2026-10-11 17:09 UTC

red-teaming

band: coolmomentum: stable score: 0.008
temperature history

Episodes (2)

The New York Times reports that a mistake by Irregular materially derailed security evaluations conducted for OpenAI, Anthropic, and Meta, exposing a need for tighter controls over third-party frontier-model testing.
expiredconvergesscott: high
Taper maintainer Walex4 reports that its published adversarial harness exposed five now-fixed authorization bypasses in adapter classification and AWS resource validation, demonstrating credential-scope failures that token verification alone does not prevent.
watchingconvergesscott: medium