swift-skeleton

Overview

swift-skeleton is an MCP server that extracts type declarations, properties, method signatures, and source locations from Swift and other programming languages. It creates a lightweight structural overview of codebases that fits in context windows, allowing AI assistants to understand code architecture without reading entire source files. Built on Tree-sitter for fast parsing, it supports multiple languages including Swift, Kotlin, TypeScript, Go, Zig, Rust, C++, Python, and Java.

Try asking AI

After installing, here are 7 things you can ask your AI assistant:

• skeleton

  Extract structural skeleton of Swift, Kotlin, TypeScript, Go, Zig, Rust, C++, Python, or Java projects

  Try:Get the skeleton for my Swift projectTry:Show me the structure of my TypeScript codebaseTry:Extract type declarations and method signatures from my project
• query

  Search for specific symbols or declarations in the project skeleton

  Try:Find all classes containing 'NetworkManager'Try:Search for function declarations named 'processData'Try:Look for all structs in the codebase
• status

  Get project status and skeleton index information

  Try:Check the status of the skeleton indexTry:Show project information
• diagnostics

  Get diagnostic information about the project skeleton

  Try:Show parsing diagnostics for the projectTry:Check for any skeleton generation issues
• files

  List files included in the skeleton index

  Try:Show which files are indexed in the skeletonTry:List all Swift files in the project
• languages

  List supported programming languages

  Try:Which languages does swift-skeleton support?Try:List all available language parsers
• index.open

  Open a project in the JSON-RPC daemon

  Try:Open my Swift project in the skeleton indexTry:Start indexing the project at /path/to/project
• index.get_skeleton

  Get skeleton data via JSON-RPC daemon

  Try:Get the skeleton for the current projectTry:Retrieve structural information via the daemon
• index.query

  Query the project skeleton via JSON-RPC daemon

  Try:Search for all 'User' related symbolsTry:Find functions with 'authenticate' in their name
• index.update

  Update the project skeleton via JSON-RPC daemon

  Try:Update the skeleton index with new filesTry:Refresh the project skeleton
• index.diagnostics

  Get skeleton diagnostics via JSON-RPC daemon

  Try:Check for any issues in the skeleton indexTry:Show parsing diagnostics via daemon
• index.status

  Get index status via JSON-RPC daemon

  Try:Check the current status of the skeleton indexTry:Get information about the indexed project

Comparable tools

Installation

mint install 1amageek/swift-skeleton

git clone https://github.com/1amageek/swift-skeleton.git
cd swift-skeleton
swift build -c release

{
  "mcpServers": {
    "swift-skeleton": {
      "command": "path/to/skltn",
      "args": ["daemon"]
    }
  }
}

skltn install-skill

FAQ

Which programming languages are supported by swift-skeleton?

swift-skeleton supports Swift, Kotlin, TypeScript, Go, Zig, Rust, C++, Python, and Java. Each language parser can be selected via package traits.

How does swift-skeleton help AI coding assistants?

It provides a structural overview of codebases (type declarations, methods, relationships) that fits in context windows, allowing AI assistants to understand architecture before exploring specific implementations.

Can I use swift-skeleton as a library in my own project?

Yes. The project offers both embedded (in-process) and sidecar (out-of-process) usage patterns, along with a JSON-RPC daemon for MCP integration.

Compare swift-skeleton with