mcpField
P
Periphery Code Audit

Periphery Code Audit

Local
github.comDeveloper Tools2Updated 49w ago

About Periphery Code Audit

Periphery Code Audit connects AI assistants to Periphery for auditing Swift code in iOS and macOS projects. It wraps the Periphery CLI and returns structured results for project scans, unused imports, and redundant public declarations.

Key features

  1. 01Seven tools for Periphery analysis
  2. 02Scan Xcode projects and Swift Packages
  3. 03Find unused imports and declarations
  4. 04Identify redundant public declarations
  5. 05Configure scans and output formats

Use cases

  1. 01Find unused code in an iOS project
  2. 02Audit a Swift Package for unused imports
  3. 03Review public declarations in a framework
  4. 04Run scans with project-specific options

MCP support

The MCP server exposes seven tools that invoke Periphery CLI for installation checks, version lookup, and configurable Swift project scans. Results are available to MCP clients as structured output.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

No install method listed yet

We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.

Repository
Packages
mcpbperiphery-mcp-server1.0.0

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

Periphery MCP Server

A Model Context Protocol (MCP) server that wraps the Periphery tool, enabling AI assistants to perform Swift code audits and detect unused code in iOS/macOS projects.

Features

  • 7 MCP Tools for comprehensive code analysis
  • CLI Integration - Uses Periphery CLI for stability and version flexibility
  • JSON Output - Structured, parseable results perfect for AI interpretation
  • Flexible Scanning - Support for Xcode projects and Swift Packages
  • Advanced Options - Full control over Periphery scan configurations

Quick Start

Using Homebrew (Recommended):

# 1. Install Periphery
brew install peripheryapp/periphery/periphery

# 2. Install Periphery MCP Server
brew tap zyntx-lab/tap
brew install periphery-mcp-server

# 3. Configure Claude Desktop
# Edit: ~/Library/Application Support/Claude/claude_desktop_config.json
# Add:
#   "periphery": {
#     "command": "/opt/homebrew/bin/periphery-mcp-server"
#   }

# 4. Restart Claude Desktop (Cmd+Q then reopen)

# 5. Test it
# In Claude Desktop: "Check if Periphery is installed"

From Source:

# 1. Install Periphery
brew install peripheryapp/periphery/periphery

# 2. Clone and build this server
git clone https://github.com/zyntx-lab/periphery-mcp-server.git
cd periphery-mcp-server
swift build -c release

# 3. Install to system path
sudo cp .build/release/periphery-mcp-server /usr/local/bin/
sudo chmod +x /usr/local/bin/periphery-mcp-server

# 4. Configure Claude Desktop
# Edit: ~/Library/Application Support/Claude/claude_desktop_config.json
# Add:
#   "periphery": {
#     "command": "/usr/local/bin/periphery-mcp-server"
#   }

# 5. Restart Claude Desktop (Cmd+Q then reopen)

# 6. Test it
# In Claude Desktop: "Check if Periphery is installed"

Prerequisites

  • macOS 13.0 or later
  • Swift 6.0 or later
  • Periphery installed
Installing Periphery
brew install peripheryapp/periphery/periphery

Or download from Periphery releases.

Installation

Option 1: Homebrew (Recommended)
# Add the tap
brew tap zyntx-lab/tap

# Install the server
brew install periphery-mcp-server

# Verify installation
which periphery-mcp-server
# Should output: /opt/homebrew/bin/periphery-mcp-server
Option 2: Download from GitHub Releases

Download the latest binary from Releases:

# Download and install
curl -L https://github.com/zyntx-lab/periphery-mcp-server/releases/download/v1.0.0/periphery-mcp-server \
  -o /usr/local/bin/periphery-mcp-server
chmod +x /usr/local/bin/periphery-mcp-server

# Verify installation
which periphery-mcp-server
# Should output: /usr/local/bin/periphery-mcp-server
Option 3: Build from Source
# Clone the repository
git clone https://github.com/zyntx-lab/periphery-mcp-server.git
cd periphery-mcp-server

# Build release version
swift build -c release

# Install to system path (requires password)
sudo cp .build/release/periphery-mcp-server /usr/local/bin/
sudo chmod +x /usr/local/bin/periphery-mcp-server

# Verify installation
which periphery-mcp-server
# Should output: /usr/local/bin/periphery-mcp-server
Option 4: Install from MCP Registry

The server is also available on the MCP Registry. Use your MCP-compatible client to install it directly.

