xcode-build-fixer

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a Python script (scripts/benchmark_builds.py) to invoke xcodebuild for building, cleaning, and extracting project settings. It also uses git ls-remote to verify Swift Package Manager tags. These commands are necessary for the skill's objective of measuring and improving build performance.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from .build-benchmark/optimization-plan.md to determine which fixes to apply. This represents an indirect prompt injection surface. The skill incorporates a safety protocol requiring the agent to only implement items with explicit developer approval via a checklist, which serves as a boundary marker. The capability inventory includes file system modifications and local command execution, but the process is gated by human review requirements.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 12:23 PM
Security Audit — agent-trust-hub — xcode-build-fixer