Platform overview

A Closer Look at the Astrix Threat Modeling Platform

Astrix is threat modeling software that takes the architecture your teams already document and turns it into a structured, scored threat model. This page walks through what goes in, what the platform does with it, and what your security and engineering teams get out.

You provide

  • API specificationsOpenAPI / Swagger
  • Infrastructure-as-codeTerraform, CloudFormation, ARM, Kubernetes
  • DiagramsVisio, draw.io, images
  • Repositories or a descriptionGitHub, plain language

Astrix

System modelDescription, components and data flows
Framework analyzersSTRIDE · PASTA · VAST · SAIF · MAESTRO
Finding enrichmentCVSS · CWE · MITRE ATT&CK · mitigation guidance
Review and collaborationRoles, comments, audit trail, snapshots

Your teams get

  • Scored threat findingsWith mitigation guidance
  • Attack paths and heat mapsWhere to focus first
  • Compliance gap viewThreats mapped to controls
  • Tickets and reportsJira, GitHub, Azure DevOps, PDF, DOCX
Conceptual model of how Astrix is organized: inputs, platform layers and outputs. It is not a product screenshot.

Introduction

Understand the platform. Strengthen the design.

Threat modeling is a structured way of asking what could go wrong with a system before attackers find out. Security practitioners often frame it as four questions. They apply to any threat modeling effort, with or without a tool.

Astrix is a threat modeling platform built around those questions. It does not replace your reviewers' judgment. It removes the manual work around them: redrawing architectures, applying a methodology by hand, scoring findings case by case, and chasing mitigations through slide decks.

The table shows where the general practice ends and where Astrix provides specific support.

  1. What are we working on?Imports specs, infrastructure-as-code, diagrams or repositories, or takes a plain-language description, to capture the system's components and data flows.
  2. What can go wrong?Runs the analyzer for the framework you choose (STRIDE, PASTA, VAST, SAIF or MAESTRO) to identify threats against that model.
  3. What are we going to do about it?Scores each finding (CVSS), maps it to CWE and MITRE ATT&CK, suggests mitigations, and pushes work to Jira, GitHub Issues or Azure DevOps.
  4. Did we do a good enough job?Your reviewers accept or adjust results, comment and track remediation. When the design changes, they re-analyze it and keep snapshots.

Platform at a glance

Four areas, one threat modeling platform

Each area has a distinct job. Together they take a threat model from first import to tracked mitigation.

Area 1

System context

Responsible for
Capturing what you are building.
Uses
API specs, IaC, diagrams, repositories or a written description.
Gives you
A system model with components and data flows, without re-entering it by hand.
Workflow stage
Steps 1–2
Area 2

Threat analysis

Responsible for
Identifying threats against the model.
Uses
The system model and the framework you select.
Gives you
A methodology-specific draft threat model, generated by a dedicated analyzer for that framework.
Workflow stage
Step 3
Area 3

Risk insight

Responsible for
Showing which threats matter most.
Uses
Identified threats and their context.
Gives you
CVSS scores, CWE and ATT&CK mapping, attack paths, heat maps and compliance gaps.
Workflow stage
Step 4
Area 4

Delivery and review

Responsible for
Turning findings into decisions and work.
Uses
Reviewed findings, roles and your issue trackers.
Gives you
Tickets, remediation tracking, reports, comments and snapshots over time.
Workflow stage
Step 5

Core platform capabilities

What each capability does, and why it matters

Six capability groups, each described by the need it addresses, what Astrix actually does, and where it fits in a threat modeling practice. For a feature-by-feature reference, see Astrix features and capabilities.

A · System context

Architecture import

Most threat models start with someone redrawing the architecture by hand, and that copy is out of date the moment the real design moves on.

Astrix starts from artifacts your teams already maintain. It accepts OpenAPI and Swagger definitions; Terraform, CloudFormation and ARM templates; Kubernetes YAML; Visio and draw.io diagrams; uploaded architecture images; and GitHub repositories. When none of those exist yet, as in early design, you can describe the system in plain language.

Where it fits"What are we working on?": establishing system context.
Why it mattersLess manual entry, and a model that reflects the system as it is defined.

B · Threat identification

Framework-specific threat analysis

Teams pick different methodologies for different systems, and results vary with whoever runs the session.

Astrix has a dedicated analyzer for each of five methodologies, each with its own detection logic: STRIDE, PASTA, VAST, SAIF and MAESTRO. You choose the framework and the analyzer generates a draft threat model from the system model.

Analysis is AI-assisted. By default it runs on an enterprise AI service hosted in the EU, and each request passes multiple guardrail layers before and after the model call. If no AI provider is configured, a built-in rule engine runs the analysis inside Astrix. Either way, the output is advisory: your reviewers accept, adjust or reject it.

