Skip to content
Cisco AI Defense logo
CiscoAI Security

AIBOM — DefenseClaw

Overview

defenseclaw aibom scan queries the live OpenClaw environment and builds a unified inventory of skills, plugins, MCP servers, agents, tools, models, and memory. The command enriches the inventory with policy state, converts it to a scan result for the audit database, and can print either human output or JSON.

What it collects

ComponentSource
SkillsOpenClaw skill inventory
PluginsOpenClaw plugin inventory
MCP serversOpenClaw MCP configuration
AgentsOpenClaw agent inventory
ToolsTool inventory exposed by OpenClaw
ModelsModel inventory exposed by OpenClaw
MemoryMemory inventory exposed by OpenClaw

The implementation lives in defenseclaw.inventory.claw_inventory and is invoked from cli/defenseclaw/commands/cmd_aibom.py::scan.

CLI

defenseclaw aibom scan
defenseclaw aibom scan --summary
defenseclaw aibom scan --json
defenseclaw aibom scan --only skills,plugins,mcp

See aibom CLI.

Output modes

OptionBehavior
defaultPrints the formatted inventory.
--summaryShows the summary table only.
--jsonPrints the full inventory JSON and stamps provenance metadata.
--onlyRestricts collection to a comma-separated category list.

Audit behavior

After building the inventory, the command calls claw_aibom_to_scan_result and logs that scan result when an audit logger is configured. That makes AIBOM part of the same scan history as skill, MCP, plugin, and CodeGuard results.

Use cases

  • Incident response. "Is model X installed anywhere?" Use JSON mode and filter the live inventory.
  • Readiness checks. Confirm skills, plugins, MCP servers, and tools are visible before enabling watcher enforcement.
  • Audit trail. Store scan rows from each AIBOM run in the audit database.
defenseclaw aibom scan --json | jq '.models'

Related