Default-credential checks
Deterministic nuclei default-logins run on demand (no LLM cost, no surprises).
Last updated May 27, 2026
What it is
A non-LLM, fully-deterministic check that runs the nuclei
default-logins template tree against an authorised service. No
reasoning loop, no API tokens spent: just the same well-known
weak-credential checks every pentester runs first, automated and
audit-logged.
What it produces
For each service checked:
- A list of any matched default-credential combinations (with the exact request/response that proved it)
- One finding per match, recorded in the same
findingstable as AI-driven discoveries, so they show up in the same dashboard and reports - The matched product’s category badge (“Web admin panel”, “DB admin”, “Router admin”, etc.)
What to use it for
- New-acquisition triage: “did they leave any boxes on default admin passwords?”
- M&A due diligence: cheap, audit-ready evidence the surface is / isn’t credential-exposed
- The pre-AI quick win: scan a target, immediately default-check, THEN spend AI tokens on the harder CVEs
Why it matters
Default credentials are still in the OWASP Top 10 and CISA’s Known-Exploited Vulnerabilities catalogue. Most security tools don’t check them because the checks look “noisy”, but tools that DO check them find them constantly. We make the check first-class and cost-free at runtime.
Why it’s valuable
- No LLM token cost. Free to run on every service.
- Deterministic. Run it twice on the same service, get the same result. Patch and re-run, the row flips green.
- Same findings pipeline. Default-credential hits get the same ticketing / sharing / replay treatment as AI verdicts.
How to use it
- Open a service detail drawer.
- Click Run default-credential test.
- Result lands in the findings dashboard within seconds.
Or via the API:
POST /api/default-logins
{ "service_id": 12345 }
You authorise the test; we record that you did. Don’t run this against services you don’t own. That’s a CFAA / Computer Misuse Act issue, and the checkbox is a record we asked.