Where it fits"What can go wrong?": identifying threats.
Why it mattersA consistent methodology across teams, and expert time spent reviewing rather than drafting.

C · Threat assessment

Scored, contextualized findings

A long list of threats without severity or context leaves teams arguing about what to fix first.

Each finding Astrix produces carries a CVSS score, a CWE identifier for the underlying weakness, a MITRE ATT&CK mapping for the attacker technique, and specific mitigation guidance. Severity-by-likelihood heat maps show where findings cluster across the model.

For regulated environments, compliance gap analysis maps threats to the controls they affect in HIPAA, GDPR, PCI DSS, SOX, FISMA, NIST SP 800-53 and ISO 27001. This shows where threats touch your control set. It is not an audit, a certification or a guarantee of compliance.

Where it fitsEvaluating and prioritizing findings.
Why it mattersShared, standard vocabulary between security, engineering and risk teams.

D · Threat analysis

Attack path visualization

Individual findings rarely show how an attacker could chain weaknesses together to reach something valuable.

Astrix maps attack paths as a graph, with kill-chain analysis that places threats along the stages of an attack, and lateral-movement detection that shows how an attacker could move between components after gaining a foothold.

Where it fitsUnderstanding attack surface and prioritizing controls.
Why it mattersA choke point that breaks several paths at once is often the best place to invest.

E · Mitigation planning

Mitigation guidance and remediation tracking

Findings recorded in documents tend to stall: no owner, no ticket and no record of what was decided.

Every finding comes with specific mitigation guidance. You can export selected findings as work items to Jira, GitHub Issues or Azure DevOps, and track remediation status in Astrix, where changes are kept in an append-only remediation log. Slack and Microsoft Teams notifications keep the right people informed.

Astrix documents and tracks mitigation decisions. It does not change your systems or apply controls itself; your teams do that through their normal delivery process.

Where it fits"What are we going to do about it?"
Why it mattersEach threat has an owner, a status and a history.

F · Collaboration and ongoing review

Collaboration, reporting and living threat models

A threat model is only useful if the right people can work on it and it keeps pace with the design.

Each threat model has owner, editor and viewer roles. You can organize users into teams and share models with expiring links. Teams edit together in real time, discuss findings in comment threads, and every change is recorded in an audit trail.

When the architecture changes, re-analyze the model and keep snapshots of earlier versions. Export threat models and reports as PDF or DOCX, or schedule reports by email for stakeholders who do not use the platform.

Where it fits"Did we do a good enough job?": review and maintenance.
Why it mattersA threat model that stays current, with a reviewable record of how it evolved.

Threat modeling workflow

From system design to structured threat analysis

The threat modeling workflow Astrix supports, stage by stage. Not every stage is automated: each row shows what your team does and what Astrix provides.

  1. Establish the system context

    Decide what is in scope: the application, its components and its boundaries.

    Your team

    Chooses the system and supplies specs, IaC, diagrams, a repository or a description.

    Astrix

    Imports those artifacts and captures the system description and components.

  2. Examine interactions and trust boundaries

    Understand how information moves and where security assumptions change.

    Your team

    Checks that the model is complete and confirms where data crosses trust boundaries.

    Astrix

    Represents the data flows between components as part of the threat model.

  3. Identify potential threats

    Apply a methodology to find what could go wrong.

    Your team

    Selects the framework that fits the system and reviews the draft threats.

    Astrix

    Runs the dedicated analyzer for that framework to generate a draft threat model.

  4. Evaluate and document findings

    Decide which threats matter and how much.

    Your team

    Accepts, adjusts or rejects findings and agrees priorities.

    Astrix

    Provides CVSS scores, CWE and ATT&CK mapping, heat maps, attack paths and compliance gaps.

  5. Plan and review mitigations

    Act on findings and keep the model current.

    Your team

    Owns and implements mitigations, and re-runs the review when the design changes.

    Astrix

    Supplies mitigation guidance, exports work items, tracks remediation, keeps snapshots and produces reports.

Architecture Threat Modeling

Bring architectural security questions into focus

Architecture Threat Modeling examines a design before it is deployed: its components, interfaces and dependencies, the data flowing between them, and the trust boundaries those flows cross. It brings out the security assumptions the design depends on, so teams can test them deliberately rather than discover them in production.

In Astrix, the system model holds the components and data flows, the selected framework's analyzer raises threat scenarios against them, and attack path visualization shows how those scenarios connect across the attack surface.

