claude-monitor
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's diagnostic scripts (api_bench.py, health_check.py, and monitor.py) utilize the subprocess module to automatically install the psutil library from the standard Python Package Index (PyPI) if it is not already available in the execution environment.- [COMMAND_EXECUTION]: The skill programmatically monitors system performance by executing psutil functions to collect metrics on CPU usage, memory allocation, and disk space. It also inspects active process lists to identify resource-heavy browser tabs and Claude Code instances to assist in troubleshooting lag.- [SAFE]: The skill implements proactive security measures, including a safe_user_path helper function in the monitoring script that validates output paths to prevent path traversal vulnerabilities when saving log files.- [SAFE]: Network activity is limited to benign connectivity and latency tests targeting official well-known services (api.anthropic.com and cdn.anthropic.com) and public DNS (8.8.8.8) to differentiate between local hardware bottlenecks and remote API latency.
Audit Metadata