ldomaradzki/xcsift

A Swift command-line tool that parses xcodebuild output into token-efficient JSON for coding agents.

What it solves

xcsift addresses the problem of verbose, human-centric build logs from Xcode and Swift Package Manager (SPM) being inefficient for AI coding agents. Standard tools like xcbeautify focus on readability for humans, which often wastes valuable LLM context window space. xcsift converts these logs into compact, structured formats that are optimized for machine parsing and token efficiency.

How it works

It acts as a command-line filter that sits in your build pipeline. You pipe the output of xcodebuild or swift build into xcsift, which then parses the raw text and outputs structured data such as JSON or a custom, highly compressed format called TOON. It also functions as an MCP (Model Context Protocol) proxy, sitting between an AI agent and Xcode's native MCP server to inject structured build diagnostics directly into the agent's workflow.

Who it’s for

This tool is designed for developers building or using AI-powered coding assistants (such as Claude Code, Cursor, or Codex) who need to provide those agents with accurate, concise, and machine-readable build and test results.

Highlights

  • Token Efficiency: Offers a specialized TOON format that reduces token usage by 30-60% compared to standard logs.
  • Structured Diagnostics: Extracts errors, warnings, linker failures, and test results with precise file and line locations.
  • MCP Integration: Includes a proxy mode to seamlessly integrate with Xcode's native MCP server for real-time agent feedback.
  • CI-Ready: Supports GitHub Actions annotations, quiet modes, and configurable exit-on-failure behavior for automated pipelines.
  • Broad Coverage: Automatically converts build artifacts like .profraw and .xcresult into machine-readable JSON.

Related

  • Project
    github/CopilotForXcode

    An AI coding assistant for Xcode that provides intelligent code completions, chat, and an agentic mode for Swift and Objective-C development.

  • Project
    AvdLee/Xcode-Build-Optimization-Agent-Skill

    A collection of AI agent skills for benchmarking and optimizing Xcode build performance, providing an automated analyze-approve-fix workflow to reduce build times.

  • Project
    ml-explore/mlx-swift

    MLX Swift provides Swift bindings for the MLX tensor library, enabling high‑performance machine‑learning (including LLMs, vision‑language models and diffusion) on Apple‑silicon GPUs (Metal) and Linux (CPU/CUDA). It offers core tensor ops, neural‑network layers, optimizers and example apps, and can be added to Xcode or SwiftPM projects.

  • Project
    jgravelle/jcodemunch-mcp

    An MCP server that uses tree-sitter AST parsing to let AI agents retrieve precise code symbols instead of entire files, reducing token costs by up to 99%.

  • Project
    getsentry/MobileBuildMCP

    MobileBuildMCP is an open‑source tool that runs an MCP server and a CLI to let AI agents (or humans) invoke Xcode build and test commands for iOS/macOS projects. Install via Homebrew or npm, start the server with `mobilebuildmcp mcp`, and use subcommands like `simulator build` to compile or test apps. Optional “skills” JSON files teach LLM assistants how to talk to the server. MIT‑licensed, macOS‑only, with Sentry telemetry.