fleet-inventory

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill follows secure design principles by separating discovery from remediation and requiring pre-flight validation of the environment via the /mcp-lightspeed-validator skill.
  • [COMMAND_EXECUTION]: Operations are limited to specific MCP tools (get_host_details, get_cve_systems) provided by the lightspeed-mcp server, which prevents arbitrary command execution.
  • [CREDENTIALS_UNSAFE]: No hardcoded credentials detected. The skill utilizes environment variables for authentication and includes documentation in references/05-error-handling.md explicitly warning against echoing these secrets.
  • [PROMPT_INJECTION]: No evidence of instructions designed to bypass agent safety guardrails or override system prompts was detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 06:05 PM