system-health

Pass

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill leverages standard UNIX utilities such as ps, top, df, free, and uptime to monitor system resources. While it includes commands to terminate processes (kill, pkill), these are standard tools for system maintenance and are appropriately documented for their intended use.
  • [EXTERNAL_DOWNLOADS]: No external downloads, package installations, or remote script executions were detected. The homepage reference points to authoritative Linux manual pages.
  • [CREDENTIALS_UNSAFE]: The skill does not access sensitive credential files (like .ssh, .aws, or .env) or contain hardcoded secrets. It focuses exclusively on public system telemetry and resource usage.
  • [PROMPT_INJECTION]: No instructions were found that attempt to override agent safety guidelines, bypass system constraints, or extract internal prompts.
  • [DATA_EXFILTRATION]: No network operations (such as curl or wget) targeting external domains were identified. All commands operate locally on the system for monitoring purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 31, 2026, 09:05 AM
Security Audit — agent-trust-hub — system-health