evaluation-design
OfficialDesign rigorous AI safety evals with a rubric.
AuthorEquiStamp
Version1.0.0
Installs0
System Documentation
What problem does it solve?
Phase 3 evaluation design for AI safety often requires structured, rubric-aligned design artifacts. This Skill helps users generate rigorous evaluation designs that align with the 7-dimension rubric and Phase 2 outputs to ensure consistent, defensible assessments.
Core Features & Use Cases
- Guides users to craft an Evaluation Question, Instrument Design, Score Against Rubric, and Operational Checklist.
- Supports both from-scratch design and review of existing evaluation designs for risk-grounded decisions.
- Produces a complete Evaluation Design Document including tier justification, inputs, scoring, and exclusion criteria, plus a final scorecard.
Quick Start
Provide Phase 2 targets and request a Phase 3 design or review, and receive a complete Evaluation Design Document.
Dependency Matrix
Required Modules
None requiredComponents
Standard package💻 Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: evaluation-design Download link: https://github.com/EquiStamp/evaluating-evaluations/archive/main.zip#evaluation-design Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 223,000+ vetted skills library on demand.