> For the complete documentation index, see [llms.txt](https://docs.medcrypt.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.medcrypt.com/msi/review-your-threat-model.md). # Manage risk MSI builds a STRIDE threat model of your device from its documents in a single automated run. It maps components, data flows, and trust boundaries, identifies and scores threats, and matches the controls that mitigate them. You then review the result, refine it with comments, and export a report for your submission. STRIDE is a standard way to sort security threats into six categories: Spoofing, Tampering, Repudiation, Information disclosure, Denial of service, and Elevation of privilege. ## Before you begin * Your organization's plan must include Threat Modeling. * Select the device and device version you want to model. Each threat model belongs to one device version. * Upload your design and architecture documents for the device. The threat model reads them, and controls are only counted as documented when MSI finds them in these documents. ## Start a threat model 1. In the sidebar, select **Threat Model**. If the selected version has no threat model yet, you see the start form. It shows the **Risk Formula** set for this device version and an estimated run time. 2. Optional: select **View device assets** to check the device's assets on **Device Knowledge** before you run. Threats are scoped to these assets. 3. Select **Generate Threat Model**. The run goes from your architecture to a scored threat and control register without stopping for approval. You review the results afterward, row by row, in the Risk Register. To start another run later, use the run selector in the page header and select **+ New Threat Model**. You can use the same selector to switch between previous runs. ### Start from an existing threat model in Excel If you already have a threat model in a spreadsheet, you can import it instead of starting from scratch. 1. On the start form, select **Import from Excel**. 2. Select an Excel (.xlsx, .xls) or CSV file, then select **Import**. Assets, threats, controls, and scores are read from the file. Anything the workbook already covers is kept as is, and MSI fills in the rest. 3. Review the import summary, which counts the assets, threats, controls, and scores that were imported. 4. If the summary says workbook elements need your decision, select **Open the Contradictions tab** to resolve them on **Device Knowledge**. Nothing in your existing architecture is overwritten. 5. Select **Start generation**. ## Understand the results When the run completes, the **Threat Model** page opens on the latest report. It has four tabs. | Tab | What it shows | | ----------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Traceability** | Whether each threat has an evidenced control and links to your device requirements. The table lists each **Threat** with its **Controls** and **Device requirements**. Use the filter to focus on, for example, **Unmitigated Threats** or **High/Critical Risk Gaps**. | | **Scope & Assumptions** | **Documented Exclusions** (threats retired with a reason; they stay in the record and a later run cannot silently bring them back) and the **Scope Assumptions** the analysis relied on. | | **Architecture** | **Data Flow Diagrams**, **Trust Boundaries**, and **Threat Actors**. | | **Comments** | Feedback you've added for the next run, and the regenerate controls. | {% hint style="info" %} A control counts as coverage only when it's backed by evidence, such as a quote MSI verified in your documents. Controls the AI suggests but can't find in your documents are shown as suggestions, not as mitigation. Review them in the Risk Register. {% endhint %} To act on individual threats (approve, retire, rescore, or add controls), select **View Risk Register**. ## Refine the model with comments Comments guide the AI when you regenerate the threat model. 1. On the **Comments** tab, select **Add Comment**. 2. Choose a **Target Type**: **General / Session**, **Asset**, **Threat**, **Control**, or **Data Flow Diagram**. For anything other than a general comment, pick the **Target**. 3. Choose an **Action Type**: **Note**, **Accept**, **Reject**, or **Modify**. 4. Enter your **Comment** and select **Add Comment**. 5. Optional: to add documents, select a collection under **Source Documents** and upload the new files there. 6. Select **Regenerate Threat Model**. When you have comments that haven't been used yet, the button reads **Regenerate with N New Comments**. Regenerating creates a new version of the threat model with your comments and any new documents applied. Comments that were used are marked **Incorporated**. ## Compare two runs 1. On the **Threat Model** page, select **Compare Threat Models**. 2. Choose the **Baseline (from)** and **Comparison (to)** runs. Only completed runs are listed. 3. Select **Compare**. The **Threat Model Comparison** page shows what changed between the two runs on the **Coverage**, **Traceability**, and **Scope** tabs. ## Export the report 1. On the **Threat Model** page, select **Download Report**. 2. Choose a format: **JSON**, **Markdown**, **PDF**, **Excel (XLSX)**, or **Word (.docx)**. ## Next steps * [Triage risks in the Risk Register](/msi/review-your-threat-model/triage-risks-in-the-risk-register.md): review, score, and approve each threat, and close control gaps. * [Answer guided questions](/msi/get-started/answer-guided-questions.md): your answers feed the next threat model run. * [Compare your threat model with pentest or scanner findings](/msi/review-your-threat-model/compare-with-pentest-or-scanner-findings.md): check whether the model predicted what testing found.