diff --git a/images/whitebox_dynamic_testing.png b/images/whitebox_dynamic_testing.png
new file mode 100644
index 0000000..f2e8d22
Binary files /dev/null and b/images/whitebox_dynamic_testing.png differ
diff --git a/white-box-pentest/quickstart.mdx b/white-box-pentest/quickstart.mdx
index e37c0ac..ee549f5 100644
--- a/white-box-pentest/quickstart.mdx
+++ b/white-box-pentest/quickstart.mdx
@@ -12,9 +12,8 @@ Before you start, make sure the organization has:
- A connected repository or an uploaded archive to assess
- The branch or commit you want Hacktron to use
-- Any target URLs that should be tested, such as staging or production URLs
-- Authentication instructions for protected areas, if the application requires
- sign-in
+- For dynamic testing, the domains or URLs of the applications to test and
+ instructions for accessing each one, even if no sign-in is required
- Enough Whitebox Scan credits, or an organization owner who can add a payment
method and buy credits
@@ -47,13 +46,38 @@ On-premises deployments can select an enabled GitHub Enterprise Server repositor
-
- Add target URLs, login steps, test credentials, areas to emphasize, and any
- exclusions. Keep credentials scoped to a test account where possible.
+
+
+ **Dynamic testing is optional.** Choose **Skip dynamic testing** to
+ continue with a source-code assessment without testing a deployed
+ application.
+
-
+ In **Define target**, add each application you want Hacktron to test:
-
+ 1. Enter its **Domain**, such as `https://staging.example.com`.
+ 2. Fill in **Auth instructions** with login steps, dedicated test credentials,
+ and how to confirm successful access. For a public site, state that no
+ login is required and describe the page that should load.
+ 3. Click **Verify access**. Hacktron follows your instructions in a browser;
+ a live access preview lets you watch the attempt. Wait for the target to
+ show **Verified** before continuing.
+
+ Use **Add target** for another application. Each target needs its own domain
+ and instructions. With multiple targets, **Verify all** checks the targets
+ that still need verification.
+
+
+
+ If verification fails, check the domain, credentials, and instructions, then
+ retry. Verification lasts for one hour. If you edit a verified target's
+ domain or instructions, verify it again or choose **Revert to last success**.
+
+
+
+ In **Context documents**, add architecture notes, threat models, API specs,
+ or other material that helps explain the application. Then review the
+ repositories, targets, and access instructions in **Review details**.
@@ -82,6 +106,42 @@ On-premises deployments can select an enabled GitHub Enterprise Server repositor
+## Writing auth instructions
+
+Write instructions for each target as if someone were signing in for the first
+time. Include:
+
+- The starting URL and the steps to reach the application
+- Test-account credentials and the role or permissions of each account
+- How to obtain an MFA or OTP code, if required, and any SSO, CAPTCHA, or
+ network access requirements
+- What successful access looks like, such as a dashboard or account page
+- Areas to focus on and actions to avoid, such as payments or deleting data
+
+For a target that requires sign-in:
+
+```text
+1. Open https://staging.example.com/login.
+2. Sign in with qa-user@example.com and [dedicated test password].
+3. Confirm the dashboard loads and the account has the Member role.
+4. Focus on project access and file uploads.
+5. Do not make payments, delete projects, or modify other users' data.
+```
+
+For a public target:
+
+```text
+1. Open https://help.example.com.
+2. No login is required.
+3. Confirm the help center and search page load.
+```
+
+
+ Auth instructions are shared with AI agents for testing. Use a test
+ environment and dedicated test accounts, not personal credentials or
+ production secrets. Revoke test credentials after the assessment.
+
+
## Scope checklist
| Item | What to provide |
@@ -93,11 +153,6 @@ On-premises deployments can select an enabled GitHub Enterprise Server repositor
| Exclusions | Systems, data, tenants, destructive actions, or rate limits that are out of scope. |
| Context documents | Architecture notes, threat models, API specs, or prior pentest reports. |
-
- Do not provide personal user credentials or production secrets. Use dedicated
- test accounts and revoke them after the assessment.
-
-
## API option
You can also start Whitebox Scans from the REST API. The API follows the