This guide documents all MCP servers configured for optimal development workflow with Claude Code in VSCode.
Configuration File: ~/Library/Application Support/Code/User/settings.json
Purpose: Web content fetching and conversion
What I Can Do:
- Fetch web pages and convert to markdown
- Extract documentation from websites
- Research APIs and libraries
- Get latest package information
Use Cases:
"Fetch the latest Python documentation for asyncio"
"Get information from https://example.com"
"Research the TradingView API"
Configuration:
{
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-fetch"]
}Purpose: GitHub repository management and operations
What I Can Do:
- Search repositories and code
- Read files from any repo
- Create and manage issues
- List PRs and branches
- Get repository information
- Manage releases
Use Cases:
"Search GitHub for Python trading libraries"
"Get the README from anthropics/mcp-python"
"List issues in my TradingViewMCPServer repo"
"Create a new issue about X"
Configuration:
{
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-github"],
"env": {
"GITHUB_PERSONAL_ACCESS_TOKEN": "ghp_..."
}
}Token Permissions:
- ✅
repo- Full repository access - ✅
read:org- Read organization data - ✅
gist- Gist access
Purpose: Project structure tracking and file operations
What I Can Do:
- Navigate project directories
- Read file contents efficiently
- Track project structure
- Search files by patterns
- List directory contents
- Monitor file changes
Allowed Directories:
/Users/levtheswag/VSCodestuff(your main projects)/Users/levtheswag/Github Forking(forked repositories)
Use Cases:
"Show me the structure of my TradingViewMCPServer project"
"Find all Python files in my project"
"List files in the scrapy directory"
"Read all config files in my project"
Configuration:
{
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-filesystem",
"/Users/levtheswag/VSCodestuff",
"/Users/levtheswag/Github Forking"
]
}Benefits:
- ✅ Efficient directory traversal
- ✅ Fast file searches
- ✅ Structured project overview
- ✅ No need for manual
lscommands
Purpose: Dynamic and reflective problem-solving
What I Can Do:
- Break down complex problems step-by-step
- Think through architecture decisions
- Plan multi-step implementations
- Reason about edge cases
- Structured debugging approach
Use Cases:
"Help me think through how to implement X feature"
"What's the best architecture for Y?"
"Break down this complex problem"
"Debug this issue step by step"
Configuration:
{
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-sequential-thinking"]
}Benefits:
- ✅ Structured problem-solving
- ✅ Clear step-by-step reasoning
- ✅ Better architecture decisions
- ✅ Thorough debugging process
Purpose: Knowledge graph-based persistent memory
What I Can Do:
- Remember context across sessions
- Store project-specific knowledge
- Track decisions and rationale
- Maintain conversation history
- Build knowledge graphs
Use Cases:
"Remember that we're using Python 3.10+ for this project"
"What did we decide about the architecture?"
"Recall our discussion about Pine Script v6"
"Store this pattern for future reference"
Configuration:
{
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-memory"]
}Benefits:
- ✅ Context persistence
- ✅ No need to re-explain project structure
- ✅ Tracks decisions and patterns
- ✅ Builds project knowledge over time
Purpose: Advanced Git repository operations
What I Can Do:
- Read commit history
- Search through commits
- Analyze code changes
- Find when bugs were introduced
- Track file history
- Blame analysis
Configured Repository:
/Users/levtheswag/VSCodestuff/TradingViewMCPServer
Use Cases:
"Show me recent commits"
"When was this function changed?"
"Find the commit that introduced this bug"
"Show me who changed this file"
"Search commits for 'Pine Script'"
Configuration:
{
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-git",
"--repository",
"/Users/levtheswag/VSCodestuff/TradingViewMCPServer"
]
}Benefits:
- ✅ Deep git history analysis
- ✅ Code archaeology
- ✅ Bug tracking
- ✅ Understanding code evolution
Purpose: Local database operations and queries
What I Can Do:
- Query SQLite databases
- Create and manage tables
- Run SQL queries
- Analyze database structure
- Data migrations
Allowed Directory:
/Users/levtheswag/VSCodestuff(can access any .db files here)
Use Cases:
"Query the database at path/to/db.sqlite"
"Show me the schema of this database"
"Run this SQL query: SELECT * FROM..."
"Create a new table for X"
Configuration:
{
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-sqlite",
"/Users/levtheswag/VSCodestuff"
]
}Benefits:
- ✅ Local data analysis
- ✅ Database debugging
- ✅ Quick queries
- ✅ Schema inspection
After updating MCP configuration:
- Press
Cmd+Shift+P(macOS) orCtrl+Shift+P(Windows/Linux) - Type "Reload Window"
- Press Enter
Close and reopen VSCode completely.
After reload, test each server:
"What MCP servers do you have access to?"
