Wiring an Agent Eval Suite Into CI
Set your eval threshold on regression from baseline, not arbitrary pass rates.
Rosalind Wexley
The Eval Register
An Audit Trail for an Agent in a Regulated Workflow
Regulated agents need audit trails that reconstruct every decision, not just log inputs and outputs.
Augustin Whitlaw
Writing Scoring Rubrics That Domain Experts Actually Agree On
Break judgments into measurable parts, then measure whether experts actually agree on them.
Malachy Winterbourne
Running a Structured Red-Team Cycle Against a Production Agent
A three-phase red-team cycle catches vulnerabilities that single-pass testing misses entirely.
Xiomara Wenzel