evaluation-protocol-design

Define metrics, datasets, seeds, baselines, and fairness constraints for evaluation protocols.

11|1|Updated Feb 22, 2026
One-click install
npx skills add https://github.com/EvoClaw/amplify --skill evaluation-protocol-design
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: evaluation-protocol-design
Source: https://github.com/EvoClaw/amplify/tree/main/skills/evaluation-protocol-design
Command: npx skills add https://github.com/EvoClaw/amplify --skill evaluation-protocol-design

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps teams design and lock a comprehensive evaluation protocol for method development projects, ensuring transparent criteria and reproducible comparisons.

Core Features & Use Cases

  • Standardizes metric selection, dataset curation, seed scheduling, and baseline listing to prevent post-hoc rationalization.
  • Guides decision gates for Type M, Type C, and Type H projects, with clear steps and justifications.
  • Enables governance and publication-ready reporting by enforcing a locked protocol document.

Quick Start

Fill in all eight protocol items and save the completed document to docs/03_plan/evaluation-protocol.yaml

Frequently Asked Questions about evaluation-protocol-design

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I design an evaluation protocol for method development experiments?▼

Design an evaluation protocol by locking metrics, datasets, seeds, baselines, and fairness constraints into a standardized document. This prevents post-hoc rationalization and ensures reproducible comparisons for method development experiments.

What should be included in an evaluation protocol to ensure reproducibility?▼

An evaluation protocol for reproducibility must include eight design items: metrics, datasets, seeds, baselines, fairness constraints, and locked decision gates. Store the completed document as evaluation-protocol.yaml.

How do I standardize baseline selection and metric decisions across different project types?▼

Standardize baseline selection and metric decisions by applying locked evaluation protocols with defined decision gates for Type M, Type C, and Type H projects. This ensures consistent criteria from metric selection through baseline inclusion.

Can I use a locked evaluation protocol for publication-ready reporting?▼

Yes, a locked evaluation protocol enables publication-ready reporting by enforcing a locked header and all eight completed design items in the protocol document. This governance ensures transparent criteria and reproducible comparisons.

When do I need to lock an evaluation protocol for my statistics and dataset experiments?▼

Lock an evaluation protocol before running method development experiments to prevent post-hoc rationalization. Defining metrics, datasets, seeds, and baselines beforehand standardizes decisions and ensures reproducible statistics.