IT Operations

What the IT team is asking for

An application got slow at eleven this morning. The answer is somewhere across a load balancer, four application servers, a database and a cluster that rolled out a change last night. Someone opens six terminals and starts pasting commands into a chat window. Two hours later there is an answer and no record of how it was reached.

• Ask a question once and have it answered on every server it concerns.
• Get back the commands that ran, not a summary of what an assistant believes happened.
• Find configuration drift across an estate before an auditor does.
• Keep the investigation record, so the next person starts from it rather than from zero.

Hosts are the execution and evidence workers

A small agent runs on each machine. The model never touches the machine; it asks the host worker to run an approved diagnostic, and the host worker returns what happened. That is why the answer can be checked.

A host joins the fleet with one install command and an enrollment token; revoking the host unpairs it.

Read-only execution policy is the default on every enrolled host; the full policy is an explicit, per-host choice.

Every answer shows the exact commands that ran, their exit codes and output, per host.

The diagnostic skills an agent may use

Skills come from a signed vendor catalogue or your own uploads, and can be enabled or disabled per workspace.

The built-in catalogue covers the estate an operations team actually runs, and each skill is a scoped set of diagnostics rather than a free-form prompt:

• linux-host-triage
• windows-host-triage
• active-directory-investigation
• exchange-server-triage
• kubernetes-cluster-triage
• virtualization-platform-triage
• network-firewall-triage
• storage-filesystem-triage
• identity-dns-edr
• file-provenance
• infrastructure-incident-rca

A skill declares what it can and cannot do, the tools it needs, the paths it may read, and how it is evaluated.

How the work is done

An agent is a model, a set of approved skills, tools and a policy, applied to a scope.

Agents reach your systems through approved connectors — APIs, databases, applications — each with its own permissions.

Changes wait for human approval before they run.

Where it runs

The whole stack self-hosts with one command; an air-gap bundle ships every image.

Bring your own model — OpenAI-compatible endpoints, Ollama, Anthropic, or Cloudflare AI Gateway / Workers AI — with keys envelope-encrypted at rest.

Sign in with your own identity provider (OIDC with PKCE); no shipped default administrator password.

Customer-controlled inference and execution boundaries.