Conceptual architecture with five security questions A client application calls a public API over the internet. The API uses an external identity provider for sign-in, calls an internal service, and the service reads a shared database and sends outbound webhooks to a partner system. Dashed lines mark trust boundaries. Numbered markers one to five correspond to the security questions listed beside the diagram. INTERNETYOUR PLATFORMIDENTITYPARTNERS Client appPublic APIInternal service Shared databaseIdentity providerPartner webhook 1 2 3 4 5
  • Data flow
  • Trust boundary
  • Numbered security question
Conceptual example of a customer system, not the Astrix platform architecture and not a product screenshot.

Questions a design review should answer

  1. Is every request to the public API authenticated?Spoofing at the internet trust boundary.
  2. What happens when the API is flooded?Denial of service at the exposed entry point.
  3. Are identity-provider tokens validated correctly?A security assumption about an external dependency.
  4. Could one customer's request read another customer's data?Information disclosure in a shared data store.
  5. Can outbound webhooks be pointed at internal addresses?Tampering and disclosure through an integration.

Architecture threat modeling looks for

  • Missing or misplaced controls in the design
  • Risky data flows and trust-boundary crossings
  • Unsafe assumptions about dependencies
  • Attack paths across components

Security testing looks for

  • Vulnerabilities in code that has been written
  • Misconfigurations in deployed systems
  • Exploitable behaviour at runtime
  • Regressions between releases

The two complement each other. Threat modeling shapes the design; testing checks the implementation. Astrix supports the first and does not replace the second.

Platform security

Designed with security in mind

A threat model describes how your systems could be attacked, so the platform that holds it must be trustworthy itself. Each item below is labelled with its status: a control implemented in the platform or a documented commitment. We also threat-model our own platform with Astrix.

  • ImplementedBuilt into the platform today
  • CommitmentDocumented policy or contractual commitment
Implemented

Tenant isolation

Each customer organization has its own dedicated database, and organization membership is verified on every request.

Implemented

Authentication and access

Single sign-on via OpenID Connect with your identity provider, multi-factor authentication, and organization, team and per-model roles.

Implemented

Encryption

Data is encrypted in transit and at rest using industry-standard encryption.

Implemented

EU hosting

Hosted in the EU on enterprise-grade cloud infrastructure. Dedicated deployments in other regions are available on request.

Implemented

Controlled AI data flow

Only the system description, components and data flows are sent for analysis, through multiple guardrail layers. Never identities, evidence files or credentials.

Implemented

Audit logging and upload checks

Per-organization audit logs that your admins can view, plus security checks on uploaded files.

Commitment

No training on your data

Neither Astrix nor its default AI provider trains models on your content or outputs.

Commitment

Data ownership and deletion

Export at any time. On exit, data is deleted within contractual timeframes and deletion is confirmed in writing. Breach notification within contractual timeframes.

Security lifecycle

Where Astrix fits in the security lifecycle

Threat modeling is most valuable during design, and it stays useful as systems change. It complements secure design reviews, application security testing and vulnerability management; it does not replace them.

  1. Supports

    Requirements and planning

    Describe the planned system in plain language to raise threats early, before the design is fixed.

  2. Primary role

    Architecture and design

    Import or describe the design, analyze it, and agree mitigations while changes are still cheap.

  3. Supports

    Development

    Mitigations arrive as Jira, GitHub Issues or Azure DevOps work items with guidance attached.

  4. Other tools

    Security testing

    SAST, DAST and penetration testing check the implementation. Findings from the threat model can help focus that testing.

  5. Supports

    Release review

    PDF or DOCX reports show which threats were identified, accepted or mitigated.

  6. Supports

    Change and maintenance

    Re-analyze when the architecture changes, and keep snapshots of earlier versions.

Use cases

How teams put the platform to work

Reviewing a new application architecture

Security architect · new customer-facing service

Starting point
An OpenAPI definition and Terraform for a service due to be built next quarter.
Objective
Agree design-level security requirements before development starts.
How Astrix supports it
Import both artifacts, run STRIDE, review the scored findings and attack paths with the engineering lead.
Output
A reviewed threat model, mitigation work items in the team's tracker, and a PDF report for sign-off.

Assessing a redesign or new integration

AppSec team · existing system adding a payment provider

Starting point
An existing threat model, and a design change that adds an external dependency.
Objective
Understand what the new data flows change about the system's exposure.
How Astrix supports it
Update the model, re-analyze it, and keep a snapshot of the earlier version for reference.
Output
New or changed findings with mitigation guidance, tracked to resolution.

Securing an AI or agentic feature

Product security · LLM-powered or multi-agent application

Starting point
A design that adds a model, tools and agents to an existing product.
Objective
Apply a methodology designed for AI systems, not only a traditional one.
How Astrix supports it
Analyze the design with SAIF or MAESTRO alongside STRIDE, using their dedicated analyzers.
Output
AI-specific threats with scores, mappings and mitigation guidance.

