openclaw-health-check

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes common system utilities (such as df, du, find, grep, and awk) and local service managers (pm2, openclaw) to gather operational metrics and service status information from the local environment.
  • [REMOTE_CODE_EXECUTION]: Although automated scans flagged a pipe to python3, the implementation uses a static command-line string to parse JSON data from a local health endpoint (http://localhost:18789/health). This approach is used for data processing rather than executing remote script content.
  • [SAFE]: The skill accesses local session log files in ~/.openclaw/ to calculate cost estimates and activity levels. This data ingestion is focused on generating the diagnostic report and does not involve exfiltration to external domains.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 02:29 PM
Security Audit — agent-trust-hub — openclaw-health-check