Alfred AI

New CVE tests, built by AI and verified by researchers

Alfred AI turns published CVEs into payload-based tests, and a Detectify security researcher reviews each one before it goes live.

Book a demo Start a trial

How Alfred AI works

Watches for published CVEs

Alfred AI monitors public CVE disclosures and various sources to find published CVEs with a working public proof of concept.

Monitors public CVE disclosures
Checks other public sources too
Looks for working public proofs of concept

Turns a proof of concept into a real test

Alfred AI builds a payload-based test from the public proof of concept, like the test modules our researchers and Crowdsource community build. We've set it up to focus on web-facing vulnerabilities.

Payload-based test modules
Built from public proofs of concept
Best at web-facing vulnerabilities with documented exploits

Every test verified by a researcher

Before an Alfred AI test goes live, a Detectify security researcher reviews it. It gets the same quality bar as the tests our researchers build by hand, so Alfred's false positive rate is in line with the rest of Detectify.

Human review before every test goes live
Same quality bar as hand-built tests
False positive rate in line with the rest of Detectify

Findings your team can act on

Findings come with a risk level, a description and references, so your team can prioritize. More and more new tests also include steps to fix the issue. Alfred AI also updates the risk, description, remediation and references on older tests, so findings carry the latest information on the vulnerability and triage stays quick.

Risk level, description and references on every finding
Remediation steps on more new critical and high-severity tests
Older tests updated over time

Alfred AI fills the gaps, researchers hunt for the rest

Alfred AI doesn't replace the tests from Crowdsource, our community of ethical hackers. It complements coverage for published CVEs, so Detectify's researchers and the Crowdsource community can focus on trending vulnerabilities and novel research.

Complements the Crowdsource community
Researchers focus on novel vulnerabilities
99% of found vulnerabilities have no CVE assigned, based on our own data over the past three years

Alfred AI covers the published CVEs. The crowd finds the rest

Our community of ethical hackers submits vulnerabilities that often have no CVE yet. Each one becomes a test that runs next to the tests Alfred AI builds.

1,765+test modules built by ethical hackers

99%of found vulnerabilities have no CVE assigned, based on our own data over the past three years

More about Crowdsource

I trust that Detectify will always outpace us in identifying new threats. It gives us peace of mind knowing that Detectify will find any new issues.

JensChief Security Officer, Kivra

Read case study

FAQs

01What does Alfred AI cover?

CVEs that have a public proof of concept. We've set Alfred AI up to focus on web-facing vulnerabilities, and it builds a payload-based test for each one.

02How does Alfred AI avoid false positives?

A Detectify security researcher reviews every Alfred AI test before it goes live, with the same quality bar as the tests our researchers build by hand. Tests built by Alfred AI have a false positive rate in line with the rest of Detectify.

03Does every finding include remediation steps?

Every finding comes with a risk level, a description and references. Remediation steps are included on more and more new tests, and older tests are updated over time.

04Where can I see Alfred AI's new tests?

New tests from Alfred AI, Crowdsource and our internal researchers are listed in the changelog.

05How does Alfred AI differ from Crowdsource?

Crowdsource is our community of ethical hackers, who find and submit vulnerabilities, many of them with no CVE. Alfred AI builds tests for CVEs that are already public. It adds to Crowdsource tests and doesn't replace them.