菜单
Set up and use XcodeBuildMCP to compile and test Xcode projects after Swift edits. Use after any edit to a `.xcodeproj` or `Package.swift`. Triggers on `xcodebuild`, simulator build, build verification, agent-driven iOS development.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | xcodebuildmcp-build-loop |
| description | Set up and use XcodeBuildMCP to compile and test Xcode projects after Swift edits. Use after any edit to a `.xcodeproj` or `Package.swift`. Triggers on `xcodebuild`, simulator build, build verification, agent-driven iOS development. |
The build is the test. After any non-trivial Swift edit, compile the project end to end on a real simulator (or macOS scheme) through XcodeBuildMCP. Fix failures before continuing. Type checker passing is not enough.
Any task that edits Swift, SwiftUI, Package.swift, asset catalogs, or Xcode project settings in a repo built with xcodebuild (.xcodeproj or .xcworkspace). Skip only for pure docs or non-Swift changes.
The Swift type checker passes long before the project actually links. SwiftUI semantics, @Observable registration, Sendable and MainActor isolation, package dependency resolution, and platform availability gates only fail at the full build or at runtime. A green editor with a red xcodebuild is a regression you will ship.
npm install -g xcodebuildmcp (or use npx xcodebuildmcp from the MCP config).~/.claude/settings.json or .mcp.json). Use the package's documented command and args (typically npx -y xcodebuildmcp).mcp__XcodeBuildMCP__* appear. If not, the MCP did not start; see references/setup-and-commands.md.mcp__XcodeBuildMCP__build_sim_name_proj for iOS / iPadOS / visionOS simulators, or build_mac_proj for macOS. Pass projectPath, scheme, simulatorName (for sim builds).mcp__XcodeBuildMCP__test_sim against the relevant per-package test scheme.Once per session, set defaults so subsequent calls stay short:
mcp__XcodeBuildMCP__session-set-defaults
projectPath: "<ABSOLUTE_PATH_TO_XCODEPROJ>"
simulatorName: "iPhone 16"
Then later calls can omit projectPath and simulatorName.
If you do not know the scheme name, list schemes first with mcp__XcodeBuildMCP__list_schems_proj.
Package.swift.mcp__XcodeBuildMCP__* tools do not appear in the session. Cause: the MCP server did not start. Fix: check the registration in ~/.claude/settings.json (or .mcp.json), run npx -y xcodebuildmcp --help in a shell to verify the binary resolves, then restart Claude Code.No such module 'X'. Cause: the target imports X but does not depend on the package that vends it. Fix: add the missing .product(name: "X", package: "X") to target.dependencies in Package.swift.xcrun simctl list devices available, copy a name verbatim (case, spacing), update session-set-defaults.clean_proj. Cause: stale DerivedData (often after a branch switch or deployment-target bump). Fix: call clean_proj first; if persistent, delete ~/Library/Developer/Xcode/DerivedData/<ProjectHash>.Eight more failure modes with fix recipes live in references/common-errors.md.
references/setup-and-commands.md: concrete tool signatures, scheme discovery, multi-package projects, MCP connectivity checks.references/common-errors.md: one-screen remediation recipes for the recurring xcodebuild failure modes.Use when building or improving HTML/CSS artifacts, React pages, dashboards, landing pages, explainers, demos, SVG diagrams, UI prototypes, or visual outputs where design quality, subject-specific treatment, responsive behavior, and render verification matter.
General Python development skill for designing, implementing, refactoring, debugging, testing, validating, and reviewing Python code. Use for Python 3.11+ projects, especially modern pyproject-based repos using uv, Ruff, pytest, mypy, Pydantic v2, pydantic-settings, Typer, and Loguru. Applies the house Python stack uniformly while discovering local contracts.
Use when the user asks to use a kanban process, track development work on a GitHub board, work from the agent board, proactively track tasks and deferred findings, runs `/kanban-process`, or when the project is already configured to use this workflow.
Use at the start of every Fable session. Fable is the brain: decompose, dispatch, judge, and talk to the user. Codex CLI does the work: exploration, bulk reads, tests, builds, implementation, reviews, and other token-heavy units.
Use when spawning Codex subagents or Codex threads and choosing among GPT 5.6 Luna, Terra, and Sol with the appropriate effort. Use this with orchestrating-subagents whenever delegation involves model overrides, reasoning effort, Explore agents, checker/fixer loops, or fan-out work.
Use when an AI agent is itself performing a pull request review (not coaching a human, not fetching the PR). Covers agent-specific tactics: context budget, tool escalation triggers, Explore subagent dispatch (single sidecar and parallel fan-out), when to emit "I cannot review this responsibly", auto-memory integration, and agent failure modes (confabulation, tool-call theater, delegated certainty laundering, evidence laundering, patch sycophancy, premature fan-out). Complements reviewing-pull-requests (universal doctrine) and pr-review (operational PR fetch + per-layer comparison).