SIDEREAL

Continuous Security Control Validation

Your controls are configured. Are they working?

A NetworkPolicy can be defined and not enforcing. An admission webhook can be configured and silently disabled. Sidereal probes your cluster's security controls on a continuous schedule and audits whether they are operationally effective.

Filling the operational validation gap

Configuration scanners, detection tools, and attack simulators are each valuable parts of a security stack. They do their jobs well. Sidereal is built for the operational validation layer that sits alongside them, by continuously probing whether controls are active, mapped to compliance frameworks, and generating the evidence your ISSO needs.

The swivel chair problem

An ISSO validating Kubernetes security controls today pivots between disconnected tools: Kubescape for posture, Falco dashboards for detection, the SIEM for audit records, manual crosswalks to NIST 800-53, hand-built reports, spreadsheet POA&Ms. Each tool covers one piece. None of them connect the pieces. The ISSO becomes the integration layer.

Sidereal is a single operator that probes controls, maps results to frameworks, exports to your SIEM, and generates the evidence package. No manual integration required.

Configuration scanners

Kubescape, kube-bench, Trivy

Verify that security configurations are present and match known benchmarks. Well-suited for posture assessment and CIS compliance checks.

Sidereal adds active probing to verify those configurations are operationally enforcing, not just defined.

Attack simulators

Stratus Red Team, Atomic Red Team

Purpose-built for adversary simulation and red team exercises. They generate realistic attack telemetry to test detection and response capabilities.

Sidereal runs on a continuous schedule and maps every result to compliance controls for ongoing assurance.

Detection tools

Falco, Tetragon

Monitor runtime behavior and alert on suspicious activity. A critical layer of any Kubernetes security stack and a direct input to incident response.

Sidereal validates that your detection pipeline is actively alerting as expected, so you know it is working before an incident occurs.

This has happened in production

The gap between a configured control and a working control is documented, recurring, and exploitable. These are real incidents, not hypothetical failure modes.

Kyverno admission bypass via finalizers

Kyverno enforcing image policy with validationFailureAction: Enforce. Any resource with a Kubernetes finalizer attached has its deletionTimestamp set, which caused Kyverno to skip policy evaluation entirely. A developer-level account was sufficient to place any resource in a permanently policy-exempt state. (CVE-2023-34091, patched in Kyverno 1.10.0.)

Sidereal's admission probe fires against the live enforcement layer. A bypassed resource causes the probe to return Ineffective regardless of what the policy configuration says.

EKS NetworkPolicies silently unenforced

NetworkPolicy objects applied to Amazon EKS clusters using the default VPC CNI plugin. The API accepted every policy and returned success. Until Amazon added native enforcement in October 2023, none of those policies filtered any traffic. Configuration scanners reported them as present. Inter-pod traffic was never blocked.

Sidereal's network policy probe attempts the connection and checks the actual packet verdict, not whether a policy object exists.

io_uring rootkit bypasses Falco and Tetragon

Falco and Tetragon deployed for runtime detection. ARMO's "Curing" rootkit conducts all operations through Linux's io_uring interface, bypassing the syscall layer both tools instrument. Full data exfiltration with zero alerts from Falco, Tetragon, or Microsoft Defender for Endpoint. (ARMO, April 2025.)

Sidereal's detection probe independently verifies the alert was raised. If the detection pipeline has a gap, the probe finds it.

See all documented incidents, organized by probe surface →

What Sidereal does

Sidereal is a Kubernetes-native operator for continuous security control validation on federal systems. It runs targeted, low-impact probes against a live cluster to verify that security controls are operationally effective, not merely configured. Results are mapped to compliance frameworks and written to an append-only, HMAC-verified audit log.

Continuous Active Validation

Probes run on configurable schedules, firing real actions against real enforcement layers. Not cached configuration state.

Multi-Framework Compliance

Every result tagged with NIST 800-53, CMMC, CJIS, IRS 1075, HIPAA, NIST 800-171, and Kubernetes STIG controls automatically.

Detection Layer Validation

Fires known-bad syscall patterns, then independently verifies the detection pipeline (Falco or Tetragon) raised the expected alert.

ISSO-Ready Reports

Continuous monitoring summaries, POA&M entries, coverage matrices, and OSCAL evidence packages generated from probe results.

How it works

1

Deploy

Install via Helm into the sidereal-system namespace. Choose a deployment profile for your CNI and admission controller. Default execution mode is dryRun.

2

Discover

The controller scans your cluster and generates SiderealProbeRecommendations. Review and promote what fits. No probe authoring required to get started.

3

Validate

Probes run on schedule. Results are written to an HMAC-verified audit log, mapped to compliance frameworks, and exported to your SIEM or report pipeline.

Graduated adoption

dryRun Validates configuration without executing probes. Safe to run immediately after install.
observe Probes execute live. Results are recorded. No incidents or alerts generated.
enforce Full operation. Incidents created on control failure, IR webhook notified.

Probe surfaces

Six built-in probe surfaces cover the control categories federal assessors most commonly examine. A custom probe interface lets agencies extend coverage to their own surfaces.

RBAC

ServiceAccount permission boundaries are enforced and denying unauthorized operations.

NetworkPolicy

East-west traffic restrictions are actively blocking unauthorized paths at the CNI layer.

Admission Control

Admission policies reject non-compliant workload specs at the API server.

Secret Access

Workloads cannot access secrets outside their authorized namespace.

Detection Coverage

Known-bad behaviors trigger expected alerts in your detection pipeline (Falco or Tetragon).

Custom

Operator-extensible probe surface. Bring your own probe binary with a standardized contract.

Ready to validate your controls?

Sidereal installs in minutes and starts discovering probe recommendations immediately.