Skip to main content

Welcome to NebCore

NebCore gives you two command-line tools that work together: NebCLI, the tool you log in with and use to get short-lived cloud, cluster, and code-host credentials, and NebGuard, policy-driven guardrails that sit between your AI coding agent and the actions it takes. You install them once, log in once, and from then on NebCLI brokers the temporary credentials your everyday tools need while NebGuard watches what your AI agent does and steps in before anything dangerous runs.

The two tools​

  • NebCLI is the single entry point for platform actions from a terminal. It confirms who you are, hands the standard tools (the cloud CLI, the cluster CLI, the chart CLI, your version-control client) short-lived credentials, and checks that infrastructure changes are valid before they ship. It is designed so you never create or store a long-lived secret yourself.
  • NebGuard is a seatbelt and a co-pilot for AI coding agents such as Claude Code. On every action the agent is about to take, it checks the action against a built-in rule library and responds in one of four ways: allow, warn, block, or guide. It also records every decision so you get an audit trail that maps to compliance frameworks.

NebCLI and NebGuard together​

Each tool works on its own: NebGuard licenses itself through nebguard setup. If you also use NebCLI, it writes the license NebGuard needs, so running nebcli login turns on the full Paid-tier guardrails with no separate signup or consent prompt. See NebCLI and NebGuard together for how this works.

Get started​

  • Installation: install both tools, verified.
  • Quickstart: log in, get a temporary cloud credential, and watch a guardrail fire, in a few minutes.
  • NebCLI: the full command-line reference.
  • NebGuard: how the guardrails engine, rules, and licensing work.
note

This site reflects the most recent public release. Some streamlined paths noted on these pages are still rolling out; each one is labeled where it appears.