Configuration

Claude Desktop
Step 1: Locate Your Config File

The config file is at: ~/Library/Application Support/Claude/claude_desktop_config.json

Step 2: Edit the Configuration

If you installed via Homebrew:

{
  "mcpServers": {
    "periphery": {
      "command": "/opt/homebrew/bin/periphery-mcp-server"
    }
  }
}

If you installed to /usr/local/bin:

{
  "mcpServers": {
    "periphery": {
      "command": "/usr/local/bin/periphery-mcp-server"
    }
  }
}

If you're using the build directory directly:

{
  "mcpServers": {
    "periphery": {
      "command": "/FULL/PATH/TO/periphery-mcp-server/.build/release/periphery-mcp-server"
    }
  }
}

If you already have other MCP servers configured:

{
  "mcpServers": {
    "xcode": {
      "command": "node",
      "args": ["/path/to/xcode-mcp-server/dist/index.js"]
    },
    "periphery": {
      "command": "/usr/local/bin/periphery-mcp-server"
    }
  }
}
Step 3: Restart Claude Desktop

Important: You must completely quit and restart Claude Desktop for changes to take effect.

  1. Quit Claude Desktop: Press Cmd+Q or use Claude Desktop → Quit
  2. Wait a few seconds
  3. Reopen Claude Desktop
Step 4: Verify It's Working

Open a new conversation in Claude Desktop and try:

"Check if Periphery is installed"

If configured correctly, Claude will use the check_periphery_installed tool and respond with installation status.

Testing with MCP Inspector

Before configuring Claude Desktop, you can test the server with the MCP Inspector:

# Install MCP Inspector (if not already installed)
npm install -g @modelcontextprotocol/inspector

# Test your server
npx @modelcontextprotocol/inspector /usr/local/bin/periphery-mcp-server

# Or if using build directory
npx @modelcontextprotocol/inspector /path/to/periphery-mcp-server/.build/release/periphery-mcp-server

The Inspector will open in your browser where you can test all tools interactively.

Other Editors
VS Code

Add to .vscode/mcp.json:

{
  "mcpServers": {
    "periphery": {
      "type": "stdio",
      "command": "/usr/local/bin/periphery-mcp-server"
    }
  }
}
Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "periphery": {
      "type": "stdio",
      "command": "/usr/local/bin/periphery-mcp-server"
    }
  }
}

Available Tools

1. check_periphery_installed

Verify that Periphery CLI is installed and accessible.

Parameters: None

Example Response:

{
  "installed": true,
  "path": "/usr/local/bin/periphery",
  "message": "Periphery is installed and ready"
}
2. get_periphery_version

Get the installed version of Periphery.

Parameters: None

Example Response:

{
  "version": "2.18.0",
  "raw_output": "2.18.0"
}
3. scan_project

Run a basic Periphery scan on a project.

Parameters:

  • project_path (required): Path to .xcodeproj or Package.swift
  • schemes (optional): Build schemes to scan (Xcode only)
  • targets (optional): Specific targets to analyze
  • format (optional): Output format: json, xcode, csv, checkstyle (default: json)

Example Response:

{
  "success": true,
  "results": [
    {
      "kind": "class",
      "name": "UnusedClass",
      "modifiers": ["public"],
      "location": "Sources/MyApp/UnusedClass.swift:10:7"
    }
  ],
  "summary": {
    "total_unused": 5,
    "by_kind": {"class": 2, "function": 3}
  }
}
4. scan_with_config

Run Periphery scan using a YAML configuration file.

Parameters:

  • config_path (required): Path to .periphery.yml config file
5. analyze_unused_imports

Focus specifically on detecting unused imports.

Parameters:

  • project_path (required): Path to .xcodeproj or Package.swift
  • schemes (optional): Build schemes to scan
  • targets (optional): Specific targets to analyze
6. find_redundant_public

Identify public declarations that could be internal.

Parameters:

  • project_path (required): Path to .xcodeproj or Package.swift
  • schemes (optional): Build schemes to scan
  • targets (optional): Specific targets to analyze
7. scan_with_options

Advanced scanning with custom Periphery flags.

Parameters:

  • project_path (required): Path to .xcodeproj or Package.swift
  • schemes (optional): Build schemes to scan
  • targets (optional): Specific targets to analyze
  • format (optional): Output format
  • retain_public (optional): Retain all public declarations
  • retain_objc_accessible (optional): Retain @objc declarations
  • disable_unused_import_analysis (optional): Disable unused import analysis
  • index_store_path (optional): Custom index store location
  • verbose (optional): Enable verbose output

Usage Examples

With Claude Desktop

Once configured, you can have natural conversations with Claude about your code:

Check Installation:

You: "Check if Periphery is installed"

Claude: [Uses check_periphery_installed]
"Yes, Periphery is installed at /usr/local/bin/periphery
Version: 2.18.0"

Scan a Project:

You: "Scan my iOS project at ~/Projects/MyApp/MyApp.xcodeproj for unused code"

Claude: [Uses scan_project]
"I found 15 unused declarations in your project:

Classes (5):
- UnusedViewController at MyApp/UnusedViewController.swift:10
- OldDataManager at MyApp/Models/OldDataManager.swift:25
...

