# Threat Model and attack paths

The Resolve Platform's Threat Model: an editable diagram of your architecture with threats placed on it, attack paths overlaid, and OSCAL and evidence exports.

**Overview › Threat Model** (`/resolve/threat-model`) lays your architecture out across trust zones, places threats where they cross a boundary, and overlays the attack paths that could reach them. Use it to show how someone would get in, and to hand the result to an auditor as evidence.

What the diagram is built from, and the annotation kinds you can attach, are described in [Asset relationships and labels](/vulnerability-management/scope/asset-relationships.md). This page covers the Resolve Platform page itself.

## Building the diagram

A threat model is a saved document, not a live view. **Build this model** seeds zones, elements and flows once from your latest [vulnetix analyze](/cli/commands/analyze.md) results. The first time you open the page, a model scoped to every analysed repository is made for you.

From then on the diagram is yours. Move elements and zones, draw and relabel flows, and undo or redo with `ctrl+z`, `ctrl+y` or `ctrl+shift+z`. **Rebuild** merges a later scan in without rearranging anything. Elements the diagram has never held land in the **New since last review** tray. Drag an item onto the canvas, place it with its button, or use **Place all** to put every tray item into the zone it was classified into.

## Reading it

| Control | Does |
|---|---|
| Kind filter | Shows or hides **Repository**, **Service**, **Container / OCI**, **Infrastructure**, **Route**, **Workflow** and **Contributor**. This changes what is on the diagram and lays it out again. |
| Methodology | Re-labels every threat card as **STRIDE**, **LINDDUN**, **DREAD**, **Attack stage** (PASTA) or **OCTAVE**. Cards keep their position. |
| Attack path | Lights the elements each stage of the selected path correlates with and dims the rest. Nothing moves. |
| **Annotate** | Arms a tool so a click places an annotation instead of selecting |

Selecting anything (element, route, flow, zone or threat card) opens a detail panel with its source, repository, when it was last observed, and every related object. Each related object has a button that moves the selection and the view to it, so you can trace a path by clicking.

Threat cards are your findings, attached to the element they affect.

## Attack paths

An attack path is a named chain of 2 to 5 stages. It is the same record the `ATTACK_PATH` risk metric ranks on, so editing a path here changes the order of your findings queue. See [attack paths in risk strategies](/vulnerability-management/prioritisation/risk-strategies.md#attack-paths) for the record itself.

The **Attack paths** picker opens showing only paths with real evidence in the repository (at least two stages correlated), strongest first, with how many stages have evidence. **Show all paths** reveals the rest.

| Action | Does |
|---|---|
| **New** | Opens the attack path editor |
| Edit | Edits an organisation path. A Vulnesis blueprint is read-only: editing one makes an organisation-owned copy first. |
| **Start from an ATT&CK campaign** | Seeds a new path from a MITRE ATT&CK campaign's techniques. The preview shows which techniques were dropped, and a campaign too narrow to make two stages is listed with the reason. |

## Exports

| Export | Contains |
|---|---|
| **PNG** | The diagram as drawn |
| **OSCAL** | The model as OSCAL JSON |
| **Evidence pack** | Diagram, attack path, annotations, correlation, OSCAL and VEX, as a zip |

## Plan

The Threat Model needs **Pro** or above. Below Pro the page shows a sample behind an upgrade prompt.
