xcode-build-analyzer_skill

Overview

This skill analyzes Xcode build logs and DerivedData to surface build timing, warnings, errors, slow compile steps, and historical trends. It reads the LogStoreManifest.plist and xcactivitylog files from ~/Library/Developer/Xcode/DerivedData to produce summaries and diagnostics. Designed for macOS developers who need fast, read-only inspection of build health and performance.

How this skill works

The analyzer parses LogStoreManifest.plist (no decompression required) to get structured build metadata: start/stop times, status, error and warning counts, and scheme info. For detailed output it decompresses .xcactivitylog files (gzip) and extracts TaskMetrics and compiler messages using simple filters (strings, grep, Python JSON parsing). All queries are read-only and operate on DerivedData without modifying files.

When to use it

• When you want a quick history of recent builds and status per scheme.
• To find top slow build tasks and investigate long-running compile steps.
• When you need an aggregated warning or error list from the latest build.
• To check disk usage of DerivedData and identify large project caches.
• Before cleaning DerivedData to verify what will be deleted (size/path).

Best practices

• Run the analyzer on macOS where Xcode has produced DerivedData; it won’t work elsewhere.
• Prefer LogStoreManifest.plist for fast history queries; only decompress xcactivitylog for details.
• Do not modify DerivedData; treat all operations as read-only, especially while Xcode runs.
• Confirm with users before suggesting rm -rf for cleanup; print size and path first.
• Use timestamps (Core Data epoch starts 2001-01-01) when presenting human-readable dates.

Example use cases

• Generate a concise timeline of build durations and spot regressions after a merge.
• List the latest build per project across DerivedData to monitor CI or local stability.
• Extract unique Swift warnings from the last build to prioritize fixes.
• Identify the top 10 slowest TaskMetrics entries to focus optimization efforts.
• Report per-project DerivedData sizes before deciding which caches to remove.