When should I trigger the skill-improve workflow?

You should use it when a skill feels shallow, provides generic advice, lacks specific implementation details, or when a self-evaluation score falls below an 8 out of 10.

What is a validation artifact?

A validation artifact is a tangible output—like a code component or a completed workflow—created using the improved skill to prove that the new instructions work as intended in a real-world scenario.

What does it mean to 'improve' a Claude Code skill?

Improving a skill involves moving beyond generic descriptions to include concrete values (like specific timeouts or constants), actionable decision trees, and integration with the user's specific development tools.

How does the research phase work?

The skill uses Ref MCP and WebFetch to read actual library documentation and source code, ensuring that the instructions are grounded in the current reality of the framework rather than old training data.

Skill Quality Optimizer

Name: Skill Quality Optimizer
Author: lnittman

bylnittman

Productivity & Workflow

Optimizes and deepens Claude Code skills by transforming shallow instructions into actionable, data-driven expert systems.

About

The skill-improve utility is a meta-layer for Claude Code designed to elevate custom skills from generic descriptions to high-precision tools. It codifies a rigorous workflow of self-evaluation, primary source research, and multi-stage validation to ensure that every skill provides actionable decision trees, concrete constants, and tool-specific integrations. This skill is essential for power users who require their AI-driven workflows to be grounded in actual source code and verified implementation patterns rather than abstract philosophy.

Key Features

Multi-model validation steps to ensure expert-level accuracy
Logic-driven decision tree generation for complex if/then scenarios
Artifact-based verification to prove real-world skill efficacy
Systematic scoring rubric for evaluating skill depth and actionability
Automated research workflow using MCP tools to extract primary source data
0 GitHub stars

Use Cases

Transforming a basic, 'hallucinated' skill into a data-backed expert guide
Fixing skills that provide vague or shallow advice during complex coding tasks
Performing periodic audits and upgrades on a library of custom Claude skills

About

Key Features

Multi-model validation steps to ensure expert-level accuracy
Logic-driven decision tree generation for complex if/then scenarios
Artifact-based verification to prove real-world skill efficacy
Systematic scoring rubric for evaluating skill depth and actionability
Automated research workflow using MCP tools to extract primary source data
0 GitHub stars

Use Cases

Transforming a basic, 'hallucinated' skill into a data-backed expert guide
Fixing skills that provide vague or shallow advice during complex coding tasks
Performing periodic audits and upgrades on a library of custom Claude skills