> For the complete documentation index, see [llms.txt](https://copilot-docs.bugbase.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://copilot-docs.bugbase.ai/enterprise/how-to-trigger-a-web-assessment/web-assessment-phase.md).

# Web Assessment Phase

Test discovered trajectories with selected attack vectors and validate the findings.

The Assessment phase actively tests what [Web Discovery](https://copilot-docs.bugbase.ai/enterprise/how-to-trigger-a-web-assessment/web-discovery.md) found. Pentest Copilot generates test cases for each trajectory and attack vector, sends them through each selected authentication context, and reports only findings it could validate.

In a **Discovery + Assessment** run, this phase starts automatically when Discovery finishes and tests every eligible trajectory that run found. You only need the separate **Assessment** intent, described below, to **hand-pick which trajectories** to test or to re-test without mapping the application again.

## Before You Start

* Discovery has completed for the domain. Assessment does not crawl for new flows; it tests the trajectories already stored.
* The trajectories you want to test are visible in **Trajectory scope** (see below). If the list is empty, run Discovery first.
* Any browser sessions you select are validated. Expired sessions block the run on **Review**.
* Active testing is approved for the target, and production-sensitive flows such as payments or account deletion are excluded.

## What Gets Tested

Assessment tests **trajectories**: recorded user flows that contain at least one API request.

| Trajectory type       | What it is                                                                  | How it is tested                                                                                    |
| --------------------- | --------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- |
| **Single API**        | A flow with exactly one API request.                                        | The request is tested on its own.                                                                   |
| **Cleaned multi-API** | A flow with two or more related requests, reduced to the steps that matter. | The requests are tested together, so state carried between steps (IDs, tokens, carts) is preserved. |
| **Grouped multi-API** | Related API requests grouped automatically into one business flow.          | Tested as a sequence, like a cleaned flow.                                                          |

Some trajectories are left out automatically and never appear in **Trajectory scope**:

* duplicates of a flow that is already listed;
* raw multi-API flows that have a cleaned version;
* single requests for static data files (`.js`, `.json`, `.jsonp`, `.xml`), which are scanned for exposed secrets instead;
* single-API trajectories blocked by your [Settings -> Trajectories](https://copilot-docs.bugbase.ai/enterprise/settings/trajectories.md) rules.

## Run an Assessment

{% stepper %}
{% step %}

### Choose the Assessment intent

Open **Modules -> Web Assessment -> Run Assessment**, select the domain on **Asset**, then select **Assessment** on **Intent**.

<figure><img src="https://232193438-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FwolEZzMm5QD9NoFKutSj%2Fuploads%2Fgit-blob-351d05b4be6b70ee22770975b4362b16962d7058%2Fweb-intent-assessment.webp?alt=media" alt="Intent step with Assessment selected"><figcaption><p>Assessment tests the existing inventory without running Discovery again.</p></figcaption></figure>

{% hint style="info" %}
**Discovery + Assessment** also runs this phase, automatically, right after Discovery. It tests every eligible trajectory found by that Discovery run, so the **Trajectory scope** picker is not shown. Use the **Assessment** intent when you want to pick flows yourself.
{% endhint %}
{% endstep %}

{% step %}

### Choose a configuration

Under **Configuration** on the same step, choose one of the following. If there is no **Configuration** choice, AI configure is not available and the run uses **Manual**.

* **Manual** to hand-pick trajectories in **Trajectory scope**. The rest of this procedure uses **Manual**.
* **AI configure** to let vetting check access and sign in with your accounts first. It has no **Trajectory scope** picker. See [AI Configure and Vetting](https://copilot-docs.bugbase.ai/enterprise/how-to-trigger-a-web-assessment/ai-configure-and-vetting.md).
* **Use existing configuration** to reuse the scan configuration saved by the domain's last scannable vetting run.
  {% endstep %}

{% step %}

### Set up access

On **Scan settings**, under **Access**, set **Authentication coverage**, the optional **Starting URL**, and the browser sessions to use. Each selected session is tested as a separate run, which lets Pentest Copilot compare what each role can do.

<figure><img src="https://232193438-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FwolEZzMm5QD9NoFKutSj%2Fuploads%2Fgit-blob-1d2496fe6962cfcb11f7118d4633e8f9a3fd09c1%2Fweb-scope-authentication.webp?alt=media" alt="Manual Access with Unauthenticated and Authenticated coverage, the starting URL, and the browser-session picker"><figcaption><p>Authentication, starting URL, and trajectory scope are set for the selected domain.</p></figcaption></figure>
{% endstep %}

{% step %}

### Choose the trajectories to test

Still in **Access**, scroll to **Trajectory scope**. It lists every testable flow Discovery found for the domain. All trajectories are selected by default.

If the domain was assessed before with a narrower selection, the panel starts from that saved selection instead and shows **Showing the selection saved from this asset's last assessment.** Use **Select all** to test every flow again.

<figure><img src="https://232193438-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FwolEZzMm5QD9NoFKutSj%2Fuploads%2Fgit-blob-ccf8a5e2e2637627f4bafdba7ba16a3ca537f15a%2Fweb-trajectory-scope.webp?alt=media" alt="Trajectory scope list with the flow type filter open on All flows, Single API, Multi API, Grouped multi-API and Cleaned multi-API, and 8 of 8 trajectories selected"><figcaption><p>Filter by flow type or HTTP method, then deselect what should not be tested.</p></figcaption></figure>

| Control                           | Use it to                                                                                           |
| --------------------------------- | --------------------------------------------------------------------------------------------------- |
| **Search trajectories**           | Find flows by description, ID, host, method, or path.                                               |
| Flow type filter                  | Show **All flows**, **Single API**, **Multi API**, **Grouped multi-API**, or **Cleaned multi-API**. |
| **Methods**                       | Show only flows that use specific HTTP methods, such as `POST` or `DELETE`.                         |
| Checkbox on a row                 | Include or exclude that trajectory.                                                                 |
| **Select page** / **Clear page**  | Include or exclude every trajectory on the current page.                                            |
| **Select all** / **Deselect all** | Include or exclude every trajectory that matches the current filters.                               |
| **View entity**                   | Open the trajectory in the Exploit Graph to see its recorded steps.                                 |

The summary line, such as `337 of 338 selected; 1 deselected`, shows exactly what will be tested.

