What is an AI bill of materials (AIBOM)?
An AI bill of materials (AIBOM) is a machine-readable inventory of the components of an AI system: models, datasets, prompts, tools, agent frameworks, MCP servers and the software they depend on, with versions and origins. It extends the software bill of materials so that AI components can be tracked and checked.
Also called AI-BOM, ML-BOM, AI SBOM. Last reviewed .
Why it matters
Two open formats carry it. CycloneDX defines a Machine Learning Bill of Materials for "model and dataset transparency for security, privacy, safety and ethical considerations", and SPDX defines an AI profile.
The use is the same as for any bill of materials: when a component turns out to be bad, you can tell whether you run it. When Postmark warned about the impersonating postmark-mcp package in September 2025, the question for every team was whether that package, at version 1.0.16 or later, was installed anywhere.
Related terms
- AI agent inventory
- An AI agent inventory lists the systems; an AIBOM lists what is inside each one.
- MCP security
- MCP security depends on knowing which MCP server packages and versions are in use.
- Tool poisoning
- Tool poisoning is easier to catch when the approved version of each tool is on file.
- AI governance
- AI governance uses the bill of materials as evidence of what a system contains.
How Guard0 handles it
The free Guard0 AI Register scans the AI, agent, MCP and model endpoints a domain exposes and exports the result as a CycloneDX 1.6 AI-BOM: model endpoints as components, agent and MCP endpoints as services, and findings as vulnerabilities.
Sources
- CycloneDX Machine Learning Bill of Materials (ML-BOM). OWASP CycloneDX.
- SPDX: AI. SPDX, The Linux Foundation.
- Information regarding malicious postmark-mcp package. Postmark.