Expected response should list:
mcp__fetch(web fetching)mcp__github(GitHub operations)mcp__filesystem(file operations)mcp__sequential_thinking(problem solving)mcp__memory(persistent memory)mcp__git(git operations)mcp__sqlite(database operations)
Fetch:
"Fetch content from https://example.com"
GitHub:
"List my GitHub repositories"
Filesystem:
"Show me the structure of my TradingViewMCPServer"
Sequential Thinking:
"Think through how to implement a new feature step by step"
Memory:
"Remember that this project uses Python 3.10+"
Git:
"Show me recent commits in TradingViewMCPServer"
SQLite:
"List databases in my VSCodestuff folder"
- ✅ Fetch Python documentation instantly
- ✅ Search GitHub for Python libraries
- ✅ Track project structure with Filesystem
- ✅ Remember Python patterns with Memory
- ✅ Query local databases with SQLite
- ✅ Fetch C# documentation
- ✅ Search GitHub for .NET libraries
- ✅ Track .NET project structure
- ✅ Remember C# patterns
- ✅ Sequential Thinking for complex problems
- ✅ Git for code history analysis
- ✅ GitHub for research and collaboration
- ✅ Memory for project-specific knowledge
- ✅ Filesystem for structure overview
- ✅ Git for commit tracking
- ✅ GitHub for issue management
- ✅ Memory for decisions and rationale
- Sequential Thinking: Plan the feature
- Filesystem: Check current structure
- Git: Review related commits
- Memory: Recall similar patterns
- Fetch: Research if needed
- GitHub: Check for existing solutions
- Sequential Thinking: Break down the problem
- Git: Find when bug was introduced
- Filesystem: Navigate to relevant files
- Memory: Recall similar bugs
- Fetch: Research error messages
- Fetch: Get official documentation
- GitHub: Search for examples
- Memory: Store findings
- Sequential Thinking: Evaluate options
To add more servers, edit settings.json and add to claude-code.mcpServers:
Brave Search (Web Search):
"brave-search": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-brave-search"],
"env": {
"BRAVE_API_KEY": "your_key_here"
}
}PostgreSQL (Database):
"postgres": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-postgres",
"postgresql://user:pass@localhost/db"
]
}Slack (Team Communication):
"slack": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-slack"],
"env": {
"SLACK_BOT_TOKEN": "xoxb-...",
"SLACK_TEAM_ID": "T..."
}
}Google Drive:
"gdrive": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-gdrive"]
}- ✅ Stored locally in settings.json
- ✅ Only accessible by your user account
⚠️ Don't commit settings.json to git⚠️ Disable Settings Sync for sensitive data
- ✅ Filesystem MCP restricted to specified directories
- ✅ SQLite MCP restricted to specified directories
- ✅ Git MCP restricted to specified repository
- Use minimal permissions for API tokens
- Restrict directories to only what's needed
- Review token scopes regularly
- Rotate tokens periodically
Problem: Server doesn't appear after reload
Solutions:
- Check JSON syntax in settings.json (no trailing commas!)
- Ensure
npxis installed (npm install -g npm) - Check VSCode Developer Console (Help → Toggle Developer Tools)
- Try restarting VSCode completely
Problem: "Access denied" or permission errors
Solutions:
- Check directory paths are correct
- Ensure you have read/write permissions
- For Git MCP, ensure repository path is correct
- For GitHub MCP, verify token has correct scopes
Problem: MCP tools are slow to respond
Solutions:
- Reduce number of directories in Filesystem MCP
- Use more specific Git repository paths
- Check network connection for Fetch/GitHub MCP
- Clear npm cache:
npm cache clean --force
Total MCP Servers: 7
✅ Fetch - Web content retrieval
✅ GitHub - Repository management (with token)
✅ Filesystem - Project structure (2 directories)
✅ Sequential - Problem-solving framework
✅ Memory - Persistent context
✅ Git - Repository analysis (TradingViewMCPServer)
✅ SQLite - Local database queries
Allowed Directories:
- /Users/levtheswag/VSCodestuff
- /Users/levtheswag/Github Forking
Git Repository:
- /Users/levtheswag/VSCodestuff/TradingViewMCPServer
- Reload VSCode Window (Cmd+Shift+P → "Reload Window")
- Test All Servers - Ask: "What MCP servers do you have?"
- Try Each Server - Test with the examples above
- Start Using! - Integrate into your workflow
- Official MCP Servers: https://github.com/modelcontextprotocol/servers
- MCP Documentation: https://modelcontextprotocol.io
- Claude Code Docs: https://docs.claude.com/claude-code
Configuration Date: 2025-01-XX Status: ✅ Ready to Use Action Required: Reload VSCode Window
Happy Coding! 🚀
With these MCP servers, I can now:
- 🌐 Research anything instantly
- 📁 Navigate your projects efficiently
- 🧠 Think through complex problems systematically
- 💾 Remember context across sessions
- 🔀 Analyze code history deeply
- 🗄️ Query databases on the fly
- 🐙 Manage GitHub seamlessly
You're all set for the ultimate development experience! 🎉