FikaWorks Argus

The map

One map of everything you run

Argus builds an ontology of your estate: every cloud resource, cluster, workload, image, repository and identity, and how they connect. Each thing has one identity across all your environments, and every change and scan keeps the map current.

What the map lets Argus do

1. 1 Fix once, everywhere. An image running in prod-eu, prod-us and staging is one thing on the map, so one fix closes the gap in every environment.
2. 2 Read attacks across your estate. An exposed workload, an over-privileged identity and an unexpected process, seen together on the map, read as an attack in progress.
3. 3 Learn from its own work. Every approval, verified fix, confirmed link and closed incident teaches it more about your estate, and that learning stays in your environment.

We are fine-tuning an open-weight model to read the map. It runs where Argus runs, in our service, in your cloud or on the private box, so your data does not leave.

Scout

It watches, and contains within the limits you set

Scout follows each environment’s own signals as they happen. Noise is filtered inside the cluster and raw signals never reach the model. Signals are grouped and checked against the map before anything wakes the agent, so an attacker can’t talk to it through a log line.

Containment in three tiers

1. 1 Responses your team approved in advance run at once, with no model in the path.
2. 2 Reversible containment you allow per environment, such as isolating a pod, scaling to zero or blocking an address, runs without asking.
3. 3 Everything else goes on a card with its exact effect. The lasting fix, such as an upgrade, a library or an architecture change, goes to the environment’s plan.

For each environment, Scout shows whether it can see: Active, Degraded or Off. Every signal source is proven every 30 minutes by a canary, a harmless event that must reach Argus within 5 minutes. A missing source becomes an issue for the plan to fix.

Prowl

It tests your defences on purpose

Prowl runs engagements against your own infrastructure. An owner authorizes each one on a card, with its scope, its time window and its rules. A kill switch stops every scan at once, and every scan is recorded in the ledger.

What it covers

1. 1 Clusters, images, networks and web apps, from inside the cluster or from outside.
2. 2 Each finding becomes an issue about one affected thing, with its evidence.
3. 3 The environment’s plan fixes it like any other gap.

Scout and Prowl together

A built-in purple team

Scout is never told an engagement is running; the platform enforces it. When an engagement closes, Argus scores every scan: did Scout detect it, how fast, and did it contain it? Each miss becomes a detection issue, so your defences improve the same way vulnerabilities get fixed.

A person answering a Scout card is told when the activity matches an authorized engagement, so nobody blocks their own test by mistake.

Works with

Argus connects to the clouds, platforms and tools you already run, with multiple organisations from day one. Among them:

• AWS
• Azure
• Google Cloud
• Kubernetes
• GitHub
• GitLab
• Helm
• Argo CD
• Flux
• Trivy
• Kyverno
• Falco
• and more

Using something else? We build custom integrations when your environment needs them.

Where it runs

• We run it for you

  Available

  Connect your environments and Argus does the rest from our infrastructure.

• In your cloud

  Available

  Argus runs inside your own cloud account or Kubernetes, installed with Helm, on a model you host or choose.

• The private box

  Next

  An appliance in your own infrastructure with our own fine-tuned model built in. Nothing leaves the building, even air-gapped.