auto-monitor-hardened

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: Potential for indirect prompt injection as the skill processes untrusted system metadata.
  • Ingestion points: Reads system status metrics, process names, and mount points from the host environment.
  • Boundary markers: Absent for input data.
  • Capability inventory: Potentially executes file deletions (rm), service restarts (systemctl), and process termination (kill) for remediation.
  • Sanitization: Explicitly instructs the agent to use safe command execution patterns (e.g., -exec rm {} +) and avoid shell interpretation of monitoring data.
  • [DATA_EXFILTRATION]: The skill handles sensitive system telemetry including IP addresses and process lists. It contains a specific guardrail that prohibits transmitting this data to external services or third-party endpoints.
  • [COMMAND_EXECUTION]: Performs automated health checks and remediation via shell commands. It includes hardening instructions to prevent construction of dynamic commands from unquoted monitoring output, mitigating shell-injection risks.
  • [EXTERNAL_DOWNLOADS]: References external safety reports and documentation hosted on the Faberlens platform; these are documented neutrally as author-provided resources.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 06:30 PM
Security Audit — agent-trust-hub — auto-monitor-hardened