Functions (8):
- helperFunction at Utils/Helpers.swift:42
...

Imports (2):
- UIKit in DataModel.swift:1
..."

Find Redundant Public:

You: "Check which public declarations could be made internal in ~/Projects/MyFramework"

Claude: [Uses find_redundant_public]
"Found 12 public declarations that are only used internally and could be made internal:
- public class InternalHelper (only used within the framework)
- public func formatDate() (only called from within the module)
..."

Analyze Unused Imports:

You: "Find unused imports in my Swift package"

Claude: [Uses analyze_unused_imports]
"Found 8 unused imports that can be safely removed:
- Foundation in Models/User.swift (not using any Foundation APIs)
- Combine in ViewModels/ProfileViewModel.swift (Combine is imported but not used)
..."
With Custom Configuration

Create .periphery.yml in your project:

project: MyApp.xcodeproj
schemes:
  - MyApp
targets:
  - MyApp
  - MyAppKit
format: json
retain_public: false
retain_objc_accessible: true
verbose: false

Then use scan_with_config:

You: "Scan my project using the custom config"

Troubleshooting

Server Not Appearing in Claude Desktop

Symptoms: Claude doesn't recognize Periphery tools after configuration

Solutions:

  1. Verify config file syntax:

    # Check for JSON syntax errors
    cat ~/Library/Application\ Support/Claude/claude_desktop_config.json | python3 -m json.tool
    
  2. Check the binary path is correct:

    # Test that the binary exists and runs
    /usr/local/bin/periphery-mcp-server
    # Or your custom path
    /path/to/.build/release/periphery-mcp-server
    
  3. Completely restart Claude Desktop:

    • Press Cmd+Q to quit (not just close the window)
    • Wait 5 seconds
    • Reopen Claude Desktop
    • Start a new conversation (old conversations won't see new servers)
  4. Check Claude Desktop logs:

    # View logs for errors
    tail -f ~/Library/Logs/Claude/mcp*.log
    
Periphery Not Found

If you get "Periphery is not installed":

  1. Install Periphery: brew install peripheryapp/periphery/periphery
  2. Verify installation: which periphery
  3. Restart Claude Desktop/VS Code/Cursor
  4. Try the check_periphery_installed tool again
MCP Inspector Connection Errors

If the Inspector can't connect to the server:

  1. Verify the server runs:

    # Server should wait for input, not exit immediately
    /usr/local/bin/periphery-mcp-server
    # Press Ctrl+C to exit
    
  2. Check for errors:

    # Run with verbose output
    /usr/local/bin/periphery-mcp-server 2>&1 | tee server.log
    
  3. Use the absolute path:

    npx @modelcontextprotocol/inspector $(which periphery-mcp-server)
    
Scan Timeout

Default timeout is 5 minutes. For large projects:

  1. Use scan_with_config with a focused configuration
  2. Scan specific targets instead of the entire project
  3. Use --index-store-path to reuse build artifacts
  4. Build your project in Xcode first to generate the index
No Results or Empty Scan

Periphery requires a compiled project with an index store:

  1. Build your project first in Xcode (Cmd+B)
  2. Ensure schemes are shared:
    • Xcode → Product → Scheme → Manage Schemes
    • Check the "Shared" checkbox for your scheme
  3. Verify the project path:
    # For Xcode projects
    ls /path/to/YourProject.xcodeproj
    
    # For Swift Packages
    ls /path/to/Package.swift
    
  4. Check Periphery can access the project:
    # Test Periphery directly
    periphery scan --project /path/to/YourProject.xcodeproj --schemes YourScheme
    
Permission Denied

If you get "Permission denied" when running the server:

# Make the binary executable
chmod +x /usr/local/bin/periphery-mcp-server

# Or for build directory
chmod +x .build/release/periphery-mcp-server

Architecture

This server uses the CLI approach rather than importing Periphery as a library for maximum stability:

  • ✅ CLI interface is Periphery's public API contract
  • ✅ Survives internal Periphery refactorings
  • ✅ Users can update Periphery independently
  • ✅ Simpler dependency management

Contributing

Contributions are welcome! Please:

  1. Fork the repository
  2. Create a feature branch
  3. Add tests for new functionality
  4. Submit a pull request

License

MIT License - see LICENSE file for details.

Related Projects

Distribution Channels

This server is available through multiple channels:

Support

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.