Threat modeling platform · by Cognexaa

Threat Modeling Software for Secure-by-Design Architecture

Astrix turns your system architecture into a structured threat model. Import an API specification, infrastructure-as-code or a diagram, analyze it with STRIDE, PASTA, VAST, SAIF or MAESTRO, and give security and engineering teams scored findings with mitigation guidance before the design is built.

  • Five threat-modeling frameworks
  • CVSS, CWE and MITRE ATT&CK mapping
  • Hosted in the EU
Example threat modelFramework: STRIDE
Example architecture with threat markers A user's browser connects over HTTPS to an API gateway, which calls an orders service. The service reads a database in the data zone and calls an external payment provider. Dashed red trust boundaries separate the internet, application, data and third-party zones. Red markers flag spoofing at the internet boundary, denial of service at the gateway, tampering on the database flow and information disclosure on the payment flow. INTERNETAPPLICATIONDATATHIRD PARTY User browserAPI gateway Orders APIDatabasePayments HTTPS S D T I
  • CVSS 9.0Denial of service at the API gatewayCWE-400. An attacker floods the system with requests to exhaust resources. Mitigation: rate-limit requests per IP at the gateway.
  • CVSS 8.0Elevation of privilegeMisconfigured access controls could let an attacker gain elevated privileges. Mitigation: enforce role-based access control under least privilege.
Illustrative example based on findings described in Astrix product material. This is a conceptual diagram, not a product screenshot.

The challenge

Threat modeling often happens too late, and too slowly

Every new service, architecture change and release needs a threat model, and many security teams still build them by hand, in workshops and slide decks, one review at a time.

Effort to fix a flaw

Effort rises at each stage, lowest at design and highest at release.

  1. Design
  2. Build
  3. Test
  4. Release

Illustrative. Manual reviews often land at test or release.

Manual reviews slow releases

Hand-run STRIDE or PASTA sessions depend on a few experts, so reviews queue up behind every release.

Models go stale

A threat model drawn for one design is out of date as soon as the architecture changes.

Findings stay on slides

Threats recorded in documents are rarely tied to owners, tickets or remediation status.

What this means for the business

  • Costlier fixes. Design flaws found after build mean reworking code and infrastructure instead of changing a design.
  • Slower security reviews. Delivery teams wait, or ship without a review.
  • Unclear ownership. No one is sure which threats were accepted, mitigated or forgotten.
  • Inconsistent results. Review quality varies with whoever runs the session.

The Astrix platform

Make security an integral part of architecture design

Astrix starts from the architecture artifacts your teams already produce. Import them, choose a threat modeling framework, and Astrix generates a threat model that your reviewers refine, prioritize and track through to mitigation.

Analysis is AI-assisted. By default it runs on an enterprise AI service hosted in the EU, and every request passes multiple guardrail layers before and after the model call. Only the system description, components and data flows you submit are sent; never user identities, evidence files or credentials. Without an AI provider configured, a built-in rule engine runs the analysis inside Astrix.

AI output is advisory. Your users review and accept every result. Take a closer look at the platform.

Start from what you already have

  • OpenAPI / SwaggerAPI definitions
  • Terraform, CloudFormation, ARMInfrastructure-as-code
  • Kubernetes YAMLWorkload manifests
  • Visio and draw.ioArchitecture diagrams
  • Diagram imagesUploaded architecture images
  • GitHub repositoriesSource repositories
  • Plain-language descriptionDescribe the system in your own words

Core capabilities

From architecture to actionable threat model

Each capability addresses a different gap in manual threat modeling: coverage, prioritization, visibility, compliance context and follow-through.

Framework-specific threat analysis

Astrix includes a dedicated analyzer for each supported methodology, with its own detection logic, rather than one generic scan relabelled five ways. Choose the framework that fits the system: STRIDE for general architectures, PASTA for risk-centric reviews, or SAIF and MAESTRO for AI and multi-agent systems.

  • STRIDE
  • PASTA
  • VAST
  • SAIF
  • MAESTRO
Compare the frameworks

Scored, actionable findings

Every threat comes with what a reviewer needs to act on it, not just a label.

  • CVSS score for severity
  • CWE identifier for the weakness
  • MITRE ATT&CK mapping for attacker technique
  • Specific mitigation guidance

Attack path visualization

Graph-based attack path mapping with kill-chain analysis and lateral-movement detection. Severity-by-likelihood heat maps show where to focus first.

Compliance gap analysis

See which controls each threat affects across HIPAA, GDPR, PCI DSS, SOX, FISMA, NIST SP 800-53 and ISO 27001.

Indicates where threats affect controls. It is not an audit, certification or guarantee of compliance.

Collaboration and DevSecOps handoff

Owner, editor and viewer roles per threat model, real-time collaborative editing, comment threads and audit trails. Push findings to Jira, GitHub Issues or Azure DevOps, notify Slack or Teams, and schedule PDF or DOCX reports by email.

