AI Agent Force - LumiOps

A team of specialized agents.

A catalog of AI agents — personal, org-shared and UnityOne AI-curated. Each one has a defined role, its own guardrails, a tracked success rate, and a Direct button to put it to work.

curated agents

capability tiers

Dry-run

before anything ships

◈Agent Team81 agents

⌕

Direct any agent to investigate, diagnose or remediate. Each agent runs within its guardrails — capability tier, scope and approval policy enforce what it can do.

No agents match that search.

CAPABILITY TIERS

Every agent earns its reach.

Agents don’t get blanket access. Each one sits in a tier that defines what it may do — and
you promote it only when its success rate says it’s ready.

discover

Look, don’t touch

Read-only investigation. It gathers evidence and reports back — nothing changes.

manage

Act, within limits

Scoped changes against tagged targets, bounded by a blast radius you
set.

automate

Run on its own

Trusted, repeatable remediation across sites — earned by a proven track
record.

GUARDRAILS & GOVERNANCE

You set the blast radius.

Scope, tier and approval policy are attached to the agent
itself — not bolted on at runtime.

1

Scope

Which sites, environments and tags an agent may ever touch.

2

Blast radius

A hard ceiling on how many targets a single run can affect.

3

Dry-run mode

It executes the whole plan and shows you the diff — changing nothing.

4

Approval policy

Act, ask, or advise — decided per agent, changeable any time.

PERSONAL · ORG · CURATED

Three places agents come from.

PERSONAL

Private to you

Agents you create or clone. They live in your workspace and nobody else sees them.

Start one, or fork anything below

ORG

Shared with your workspace

Team-specific agents wired to your own systems — your proxies, your metrics, your runbooks.

1 in this workspace

UNITYONE

Curated by UnityOne

Production-tested agents across cloud, Kubernetes, identity, certificates, spend and infrastructure health.

79 ready to direct

CAPABILITIES

What makes it a team, not a
toolbox.

◈

Domain AI agents

Specialists for compute, network, storage, cloud, database, identity and security.

👤

Persona & role-based

Agents shaped around a job — on-call, auditor, investigator, sentinel.

🛡

Guardrails & governance

Scope, tier and approval policy travel with the agent, enforced on every run.

◎

Blast-radius controls

A hard cap on targets per run, so a bad call stays small.

▷

Dry-run mode

See exactly what would change before a single thing does.

📈

Success tracking

Every agent carries a live success rate. Promote the strong ones, retrain the rest.

⑂

Cloning & forking

Fork any org or UnityOne agent into your own and tune it without touching the original.

💬

Natural-language direction

Hit Direct and tell it what you need in a sentence.

▤

Full auditability

Every direction, decision and change is logged and attributable.

Get started

Put an agent
to work today

Start in discover mode on your own stack. Promote what proves itself.

discover

manage

automate

1

2

3

4

Build

Orchestrate

Cowork

Resolve

About Us

About Us

Product

Coworker

Orchestrator

Product

Orchestrator

Coworker

Build

Platforms

Estate Models

Agents

Observability

Service Desk