Black Box, Gray Box and White Box
Every pentest is labeled with its testing access, set from what you provide when the pentest starts:
| Testing access | What you provide | What gets tested |
|---|---|---|
| Black box | No credentials, no source code | What an unauthenticated visitor can reach |
| Gray box | Test credentials, no source code | The authenticated app behind the login |
| White box | Source code (with or without credentials) | Everything above, plus static analysis of your code |
The label is locked when the pentest starts and appears on the pentest page, the report and the attestation letter.
Black box (no credentials, no source)
Tests your application from the outside, like an attacker would. No credentials and no access to source code.
Tools included (11):
| Tool | Category | What it tests |
|---|---|---|
| Port Scanner | Network | Open ports |
| Web Scanner | Web app | XSS, injection, CSRF, misconfigurations |
| Vuln Scanner | Web app | Known CVEs, misconfigurations, exposures |
| Server Audit | Web app | Dangerous files, outdated server software |
| Enumerator | Web app | Hidden directories, backup files |
| Net Scanner | Vulnerability | Full network vulnerability assessment |
| TLS Analyzer | SSL/TLS | Certificate issues, weak ciphers, protocol vulnerabilities |
| Sub Hunter | Recon | Subdomain enumeration |
| Web Probe | Recon | HTTP response probing, tech detection |
| WAF Detect | Recon | WAF detection and fingerprinting |
| Security Checks | Multi | Additional vulnerability checks |
When to use: You do not have a GitHub repository, or you only want to test the external attack surface.
Gray box (credentials, no source)
Everything in black box, run with test credentials you provide, so the AI agents also test the authenticated app behind the login instead of only the public pages.
How to enable: open Add test credentials when you create a pentest and enter a test account (username and password with the login page URL, an API key or header, or OAuth client credentials). Credentials are encrypted at rest and destroyed automatically when your pentest completes.
When to use: Your app is behind a login and you want the authenticated surface covered, but you are not attaching source code.
White box (source code)
Everything in black box, plus static analysis of your source code via GitHub. Attaching source makes a pentest white box even when credentials are also provided; add credentials as well to cover the authenticated app.
Additional tools (3):
| Tool | Category | What it tests |
|---|---|---|
| Code Scanner | SAST | Code-level vulnerabilities, insecure patterns |
| Dep Scanner | SCA | Vulnerable dependencies in package files |
| Secret Scanner | Secrets | Hardcoded API keys, tokens, passwords in code |
Requirements:
- A
repoUrlsupplied on the pentest request - GitHub repository connected via OAuth or GitHub App
- Repository access granted to TurboPentest
The repository is cloned at depth 1 (latest snapshot only), so the code tools scan the current working tree rather than full git history.
When to use: You have a GitHub repository and want comprehensive coverage including code-level vulnerabilities that cannot be found from outside.
Setting up GitHub access
See GitHub Connection for setup instructions.
Running a Pentest
Start a TurboPentest pentest from the dashboard or API - enter a target URL, optionally add a GitHub repo for white-box analysis, and launch the scan.
Multi-Target Engagements
Test several targets in one TurboPentest engagement - each tier includes a target allowance, extra targets use one more credit each, and you get one combined report.