Architecture Threat Modeling

Architecture Threat Modeling that starts with system design

Many serious security weaknesses come from how a system is put together, not from a single line of code. Architecture Threat Modeling examines the design itself: what talks to what, which data moves where, and where trust changes along the way.

Conceptual architecture threat model Web and mobile clients on the internet reach a WAF and API gateway at the edge. The gateway calls an auth service and an orders API in the application tier; the orders API hands work to a background worker. The auth service reads a customer database, the orders API writes to object storage, and the worker calls an external payment provider. Dashed red lines mark five trust boundaries. Threat markers flag spoofing on client traffic, denial of service at the gateway, elevation of privilege on the auth call, tampering on the orders call and information disclosure on the database flow. INTERNETEDGEAPPLICATION TIER DATA TIERTHIRD PARTY Web clientMobile app WAF +API gateway Auth serviceOrders APIBackground worker Customer DB (PII)Object storagePayment provider S D E T I
  • Data flow
  • Trust boundary
  • STRIDE threat: Spoofing, Tampering, Information disclosure, Denial of service, Elevation of privilege
Conceptual example of an architecture threat model, not a product screenshot.

What an architecture-level model helps you reason about

Components and their interactions
Services, clients, gateways and stores, and which ones call each other.
Data flows and trust boundaries
Where sensitive data travels, and where it crosses from one level of trust to another, which is where many threats concentrate.
External systems and dependencies
Third-party services and identity providers your design relies on.
Attack surface and attack paths
How an attacker could move from an entry point towards valuable assets.
Mitigation decisions
Which controls address each threat, and who owns the work.

In Astrix

Each threat model captures the system description, components and data flows you import or describe. The selected framework's analyzer identifies threats against them, and attack path visualization and mitigation guidance turn the model into decisions.

How Astrix works

Five steps from design to mitigation

  1. Import

    Bring in an API spec, infrastructure-as-code, a diagram or repository, or describe the system in plain language.

    Start from real design artifacts instead of redrawing the architecture by hand.

  2. Analyze

    Choose STRIDE, PASTA, VAST, SAIF or MAESTRO. The matching analyzer identifies threats against your components and data flows.

    Every team applies the same methodology the same way.

  3. Prioritize

    Each finding carries a CVSS score with CWE and MITRE ATT&CK mapping. Heat maps and attack paths show where risk concentrates.

    Effort goes to the threats that matter most.

  4. Mitigate

    Apply mitigation guidance, push findings to Jira, GitHub Issues or Azure DevOps, and track remediation status.

    Findings become owned work items, not slides.

  5. Review

    Re-analyze as the design changes, keep snapshots of each model, and export PDF or DOCX reports.

    The threat model stays in step with the architecture.

Methodologies and frameworks

Five threat modeling methodologies, one platform

Different systems call for different approaches. Astrix supports established methodologies for traditional applications alongside newer frameworks for AI and multi-agent systems, each with its own analyzer.

Threat modeling methodologies supported by Astrix
MethodologyWhat it isTypically used for
STRIDESix threat categories: spoofing, tampering, repudiation, information disclosure, denial of service and elevation of privilege.General application and system architectures
PASTAProcess for Attack Simulation and Threat Analysis: a seven-stage, risk-centric method that ties business objectives to technical security requirements.Risk-driven reviews of business-critical systems
VASTVisual, Agile and Simple Threat modeling, built to scale across large organizations by separating application and operational threat models.Enterprise-wide threat modeling programs
SAIFGoogle's Secure AI Framework, a six-pillar framework for AI and machine-learning systems.AI/ML systems and LLM-powered applications
MAESTROAn emerging seven-layer framework for securing multi-agent AI architectures and autonomous systems.Agentic and multi-agent AI architectures

Finding enrichment

  • CVSS scoring
  • CWE
  • MITRE ATT&CK

Compliance gap analysis

  • HIPAA
  • GDPR
  • PCI DSS
  • SOX
  • FISMA
  • NIST SP 800-53
  • ISO 27001

Framework and standard names belong to their respective owners. Support means Astrix includes an analyzer or mapping for them; it does not imply endorsement or certification by their authors.

Why Astrix

From one-off workshops to a repeatable threat modeling practice

Astrix keeps the judgment of your security reviewers and replaces the manual work around it.

Manual threat modeling

  • Architecture redrawn by hand for every review
  • Method depends on who runs the session
  • Severity judged case by case
  • Findings kept in slides and spreadsheets
  • Model outdated after the next change

With Astrix

  • Import specs, IaC and diagrams you already maintain
  • A dedicated analyzer per framework, applied consistently
  • CVSS scores with CWE and MITRE ATT&CK mapping
  • Findings pushed to Jira, GitHub Issues or Azure DevOps and tracked
  • Re-analysis and snapshots as the design evolves

Use cases

Where teams use Astrix

Explore the threat modeling use cases in detail.

