Skip to Main Content
Off-by-1 Labs at 1Password

Closing the
gap between assumption and reality

Security research with scientific rigor

Data or it didn’t happen

At Off‑by‑1 Labs we believe in producing human‑driven security research, sharpened by the intentional use of technology, resulting in robust and repeatable findings.
To accept anything less would be a regression to the mean.
1Password Off-by-1 Labs
Recent research

AI-generated vulnerability patches still require expert human review

Read the blog article

54%of AI-generated patches either do not fix a vulnerability, add a new vulnerability, or both.

33%+of AI-generated patches that fix a vulnerability are fragile.

20%of AI-generated patches change the underlying application logic when they successfully fix a vulnerability.

17%of AI-generated patches fix the underlying vulnerability in a robust manner.

Resources to explore

The Access-Trust Gap Report 2025

34% of apps operate outside your identity provider. Learn what that means for businesses navigating regulatory scrutiny and AI adoption.

Read the report

Enforce device posture everywhere

Block untrusted devices from any web app – whether managed, unmanaged, or outside SSO.

Learn more

Explore product demos and guided tours

Secure your company’s data from credential attacks and manage shadow IT risks with our industry‑leading security model.

View demos
Abstract illustration with diagonal lines and geometric shapes in teal and gold tones

Discover high‑risk accounts

1Password Enterprise Password Manager (EPM) and SaaS Manager extend governance beyond SSO and reduce risk from shared and sensitive credentials.

See how

When AI patches vulnerabilities,
what can go wrong?