Ritesh Patel | Nirmata
Ritesh Patel
Ritesh Patel is a founder of Nirmata and leads its products. He writes about governing AI coding agents, MCP security and Kubernetes policy as code.
RBAC for AI agents in Kubernetes
Give each AI agent its own service account, bind it to a narrow Role, and point its pod at that account. RBAC has no deny rules, so grants are the limit.
November 5, 2026 · 5 min read
How to sandbox AI agents on Kubernetes
Run each AI agent in its own Pod with a hardened security context, default-deny egress and a RuntimeClass, with lifecycle managed by Agent Sandbox CRD.
November 5, 2026 · 6 min read
How to map AI agents and humans to one identity model with OIDC
Put AI agents and people behind one OIDC issuer, register each agent as its own client, and map claims to the username your Kubernetes API server records for delegated…
November 5, 2026 · 7 min read
What is agent identity?
Agent identity gives each AI agent its own auditable principal. A Kubernetes service account names a workload, but cannot show which agent acted or for whom.
November 4, 2026 · 5 min read
Kyverno policies for AI agents
Kyverno policies for AI agents: enforce Pod Security Restricted controls on agent pods at admission with a ValidatingPolicy, then test the policy with the CLI.
November 4, 2026 · 6 min read
How to collect Kubernetes compliance evidence for auditors
Pair Kyverno policy reports, which show current state, with the API server audit log, which shows history. Reports alone are not an audit trail by themselves.
November 3, 2026 · 6 min read
How to allowlist MCP servers in Cursor
Allowlist MCP servers in Cursor by command or URL pattern in the team dashboard, then limit tools and network access for each server. Enterprise plans only.
November 3, 2026 · 6 min read
Cursor rules vs enforced policy: what stops an agent
Cursor rules are guidance, not enforcement. Learn which Cursor controls block a tool call or command, which only advise, and how to configure each one of them.
November 3, 2026 · 8 min read
Measuring the value of AI coding agents
Claude Code's docs count usage, cost, PRs and accepted lines, not value. Record an outcome baseline before rollout, then compare the vendor counts against it.
November 1, 2026 · 5 min read
How to onboard 1,000 developers to Claude Code without manual setup
Onboard many developers to Claude Code with one managed policy, delivered by server, MDM or file, then checked per machine with /status and claude doctor.
November 1, 2026 · 5 min read
How to govern Claude Cowork in the enterprise
Govern Claude Cowork with managed keys for egress, folders and tools, per-group policies, OpenTelemetry export and Compliance API transcript retrieval.
November 1, 2026 · 7 min read
Using Claude on Amazon Bedrock with Claude Code and Cowork
Set up Claude Code and Claude Cowork on Amazon Bedrock with AWS sign-in, pinned model IDs, a managed model allowlist and the IAM policy that each one needs.
October 31, 2026 · 10 min read