mac-analyze

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple macOS shell commands to perform diagnostics.
  • Hardware detection: sw_vers, uname -m, sysctl, and system_profiler.
  • Memory analysis: vm_stat, sysctl vm.swapusage, memory_pressure, and ps (with awk for aggregation).
  • Disk and file analysis: df, du -sh, and find (to identify large or old files in the Downloads folder).
  • System state: pmset (power/thermal), lsof/netstat (via listening ports check), and Homebrew/Docker CLI tools.
  • [DATA_EXPOSURE]: The skill accesses system metadata and configuration file paths for diagnostic purposes.
  • Analyzes directory sizes for node_modules, Xcode artifacts, and browser caches.
  • Lists LaunchAgents and LaunchDaemons to identify startup items.
  • Scans the Downloads directory for file metadata (size and age).
  • This behavior is consistent with the stated diagnostic purpose and does not involve exfiltrating data to external servers.
  • [INDIRECT_PROMPT_INJECTION]: The skill uses external files (resources/diagnostic-commands.md and templates/diagnostic-report.md) to drive its logic.
  • Ingestion points: Reads local markdown files for command references and reporting templates.
  • Boundary markers: None present in the instructions to delimit the content of these files.
  • Capability inventory: Executes shell commands and file system reads based on the loaded instructions.
  • Sanitization: None specified for the content of the referenced diagnostic files.
  • This represents a low-risk attack surface common to modular agent skills.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 02:40 PM