Running consulting engagements

Security consultancy or channel partner

Starting point
Several client organizations, each needing its own architecture assessment.
Objective
Run a consistent method for each client while keeping client data separate.
How Astrix supports it
Partner administration across client organizations, each held in its own isolated database.
Output
Repeatable threat models and PDF or DOCX deliverables for each client.

What sets Astrix apart

Strengths within a focused scope

Astrix is a threat modeling platform, not a general security suite. These are the strengths that follow from that focus.

  1. A dedicated analyzer per methodology

    STRIDE, PASTA, VAST, SAIF and MAESTRO each have purpose-built detection logic, including frameworks written for AI and multi-agent systems.

  2. Starts from real design artifacts

    Specs, IaC, diagrams and repositories become the threat model's input, so the model reflects the system as it is defined.

  3. Findings carry standard context

    CVSS, CWE and MITRE ATT&CK give security, engineering and risk teams a shared vocabulary, with mitigation guidance attached.

  4. Built for review, not just generation

    AI output is advisory. Roles, comments, audit trails and remediation logs keep people accountable for every decision.

  5. Designed for sensitive architecture data

    A database per customer, EU hosting, controlled AI data flow, and no training on customer data.

Capabilities and outcomes

From platform capability to practical value

The capability is what Astrix provides. The value is what teams can achieve when they use it as part of their own process.

Astrix platform capabilities and the potential value to teams
Platform capabilityWhat it providesPotential value
Architecture importA system model from specs, IaC, diagrams or a descriptionLess time redrawing systems, more time reviewing them
Framework analyzersA methodology-specific draft threat modelMore consistent analysis across teams and reviewers
Scored, mapped findingsCVSS, CWE, ATT&CK and mitigation guidance per findingClearer prioritization and communication of design-level risk
Attack path visualizationGraph-based paths, kill-chain stages and lateral movementBetter visibility into how weaknesses combine
Remediation trackingWork-item export, status tracking and an append-only logVisible ownership and history of security decisions
Re-analysis and snapshotsUpdated models and earlier versions kept for referenceThreat models that keep pace with the architecture

Astrix supports threat modeling and security design review. It does not eliminate threats, prevent breaches or guarantee regulatory compliance.

Resources

Platform FAQ

What is the Astrix Threat Modeling Platform?

Astrix is threat modeling software from Cognexaa. It builds a threat model from your architecture artifacts, analyzes it with a framework-specific analyzer, and produces scored findings with mitigation guidance that teams can review, track and keep up to date.

What is included in the Astrix platform?

The platform covers:

  • architecture import
  • five framework analyzers
  • CVSS, CWE and MITRE ATT&CK enrichment
  • attack path visualization and heat maps
  • compliance gap analysis
  • remediation tracking with Jira, GitHub Issues and Azure DevOps export
  • collaboration features, snapshots, and PDF/DOCX reporting

It is offered as Starter, Professional and Enterprise plans. Usage limits differ by plan, and report export is included in all three. Pricing is available on request.

How does Astrix support Architecture Threat Modeling?

Astrix captures a system's description, components and data flows from what you import or describe, then analyzes that model rather than individual lines of code. Attack path visualization shows how threats connect across components. See the architecture example.

How much of the threat modeling workflow is automated?

Astrix automates the drafting:

  • importing artifacts
  • generating a threat model with the selected analyzer
  • scoring and mapping findings
  • exporting work items

People stay responsible for scope, for accepting or adjusting findings, for agreeing priorities and for implementing mitigations. AI output is advisory. See who does what at each stage.

Which threat modeling methodologies does Astrix support?

Five, each with a dedicated analyzer:

  • STRIDE
  • PASTA: Process for Attack Simulation and Threat Analysis
  • VAST: Visual, Agile and Simple Threat modeling
  • SAIF: Google's Secure AI Framework
  • MAESTRO: for multi-agent AI architectures
Does Astrix replace penetration testing or code scanning?

No. Threat modeling examines the design; penetration testing, SAST and DAST examine the implementation. Astrix complements those activities and can help focus them on the areas of highest design-level risk. See where Astrix fits.

How does Astrix protect our architecture data?

Each customer organization has its own database, hosted in the EU on enterprise-grade cloud infrastructure. Data is encrypted in transit and at rest. Access is controlled through SSO, MFA and role-based permissions, with audit logs. Only architecture metadata is sent for AI analysis, and customer data is not used to train models. See the security details.

How can an organization evaluate the platform?

Book a demo by emailing sales@cognexaa.com, ideally with a system you would like to see modeled. The Cognexaa team can also walk your security reviewers through the platform's architecture and controls under NDA.

See how Astrix fits your threat modeling workflow

Explore the platform, understand its capabilities, and evaluate how it could support your security architecture and threat modeling work.

sales@cognexaa.com · Back to the Astrix homepage