scheduling**:
   ```bash
   # Development/staging environments
   # Run only during business hours (9AM-6PM, Mon-Fri)
   # Savings: $245/month
   ```

3. **S3 Storage Optimization**:
   ```bash
   # Move infrequently accessed data to IA
   # 8TB × $0.0125/GB vs $0.023/GB
   # Savings: $102/month
   ```

4. **RDS Optimization**:
   - Downsize staging: db.r5.large → db.t3.large
   - Schedule dev environment: 40 hours/week vs 168
   - **Savings**: $231/month

**Medium-term Actions** (Est. savings: $450/month):

1. **Reserved Instances**: Purchase 1-year RIs for production
2. **CloudFront**: Reduce data transfer costs
3. **EBS Optimization**: Convert unused volumes to gp3

**Implementation Plan:**
Let me help you implement these changes safely:

1. **Week 1**: Resize non-production instances
2. **Week 2**: Implement auto-scheduling
3. **Week 3**: S3 lifecycle policies
4. **Week 4**: Purchase reserved instances

Would you like me to help create the specific scripts and policies for any of these optimizations?

You: Can you analyze the code quality of my main application file?

Claude: I'll use the custom file processor to analyze your code quality:

[Executes custom MCP tool]

**Code Quality Analysis Results:**

**File**: `src/main.js`

**ESLint Issues**:
- 3 errors, 7 warnings
- Most common: unused variables (4 instances)
- Security concerns: eval() usage detected

**Complexity Metrics**:
- Cyclomatic complexity: 23 (high)
- Lines of code: 847
- Function count: 34
- Average function length: 25 lines

**Recommendations**:
1. **Break down large functions**: 3 functions exceed 50 lines
2. **Remove unused code**: 4 unused variables detected
3. **Eliminate eval()**: Security risk on line 234
4. **Add error handling**: 12 functions lack try-catch blocks

**Priority Actions**:
1. Refactor the `processUserData` function (complexity: 15)
2. Remove eval() and use safer alternatives
3. Add JSDoc comments for public functions
4. Extract utility functions to separate modules

Would you like me to help implement these improvements or analyze another file?