When should I use the Evaluative Research skill?

Use it when you need to decide between multiple viable technical options, such as libraries, frameworks, or architectural patterns, and want a structured, documented rationale for your team.

How does the skill ensure objective recommendations?

The methodology forces the AI to prioritize evidence over opinion, requiring every rating in the comparison matrix to be traced back to documentation, benchmarks, or real-world usage data.

Does it support custom decision criteria?

Yes, Phase 1 of the framework specifically identifies 'Decision Drivers' and weights them according to your specific project needs, such as performance, cost, or developer experience.

Can I use this skill for debugging or root cause analysis?

No, this skill is optimized for forward-looking decisions. For debugging or finding root causes in existing code, use an investigation-methodology or gap-analysis skill.

What does the final output look like?

It generates a comprehensive 'Investigation' document in Markdown format, featuring a summary, decision context, candidate evaluations, a comparison matrix, and a clear primary recommendation.

Evaluative Research & Tech Comparison

Name: Evaluative Research & Tech Comparison
Author: ichabodcole

byichabodcole

0•

生産性とワークフロー

Conducts structured, evidence-based evaluations of technologies, libraries, and architectural patterns to facilitate informed decision-making.

This skill provides a rigorous framework for Claude to help users navigate complex technical choices by moving beyond surface-level comparisons. It guides the AI through a multi-phase process: defining non-negotiable constraints, scanning the technology landscape for viable candidates, performing deep-dive evaluations of finalists, and synthesizing findings into a weighted comparison matrix. By emphasizing evidence-backed rationale and explicit trade-off analysis, this skill ensures that architectural shifts and library selections are documented, objective, and aligned with specific project drivers like performance, maintenance, or team expertise.

主な機能

01Comprehensive risk and effort assessments for every candidate option.

02Standardized 'Investigation' documentation output for project records.

030 GitHub stars

04Four-phase evaluation framework: Define, Scan, Evaluate, and Analyze.

05Weighted comparison matrices for objective, multi-criteria scoring.

06Logic for identifying second-order effects and 'when to reconsider' triggers.

ユースケース

01Choosing between competing libraries (e.g., selecting a rich-text editor or state management tool).

02Assessing long-term architectural migration paths and technology stacks.

03Evaluating 'build vs. buy' decisions for internal infrastructure or services.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ichabodcole/project-docs-scaffold-template evaluative-research

For use in Claude.ai and ChatGPT

主な機能

01Comprehensive risk and effort assessments for every candidate option.

02Standardized 'Investigation' documentation output for project records.

030 GitHub stars

04Four-phase evaluation framework: Define, Scan, Evaluate, and Analyze.

05Weighted comparison matrices for objective, multi-criteria scoring.

06Logic for identifying second-order effects and 'when to reconsider' triggers.

ユースケース

01Choosing between competing libraries (e.g., selecting a rich-text editor or state management tool).

02Assessing long-term architectural migration paths and technology stacks.

03Evaluating 'build vs. buy' decisions for internal infrastructure or services.