Agent BOM Monitor provides fleet-wide visibility into agents and AI infrastructure. It collects local scan state to compute trust scores, lists agents and health status, and can run a local dashboard for continuous monitoring. The tool is read-only by default and focuses on visibility and risk surfacing rather than automated remediation.
Use this skill when you need to assess the health and trustworthiness of your agent ecosystem, perform periodic fleet syncs, or run a local dashboard for audits and investigative work. It's useful for security engineers, platform teams, and operators maintaining multiple agents or services.
agent-bom fleet sync to refresh state, agent-bom fleet list to view trust scores, and agent-bom serve to open the dashboard locally.Designed to be used alongside OpenClaw / local operator agents and works on macOS, Linux, and Windows. It requires Python 3.11+ and is oriented toward security and ops workflows.
Agent BOM Monitor provides fleet monitoring, trust score tracking, and a local dashboard for AI agent infrastructure. No bundled scripts — relies entirely on the external `agent-bom` CLI tool installed via pipx. SKILL.md is well-structured with comprehensive frontmatter and clear guardrails, but the examples use pseudo-function syntax inconsistent with the bash commands.
agent-bom (pipx package, not installed)No scripts to execute. Skill is essentially a wrapper that instructs the agent to use an external CLI tool. Frontmatter metadata is thorough (credential_policy, data_flow, telemetry=false, privilege_escalation=false). Clean security posture — localhost-only dashboard, no data exfiltration, read-only fleet monitoring. Niche but legitimate use case for multi-agent ops teams.