claude-monitor

Pass

Audited by Gen Agent Trust Hub on Apr 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The scripts api_bench.py, health_check.py, and monitor.py verify the presence of the psutil library and automatically install it from the official Python Package Index (PyPI) using pip if it is missing. This is a standard practice for utility scripts to manage dependencies for portable diagnostics.\n- [COMMAND_EXECUTION]: The skill uses subprocess.check_call to execute the system's Python interpreter for library installation and resource monitoring tasks. These operations are restricted to the local environment and the intended diagnostic functionality.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted system data (e.g., process names) using psutil.process_iter in health_check.py and monitor.py. While the data is passed to the agent to generate reports, the impact is minimal as the skill performs no sensitive external network operations or file-system changes based on this data.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 28, 2026, 09:56 AM