<figure><img src="https://232193438-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FwolEZzMm5QD9NoFKutSj%2Fuploads%2Fgit-blob-96aed3b518156e052206a3e01109cfbb646f1503%2Fweb-trajectory-multi-api.webp?alt=media" alt="Trajectory scope filtered to Multi API showing cleaned and grouped multi-API flows"><figcaption><p>Multi-API flows test business logic across several requests, such as checkout or complaint submission.</p></figcaption></figure>

{% hint style="warning" %}
Deselect flows that must never be replayed, such as real payments, account deletion, or sending email to real users. To keep a third-party API out of **every** run, add a [trajectory blacklist rule](https://copilot-docs.bugbase.ai/enterprise/settings/trajectories.md) instead.
{% endhint %}

Your selection is saved for the domain and becomes the starting point for its next Assessment run.
{% endstep %}

{% step %}

### Choose attack vectors

Further down **Scan settings**, the **Attack vectors** section lists the vulnerability categories to test. All available categories are selected by default for the broadest coverage. Clear any category that is outside the engagement.

<figure><img src="https://232193438-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FwolEZzMm5QD9NoFKutSj%2Fuploads%2Fgit-blob-7719fc0dde84f2b6214f435b7bd75349685ee5cb%2Fweb-attack-vectors.webp?alt=media" alt="Attack vectors, 16 of 16 selected, grouped into Access control and identity, Injection, Client-side, Server-side and request, and Information and application logic"><figcaption><p>Attack vectors appear only for intents that include Assessment.</p></figcaption></figure>

See [Vulnerability Testing Coverage](https://copilot-docs.bugbase.ai/enterprise/vulnerability-testing-coverage.md) for what each group covers. The rest of **Scan settings**, such as egress and rate limits, is shared with Discovery. See [Common Web Settings](https://copilot-docs.bugbase.ai/enterprise/how-to-trigger-a-web-assessment/configure-scan-settings-for-web-assessment.md).
{% endstep %}

{% step %}

### Review and start

In **Automation**, choose **Run now** or **Schedule**. On **Review**, confirm the domain, intent, and settings, then select **Start run** or **Schedule run**.

<figure><img src="https://232193438-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FwolEZzMm5QD9NoFKutSj%2Fuploads%2Fgit-blob-11d6c7394c0cc46fc7e1c5565de2ebf6198b8b21%2Fweb-review-manual.webp?alt=media" alt="Final configuration review for Manual with Discovery + Assessment: Web Assessment, the domain, the intent, Manual, Run now, and Scan settings cards saved for this domain"><figcaption><p>Review is the last chance to check the domain, intent, and settings before the run starts.</p></figcaption></figure>

If a selected browser session is pending, failed, or expired, **Review** shows **Browser session validation is required** and hides the start button. Validate or re-record the session, then return.
{% endstep %}
{% endstepper %}

## What Happens During the Run

1. **Session check.** Each selected browser session is validated before testing, unless you turned on **Skip browser session validation**.
2. **Test generation.** For each selected trajectory, Pentest Copilot works out which attack vectors apply and generates test cases.
3. **Testing.** Test cases are sent through the trajectory's authentication context at the configured rate.
4. **Validation.** A validator re-checks each candidate finding. Unconfirmed candidates are marked false positives, and repeats are marked duplicates.

Each authentication context appears as its own `WEB_ASSESSMENT` row in **Activity -> Activity**.

## Review the Results

* **Web Assessment -> Attack Paths** lists validated findings with evidence, AI reasoning, remediation, and the full path from the domain to the vulnerability.
* **Web Assessment -> Analysis** shows each domain's scan runs, their outcome and alerts, and statistics by severity and category.
* Use **Retest** on a finding after a fix to replay its path.

See [Analysing Scan Results](https://copilot-docs.bugbase.ai/enterprise/analysing-scan-results.md).

## Troubleshooting

| Problem                                                                                    | Check                                                                                                                                                                                                                  |
| ------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Trajectory scope** says "Run discovery first to create API trajectories for this asset." | Discovery has not produced testable flows for this asset. Run Discovery, add authenticated coverage, or record flows with the Manual Crawler.                                                                          |
| An expected flow is missing from the list                                                  | It may be a duplicate, a static data request, or blocked by a **Settings -> Trajectories** rule. Check the rules with the **Examples** panel.                                                                          |
| The start button is missing on **Review**                                                  | A selected browser session needs validation. Use **Re-configure browser session**.                                                                                                                                     |
| Review shows **Saved for this domain**                                                     | The run reuses the domain's last saved values for that section. Edit the section to apply new values. See [Review](https://copilot-docs.bugbase.ai/enterprise/run-assessment-wizard.md#saved-for-this-domain).         |
| Few or no findings                                                                         | Check the run's outcome in **Analysis**, and confirm in **Activity** that no steps failed and that the run was not cut short by **Max module runtime**. A clean result is not proof that every category was exercised. |