New service and platform design

Threat-model a design before it is built, while changes are still cheap to make.

Release and change reviews

Re-analyze when the architecture changes and compare against earlier snapshots.

Cloud and infrastructure-as-code reviews

Analyze Terraform, CloudFormation, ARM and Kubernetes definitions for design-level threats.

API and integration design

Start from an OpenAPI specification to review how an API exposes data and functionality.

AI, LLM and agentic systems

Apply SAIF and MAESTRO to AI applications and multi-agent architectures.

Security consulting and partners

Run threat models for multiple client organizations, each kept in its own isolated database.

Who it is for

Built for the people accountable for secure design

CISOs and security leaders

Visibility into design-level risk

Consistent, structured threat models and reports across teams, so you can see which threats were found, scored and addressed.

Security and enterprise architects

A structured method, every time

Choose the right framework per system and reason about components, data flows and attack paths in one place.

AppSec and product security

Scale reviews beyond a few experts

Generate a first-pass threat model from existing artifacts, then spend expert time on review and prioritization.

Engineering and DevSecOps

Findings where you work

Mitigations arrive as Jira, GitHub Issues or Azure DevOps items with clear guidance, plus Slack or Teams notifications.

About Astrix

Threat modeling you can put through your own security review

Astrix is a threat modeling platform from Cognexaa. Because customers share details of their architecture with it, the platform is designed to keep that information isolated and under the customer's control. We also threat-model our own platform with Astrix.

  • A dedicated database per customerEvery customer organization gets its own database: no shared tables and no cross-tenant queries.
  • Hosted in the EURuns on enterprise-grade cloud infrastructure. Dedicated deployments in other regions are available on request.
  • Never used to train AINeither Astrix nor its default AI provider trains models on your content or outputs.
  • Encrypted in transit and at restData is encrypted in transit and at rest using industry-standard encryption.
  • You control accessSingle sign-on via OpenID Connect, multi-factor authentication, role-based access and audit trails.
  • You own your dataExport threat models and reports at any time. On exit, your data is deleted and deletion is confirmed in writing.

Resources

Threat modeling FAQ

What is threat modeling software?

Threat modeling software helps teams describe how a system is built, identify what could go wrong from a security perspective, and decide what to do about it. Astrix does this from the architecture artifacts your teams already maintain, such as API specifications, infrastructure-as-code and diagrams, and produces a structured threat model with scored findings and mitigation guidance.

What is architecture threat modeling?

Architecture threat modeling examines a system at the design level: its components, how data flows between them, where trust boundaries sit, and which external services it depends on. It identifies threats that come from how the system is put together, which code-level testing alone can miss. See an example.

Why should threat modeling happen during system design?

Design decisions are easiest to change before they are built. Finding a missing control or an exposed data flow at the design stage lets teams fix it in the architecture rather than reworking code and infrastructure after release.

How does Astrix support the threat modeling process?

You import or describe your architecture, choose a framework, and Astrix generates a threat model. Each finding carries a CVSS score, CWE and MITRE ATT&CK mapping and mitigation guidance. You can visualize attack paths, push findings to Jira, GitHub Issues or Azure DevOps, track remediation, re-analyze as the design changes and export PDF or DOCX reports. See how it works.

Which threat modeling methodologies does Astrix support?

Astrix includes a dedicated analyzer for each of five methodologies: STRIDE, PASTA, VAST, SAIF and MAESTRO. Findings are enriched with CVSS scoring, CWE identifiers and MITRE ATT&CK mapping. Astrix also offers compliance gap analysis against HIPAA, GDPR, PCI DSS, SOX, FISMA, NIST SP 800-53 and ISO 27001; this indicates where threats affect controls and is not a certification or guarantee of compliance. Compare the frameworks.

Does Astrix use AI, and what data does the AI see?

Yes. By default, analysis uses an enterprise AI service hosted in the EU, and every request passes multiple guardrail layers before and after the model call. Only the system description, components and data flows you submit are sent; user identities, evidence files and credentials are never sent. Customer data is not used to train AI models. AI output is advisory and is reviewed and accepted by your users. If no AI provider is configured, a built-in rule engine runs the analysis inside Astrix.

Who should use Astrix?

Astrix is built for security architects, application and product security teams, enterprise and solution architects, DevSecOps and engineering teams, security leaders who need consistent visibility into design-level risk, and security consultancies that run threat models for clients. See how each team benefits.

How can my organization see and evaluate Astrix?

Book a demo by emailing sales@cognexaa.com. Astrix is offered as Starter, Professional and Enterprise plans, with pricing available on request. The Cognexaa team can also walk your security reviewers through the platform's architecture and controls under NDA.

Bring threat modeling into your architecture workflow

See how Astrix can support your team's approach to architecture security and threat analysis, using a system that matters to you.

sales@cognexaa.com · Already a customer? Sign in