Loading story…
Google Cloud has open-sourced k8s-aibom, a Kubernetes controller built to detect and document AI workloads running in cloud environments, including inference, training, agentic systems, evaluations, and retrieval-augmented generation deployments. The tool is designed to surface shadow AI activity and generate namespace-scoped AI bills of materials for auditors and security teams without modifying workloads, using a single unprivileged action in its own namespace rather than sidecars, eBPF, or root-level DaemonSet agents.
The controller outputs artifacts aligned with the CycloneDX ML-BOM standard, part of a broader software supply chain transparency ecosystem that also includes formats such as SBOM, VEX, and related attestations. Google said the project supports AI governance, compliance, and change detection efforts tied to frameworks including the EU AI Act, NIST AI RMF, and ISO/IEC 42001, extending bill-of-materials practices into machine learning environments.
Track how attackers are adapting to this technology.
2 events from the most recent confirmed update back to the earliest known activity.
Google Cloud open-sourced k8s-aibom, a Kubernetes controller that identifies AI workloads and generates namespace-scoped AI bills of materials using the CycloneDX ML-BOM standard. The tool is positioned for AI transparency, compliance, and change detection in cloud environments.
CycloneDX published documentation describing Machine Learning Bill of Materials (ML-BOM) as one of several software supply chain transparency artifact types and related standards. The reference is descriptive and does not report a security incident or exploitation event.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
2 references tracked. Mallory keeps watching after this page renders.
Map indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
CISA and partner agencies in the United States and 13 allied countries released updated guidance for the minimum elements of a Software Bill of Materials, replacing the 2021 NTIA baseline with broader requirements intended to improve software transparency and supply chain risk management. The revised framework preserves the original core principles but adds fields for component hash information, license data, author or digital signature, SBOM version, data format details, generation context, and tool metadata, while removing elements such as Access Control and SWID Tags. Officials said the changes were shaped by software supply chain security developments and feedback gathered during a 2025 public comment period. The update is designed to make SBOMs more useful for automated vulnerability management and procurement decisions, and it explicitly notes that some software categories need additional treatment beyond the general baseline. In parallel, the G7 Cybersecurity Working Group published guidance on the minimum elements for an AI-focused SBOM, providing actionable recommendations for public- and private-sector stakeholders to improve transparency across the AI supply chain. The combined guidance signals a broader international push to standardize software component disclosure for both conventional software and AI systems, with future work expected on SaaS, AI-specific components, SBOM trustworthiness, and links to formats such as **VEX** and **CSAF**.
Jul 31, 2026
A growing divide between open-weight and proprietary AI models is shaping industry competition, cybersecurity policy, and model-governance decisions. Supporters including Microsoft, Amazon, Nvidia, Google, and startups argue that releasing model weights expands access, lowers development costs, accelerates innovation, and enables independent inspection and security testing. Moonshot AI's Chinese open-weight `Kimi K3` has become a focal point after U.S. officials alleged improper model distillation, while others contend that aspects of the practice may be legitimate. Open weights do not constitute fully open-source AI: providers may withhold training code, datasets, intermediate checkpoints, and data provenance. The Open Source Initiative warns those omissions limit reproducibility, bias auditing, regulatory assurance, and meaningful modification of a model, particularly for high-stakes deployments. Anthropic and OpenAI favor stronger release controls and safety evaluations, citing risks that capable models could be exploited by authoritarian governments or replicated through industrial-scale distillation; CISOs should therefore assess both a model's independently verifiable artifacts and its abuse safeguards rather than treating “open weights” as a security or transparency guarantee.
Aug 27, 2026
Broadcom has launched **VMware Private AI Cloud**, combining VMware Cloud Foundation 9 (VCF 9) with the VMware AI Factory software layer to operate enterprise AI workloads in private-cloud environments. The platform targets AI infrastructure capital costs, operational complexity, and token-related spending through storage optimization, token monitoring, and GPU/vGPU usage tracking, while providing access to more than 150 open-source and commercial AI models. The service extends VCF Private AI governance with an **AI Gateway** for centralized model access, prompt routing, and policy controls across on-premises and cloud deployments. Broadcom also introduced agent-security capabilities including AgentMinder for authentication, authorization, monitoring, and audit of autonomous-agent actions, plus sandboxing and runtime restrictions intended to limit agent permissions, network communications, and resource consumption; TrueSource supplies curated, hardened open-source artifacts for AI deployments.
Sep 2, 2026