Skip to content

All newsletters

Newsletter

Nirmata Pulse: August 2026

Governance Beyond the Admission Webhook

News from Nirmata, August 2026

All new newsletter image

For years, Kubernetes policy came down to one question: Does this workload specification look right?

August pushed that model further.

From kernel-level enforcement and automated remediation to AI-agent authorization and deterministic AI workflows, the cloud-native ecosystem is expanding where and how policy can be applied.

Here’s what happend last month.

AI Passed the Gate. Runtime Governance Starts Here.

Admission control decides whether a workload can enter the cluster. But once it starts running, what it actually does is another story.

Nirmata Runtime for Kyverno extends policy enforcement to the kernel using BPF-LSM and eBPF, giving teams control over runtime behaviors like process execution, file access, network activity, and DNS resolution.

For AI workloads and autonomous agents, this matters: governing what an image can do isn’t enough you need to govern what it actually does.

Read the technical breakdown

THIS MONTH AT NIRMATA

🚀 Open Source

OttoFlow: AI Workflows for Kubernetes

Nirmata open-sourced OttoFlow, a declarative and deterministic workflow engine designed to run AI workflows as Kubernetes resources.

It brings production-oriented workflow automation to Kubernetes while keeping execution reviewable, predictable, and GitOps-friendly

Blog Post

🤝 Partner Ecosystem

N4K Certified on Nutanix Kubernetes Platform

Nirmata Enterprise for Kyverno (N4K) is now certified to run on Nutanix Kubernetes Platform (NKP) and is available through the NKP Partner Catalog.

This enables governance to be provisioned alongside Kubernetes infrastructure rather than added as a separate layer afterward.

Blog Post

🔐 AI Governance

Standards-Based Authorization for AI Agents

Nirmata AIControls now supports Okta Cross App Access (XAA / ID-JAG), enabling standards-based authorization for AI agents.

AIControls handles the ID-JAG exchange on behalf of agents while applying policy, budget, and attribute controls to delegated MCP and LLM calls.

Blog Post

⚙️ Automation

Turning Kyverno Violations Into Pull Requests

Finding policy violations is only the first step.

The new Remediator Agent automatically converts Kyverno policy violations into pull requests, helping development teams move from detection to remediation without manually fixing every finding.

Blog Post

NIRMATA IN THE COMMUNITY

From KubeCon + CloudNativeCon India

Two Nirmata/Kyverno sessions from KubeCon + CloudNativeCon India 2026 are now available through the CNCF ecosystem:

Automating RBI Compliance With Unified Policy-as-Code

Jim Bugwadia and Ashwath Kumar discuss using unified policy-as-code to automate regulatory compliance in financial services.

[Watch the keynote]

The Kyverno Five! Policy Types for the Full Resource Lifecycle

Neha Jaju explores Kyverno’s five policy types and how they can be applied across the Kubernetes resource lifecycle.

[Watch the session]

INDUSTRY WATCH

CNCF Graduates Kubeflow

On August 17, CNCF announced the graduation of Kubeflow, recognizing its maturity as a production platform for AI and machine-learning workloads on Kubernetes.

The project has nearly 260 million PyPI downloads and more than 6,600 contributors across 1,000+ organizations. Its roadmap also includes LLM orchestration and agentic workloads.

[Read the announcement]

KubeCon North America Adds an AI Inference + Agentic Track

CNCF’s upcoming KubeCon + CloudNativeCon North America 2026 introduces a dedicated AI Inference + Agentic track covering production AI, agentic workflows, GPU scheduling, model serving, and observability.

The conference takes place November 9–12 in Salt Lake City.

[Explore the schedule]

FROM THE KYVERNO ECOSYSTEM

Kyverno Is a Platform Primitive, Not a Security Tool

A recent CNCF Ambassador post explores a broader way to think about Kyverno: not simply as a security gate, but as a platform primitive for turning organizational intent into automated platform behavior.

The perspective is especially relevant as platform teams increasingly use policy-as-code for security, operational consistency, self-service, and governance.

CNCF Post

WHAT’S NEXT

KubeCon + CloudNativeCon North America

November 9-12, 2026 · Salt Lake City

With AI inference, agentic systems, platform engineering, and cloud-native security all receiving dedicated attention, the next KubeCon promises to continue the conversation around how Kubernetes infrastructure must evolve for production AI.

CNCF Announcement

Stay Ahead of Kubernetes Governance

The latest thinking around Kyverno, AI governance, runtime security, and policy-as-code.

Explore our Projects on GitHub to see the latest developments and contribute [GitHub].

See Nirmata in Action